Rules and standards discussions/Archive/Archive05

From ISFDB

Jump to: navigation, search

This is an archive page for the Rules and standards discussions page. Please do not edit the contents. To start a new discussion, please click here.
This archive includes discussions from January - October 2008.

Archive Quick Links
Archives of old Rules and standards discussions.


1 · 2 · 3 · 4 · 5 · 6 · 7 · 8 · 9 · 10 · 11 · 12 · 13 · 14 · 15 · 16 · 17


Expanded archive listing


Contents

Canonical name for Artists

Authors are pretty simple - most commonly used or recognized. The standard that seems to be developing for artists (now that we have significant data in the system) is the fullest name with credited artwork (Frank Kelly Freas instead of Kelly Freas/Ed Emshwiller instead of Emsh). Does that sound right? If so, Help should be updated.--swfritter 18:26, 14 Jan 2008 (CST)

One reason to go with the longest used form of the name is to avoid disambiguation problems down the road when another, say, "Knight" shows up. On the other hand, there is a fair number of artists who have used an abbreviated form of their name almost exclusively (as far as we know), e.g. Blair. Using a very infrequent form of an artist's name as his/her canonical name just because it's the longest one may lead to a lot of variant titles that could be otherwise avoided. In other words, I am firmly undecided on this one. Ahasuerus 18:58, 14 Jan 2008 (CST)
Perhaps a threshhold? A percentage or "significant number of credits" but not necessarily the most. That would probably mean that Ed Emshwiller is out (Emsh in - my preference) even though he is categorized as the canonical author for four signatures. There are a number of covers assigned to Ed Emshwiller but I think that is because they were entered from secondary sources which is a factor that has influenced perceptions of canonical artist name selection. Frank Kelly Freas, on the other hand, probably is more commonly credited as Kelly Freas, and I think by a wide margin, but early in his career was many times credited as Frank Kelly Freas. Enough times? Sometimes it will just have to be a judgment call. But once someone makes the decision it is hard to go back.--swfritter 20:21, 14 Jan 2008 (CST)

Interior artwork attributions

Artists are generally credited in as many as three places in a magazine - the table of contents, title page of the story being illustrated, and the author's signature on the work. I generally use the fullest name used to credit the author. For instance, Virgil Finlay might be credited as V. Finlay on the table of contents, Finlay on the title page, and sign his name Virgil Finlay. I use Virgil Finlay. If an author is credited on the table of contents with a full name and signs his last name or initials that clearly correspond to that name I use the name in the table of contents. If the fullest credit is a last name I use only the last name. Initial are tricky and require a judgment call. Most editors seem to be using the artist's name for commonly known artists rather than entering initials. In any case, it seems like a good idea to always make a note in the notes of the magazine when signatures and initials are used to determine artist attributions. How close is that to a policy we can agree on? Any additions?--swfritter 18:45, 14 Jan 2008 (CST)

Generally I'm inclined to agree.
An additional issue I've run into is how far credits can be applied. Two different questions come to mind.
  • In Analog for quite a few years, normal practice was to list a credit (normally after the author's name) for illustrations as "Illustrated by Kelly Freas", with no other credits. At other times normal practice was to simply put a credit (typically in italic type) Kelly Freas by the first illustration in the story; sometimes this was done even in those years when an "Illustrated by" credit was normal. Consistently, where I could verify (by signature/initials or by clear style), the illustrations have all been by the artist credited on the first one. I've thus been taking this as a credit for the whole story, without bothering to add notes to a pretty large number of issues for a large number of stories. Is that OK? (I can imagine that in a different magazine this might be done differently, though it would strike me as kind of perverse.)
  • In a very few cases, even where there's an "illustrated by" credit, there are special or unusual illustrations which aren't (to me, anyway) clearly covered by the credit. The two crude diagrams of time travel trips in "The Time-Machined Saga" (serial in Analog), & a couple of maps in other stories, come to mind. I don't have the book The Technicolor® Time Machine to check, so I'm only guessing that those diagrams were Harrison's & appear in the book. I'm not sure I've been at all consistent with this kind of thing, except that I've tried to remember to add a pub note whatever I did.
I own it, and the 'art' is there on pages 57 (diagram), 109 (map), 148 (diagram). All uncredited, so I'd assume Harrison: there's no other artwork so I presume the Schoenherr work is omitted. I'm not keen on recording such minutiae for books (I'll record major interior-art illustrators like Pauline Baynes, but not every picture) but if it helps with a magazine I'm happy to check books for you. BLongley 13:08, 15 Jan 2008 (CST)
Ooh, cool! I'll go look at what I entered in those issues, & at least clear up the pub notes. Thank you!! -- Dave (davecat) 13:47, 15 Jan 2008 (CST)
I'm not trying to quibble, only to toss in some additional complexities before anything is carved in stone. If we were to decide (say) that any illustration in a story with a general credit is covered, barring contrary credit or signature/initials (or outside evidence), I'd not complain. Ditto if we decided that to call it uncredited but leave a note if in any real doubt. Probably ditto for most anything you're likely to suggest, I think. Thanks. -- Dave (davecat) 09:45, 15 Jan 2008 (CST)

Prologues and Epilogues

There's a discussion going on here about how to handle prologues and epilogues. If you have any thoughts or comments that you'd like to add, please check it out. Mhhutchins 17:14, 19 Jan 2008 (CST)

Manga and Takahashi

I've accepted some submissions from Dcarson that delete pubs, which we agree are manga, thus not in the Rules of Acquisition. But someone has gone to a hell of lot to create this author's summary page. Removing all of this info seems a downright shame, seems an awful waste (to quote Sondheim). But I fear we have no choice. Is the person who created this lovely page still around, and can he defend keeping it? Mhhutchins 18:30, 20 Jan 2008 (CST)

We have a number of well researched and lovingly compiled bibliographies that are only tangentially relevant, e.g. William W. Johnstone's page, which consists mostly of westerns. I wouldn't mind having them listed here if we had support for Non-genre Series, but as it is, it is impossible to distinguish bona fide SF titles from other genres :( Similarly, if we had support for "Graphic Novels", we could list them as such and it wouldn't be as misleading as it is now, but significant software changes don't seem to be in the cards at the moment.
Anyway, given the software limitations that we have, perhaps the safe thing to do would be to list all of these "beautiful but alien" bibliographies on a separate Wiki page and revisit them in, say, another 6 months. By then either:
  • the software improves, or
  • we find a loving home for these biblios, or
  • we finally give up and zap them en masse. Ahasuerus 19:17, 20 Jan 2008 (CST)
Another example: Jonathan Gash's Lovejoy books. I know I verified some, can provide cover images, sorted the titles, created some variants, added some pubs: I can't justify keeping them here though. It's useful to me as I still have to find one or two titles, but I have no idea where a "loving home" will eventually be. Again, a candidate for us reaching out to other sites and asking if it's of interest. I'm not going to delete a single one of the pubs, let alone the titles, as they're based around the town I grew up in (for major variations in the value of "grew up"). If we send them away, I want to know where they went. BLongley 20:21, 20 Jan 2008 (CST)
Wow! That is the most impressive Manga-organisation I've seen here, I don't know why I've not come across it before! (I used to practice my editing/moderating/serialisation skills on unverified Japanese authors, it was always a good way to clear up some translators and artists from the Author directory at least.) I agree, it's a waste of effort if we delete it: but if the creator wants the data still we can probably provide a backup to extract it from. BLongley 19:21, 20 Jan 2008 (CST)
I'm not in principle AGAINST Manga being included here: I think I've only actively deleted Non-SF Manga. (E.g. I've obliterated some Basketball Manga.) But it often falls into the "The Moderators are unlikely to be qualified to deal with such new submissions" category that I feel are unwanted here: same as I feel about most foreign titles really, but whereas I'll have a go at including a French or German title, and make an effort over accents and suchlike, I cannot cope with Japanese, Chinese, Russian, or even Polish titles. If there's an English version, I don't mind if it's pictorial so long as it's SF - we could delete, for instance, a lot of Alan Moore and Neil Gaiman works if we enforce the "Graphic Novels are out" rule, only to have to put them back when it's turned into a film and then novelised. I've repeatedly said that we ought to concentrate on what we all do well (basically, works published in English, of all varieties) and forge links to other sites that can do our non-specialities better - so if isfdb.org.jp or isfdb.org.pl or isfsb.org.de want some of our data, let them have it and we can go back to doing what we do best. (Arguing about whether certain titles are missing a "u" in the word "colour" or not, maybe - but at least we don't have to install a new font to argue over those!) BLongley 19:21, 20 Jan 2008 (CST)
If a non-genre author's bibliography is small and looks like an accidental addition to ISFDB then I delete the titles. For example, I recently deleted Kazuo Koike as only a couple of his titles were in ISFDB. If it looks like a human being put a lot of effort into organizing an author's page then I'll let it be. In the past we have talked about making backups and letting people know they are available but nothing has come of it. It does mean ISFDB could creep into comic books, romance, and macro economics in the long run but I suspect it'll also break down into groups of editors that specialize in a genre or author meaning the only impact is server resources. Marc Kupper (talk) 22:31, 22 Jan 2008 (CST)
I'm not against manga being in the ISFDB. But that's what the current policy is and when I saw it going through the author list I figured I'd clean it up. Not a big manga or anime fan but I have enjoyed what I've seen of her work. Dana Carson 20:02, 25 Jan 2008 (CST)

Full artist signature over last name editorial attribution

One of my recent posts included a number of topics. One of them concerned situations where an artist is editorially credited only by last name but signs a fuller version of his name in a legible manner. Virgil Finlay for instance almost always signs his full name but is often credited as Finlay - same with Freas/Kelly Freas. It makes sense to me to use the more fully qualified signature. Perhaps we can come to some kind of agreement on this one issue.--swfritter 22:08, 22 Jan 2008 (CST)

Entering the full name is fine with me as a citation can be added for the reliable "secondary" source. If I was entering the publication I would use Virgil Finlay and add the publication note "The copyright page credits 'Finlay' as the cover artist but the artist' full name, 'Virgil Finlay', is visible on the front cover and so that was used for this ISFDB record."
In other words, I generally only credit exactly what's stated as the artist credit. If I enter something else for the credit then I'll explain the source.
One exception is if an ISFDB record already credits someone and I'm unable to verify that (either in the pub or via signature) then I'll leave the credit in place but add a note "The cover artist is not credited nor is a signature visible. However, prior to verification this ISFDB record credited /artist-name/. The source of this credit is unknown but is assumed to be accurate." This goes on the assumption that the data already in ISFDB tends to be accurate. If I have a reason to feel the artist credit is invalid then I will remove it but add a note explaining that prior to verification that the ISFDB record used to credit /artist-name/ but it was removed because of /insert-reason/. Marc Kupper (talk) 02:54, 23 Jan 2008 (CST)
What about cases where the credit is for a last name only (say, "Summers") but initials (say, "LRS") are signed? -- Dave (davecat) 10:05, 23 Jan 2008 (CST)
I would go with Summers. In either case you would have to do the variant author thing. As far as the data being accurate. Much of the original data was entered from secondary sources and those secondary sources had a tendency to list legal names rather than the actual credited names. That is probably why so many covers are credited to Ed Emshwiller despite the fact that he was rarely credited by that name. Same with Kelly Freas/Frank Kelly Freas. The interior art, although entered somewhat inconsistently while we are developing standards, is much more accurate. I have been going through some of the coverart records to correct artist attribution errors and also to check that the magazine titles are correct. I might note when you update the cover artist's name in a publication the artist's name for the coverart record is updated but if you update the title of a publication the title of the corresponding coverart record is not updated. Not as bad as Editor records. If you update the name of an editor in a publication nothing else anywhere else is updated. More projects!--swfritter 14:53, 23 Jan 2008 (CST)
I handle it like this
  • I enter the Cover Artist as "Summers"
  • Add a publication note that explains the credit is "Summers" on the copyright page, back cover, etc (I always say where I found it), that it's signed "LRS" on the front cover, and that this is probably Leo Summers (the canonical name in ISFDB). The only time I don't enter a publication note is if the artist is credited on either the copyright page or back cover with one of their well known names but maybe I should start getting into the habit of always entering a note so that someone later looking at the record will know exactly what was stated and won't wonder if I perhaps translated Kelly Freas into Frank Kelly Freas (I'd never do that but as Swfritter just noted, many people do translate an abbreviation or short name into a fuller or more well known version of the artist name without also noting citations).
  • Once approved I then make the cover-art title record for Summers a variant title of Leo Summers (the canonical name). This part is optional and there are many artist records that have never been linked up to the canonical name for that artist. I also do this if I'm fairly certain the pseudonym is in fact the same person as the canonical name.
Considering Summers I see
Marc Kupper (talk) 15:58, 23 Jan 2008 (CST)
<sigh> I'll have to start adding a lot more pub notes, I guess; I've just been using the credited name. (For Analog/Astounding it's (in my experience always) either an "illustrated by" credit at the beginning of the story, or a credit (in italic type) with the first illustration in the story. Since swfritter basically told me to not assume that canonical names are established for illustrators (except (he grumbles, I can almost hear his teeth gnashing) for Kelly Freas), I've been doing aliases only for Freas (to Frank Kelly Freas).
Somewhere along the line I started quietly changing Analog cover art credits (mostly ones that say "Frank Kelly Freas") to what's actually in the magazine; but I haven't been thorough or systematic about it. At least one more pass through on that, I guess.
-- Dave (davecat) 16:38, 23 Jan 2008 (CST)
Good. I think we are pretty much on the same wavelength. Even if the canonical name is not a perfect choice people are still going to be able to find the data. Artist data is important but it is probably more critical that we get the author data correct. Frank Kelly Freas is fine as canonical with me - at least it was a name he was actually credited with. I think I have been remiss in double-checking the cover artist names - I am going the route of checking the actual coverart records so I can concentrate on one thing.--swfritter 16:44, 23 Jan 2008 (CST)
There are probably more interior art permutations that we could realistically document in the Help pages. For example, suppose Finlay is explicitly credited as "Finlay" at the beginning of a story, but a subsequent illustration, although not formally credited, is signed "Virgil Finlay". Should we use "Finlay" or "Virgil Finlay" in this case? My rule of thumb is that whenever I run into something unusual, I just use the name that seems to make the most sense, set up a VT and then document the details in the Notes field. Ahasuerus 23:26, 23 Jan 2008 (CST)
As far as Cover Art and Editor records go, it's a known minefield. I have also noticed that some folks use the "Make This Title a Variant Title or Pseudonymous Work" option for existing Variant Titles apparently without realizing that it will create a new parent title, but will not delete the old parent title, which will then become an orphan. Another script that I need to write, I suppose... Which reminds me that I really need to resume writing scripts now that I seem to be running on all cylinders again, but the Wiki has been so active lately that I can barely keep up :( Ahasuerus 23:26, 23 Jan 2008 (CST)

Unpaginated pages in page count

When a book has 6 unpaginated pages of "Author Notes" and "Acknowledgements" prior to page 1, ad then has 317 paginated pages, should the page count be entered as "317" or "6+317". I know what would be done with lc roman page numbers, but what if there are none at all? -DES Talk 15:38, 25 Jan 2008 (CST)

You just reminded me that I have not been paying attention to this. Even with the roman and then decimal publications I have not been entering x+100. Fortunately, I have list of which books I've verified...
Your question is a good one and your suggestion of 6+317 seems reasonable. Some downsides are figuring out what exactly is a "page" and now we'd be spending time counting them. With hardcover books there's a sheet of paper glued to the cover that forms a page. I forget what the technical term is for this sheet but don't think we want to count it. What if a publication has an accordion folded thing? Is that one "page" or do you count each fold? Sometimes a paperback will have a glossy page just inside the cover with a painting. Does that count?
Thus, I'd say the unnumbered pages are too much of a minefield and would get us even further off ISFDB's mission of indexing specfict. Marc Kupper (talk) 03:17, 26 Jan 2008 (CST)
One of the major benefit of entering the page count is to identify editions. Entering the roman page numbers may help in that matter. I don't believe I have seen any other bibliographic data elsewhere that attempts to document the unpaginated pages.--swfritter 13:12, 27 Jan 2008 (CST)
When there is content on unpaginated pages which I will add to the pub's record, I use the same method as OCLC. Place the number of unpaginated pages in brackets, followed by the paginated pages, e.g. [7]+245 (or 245+[7] if the material follows the paginated pages.) If the content is not going to be individually credited (such as "Acknowledgements", which I never enter) I don't include the extra pages in the page count. IMHO. Also, many artbooks (for example) are unpaginated, but OCLC gives a page count which is usually bracketed. Mhhutchins 13:26, 27 Jan 2008 (CST)
The OCLC system seems reasonable. The common pagination patterns seem to be
  1. Pages seem to be counted starting at 1 at the very beginning of the publication but are unnumbered until the start of story meaning the story often starts on page 7 or 9.
  2. Pages seem to be counted starting at i at the very beginning of the publication but are roman-numbered, or unnumbered, until the start of story meaning the story often starts on page 7 or 9.
  3. The first few pages are unnumbered and the story starts on page 1. This is what triggered this thread and you are suggesting [#]### which is the OCLC format.
  4. The first few pages are roman-numbered, or unnumbered, and the story starts on page 1. We have been using x+### for this.
We we need special notations for #1 and 2 above though or is noting ### ok? Marc Kupper (talk) 13:47, 27 Jan 2008 (CST)
When I rechecked my Bored of the Rings I found it was an example of Roman-numbered continuing through the Arabic-numbered, thus story started on page 23. When there's THAT much pre-main-content material it might be useful to note the start page of the main work even if none of the other contents are worth recording: a distortion of 7-9 pages might be OK, 22 pages might make people start thinking they've got an expanded or abridged edition. BLongley 14:00, 27 Jan 2008 (CST)
And I get teased because of the level of detail used when entering magazines. As long as the second part of the page count is the last numbered page that's fine by me - otherwise we've got a lot of work to do.--swfritter 14:21, 27 Jan 2008 (CST)
If I had my way you would - those magazines that continue a story on a completely different set of pages you don't record are misleading to anyone trying to establish a page count for the story. You don't get that problem in most books! :-/ BLongley 14:29, 27 Jan 2008 (CST)
Guess we are going to have to start including ads and font sizes, too.--swfritter 14:55, 27 Jan 2008 (CST)
Don't you dare! Our guesstimates are close enough for me, mostly, but in cases like THIS, where a story is considered to be under 40 pages by one of the most experienced people here, and turns out to be pages 8-45 PLUS pages 102-167, I have to wonder whether we're recording enough about magazine entries of the Speculative Fiction we're supposedly all interested in. I'm used to messy continuations of some columns in British magazines, but a typical Interzone or such just has a couple of pages of "continued from page [x]" for a lot of paragraphs that didn't quite fit. Losing length details of two-thirds of the story seems a bigger issue to me than me not recording an uncredited artwork somewhere. I know I can't win any sort of argument: I've recorded a piece of art that's just fancy text really, in a BOOK, and ADMITTED to it here: I blame Marc Kupper for that though. ;-) BLongley 15:47, 27 Jan 2008 (CST)
Unfortunately, chopping up a story that way was a common practice in the pulp era, which, as Bill points out, makes it difficult to tell whether it was an ss, nt, nv or even a "(complete novel)". However, keep in mind that there are other gotchas when trying to derive this information from page numbers alone. For example, anything printed closer to the end of the magazine is automatically suspect because it likely shared pages with ads, so what may look as a "nt" at fist glance is likely to be an "ss". Also, some editors had a habit of putting half page essays in the middle of novelettes, so even though they appeared contiguously, you couldn't tell whether it was an "ss" or an "nt" from the table of contents. Ahasuerus 16:59, 27 Jan 2008 (CST)
I suppose the bottom line is that deriving story length information from page numbers sight unseen is inherently risky. Perhaps adding a sentence to the Help pages to encourage our editors to document potentially confusing split stories in the Notes field would be the easiest-to-implement compromise? Ahasuerus 16:59, 27 Jan 2008 (CST)
If the total page count becomes important, then we'll probably need a separate field for data entry. Amazing Stories split not only novels, but novelettes, and sometimes a story could have as many as three, full-page illustrations. Not trivial, when a bedsheet could contain >900 words/page, and pulps could run >600 words/page. It probably would hurt to have a field for typical word count for each publication.--Rkihara 17:40, 27 Jan 2008 (CST)
Or we could trust the judgment of the editors, those who enter the original data and those who double check the data and those who find discrepancies in other ways. Wish all the mags could be like Fanastic Universe. No artwork, no continued stories, the same font size used throughout the run of the magazine. If it's 16 pages it's a short story, if it's 18 pages it's a novelette, if it's 17 pages you toss a coin.--swfritter 17:59, 27 Jan 2008 (CST)
Anyway, back to annoying Book Editors ;-) :Another example: here I included the unnumbered pages as there is a Novel "Starfleet: Year One" spread across several publications, chapter by chapter, that I want to find. (When I've found them all, then "Serials in Novels" is a discussion to have.) But it's always on un-numbered pages. Here I just noted that there were Short Story Competition Rules (presumably removed from later publications once the competition had closed) and an excerpt from the next book in the series. I'm a bit inconsistent on "excerpts": when it's just a few pages from the obvious next title, I don't record it. When it's from a surprising work, I probably do - e.g. an extract from a non-genre work by the same authors. When it's a complete short-story as an example of a forthcoming collection, I do. One title I entered this weekend has an "excerpt" I didn't enter, plus a chapter of the Serial that I did. I just can't recall WHICH though. (I picked up a LARGE box of free Star Trek books yesterday and have been processing in a fairly random manner). I can't recall an example of TWO sets of record-worthy details on un-numbered pages, but I'm sure we have some somewhere. BLongley 14:29, 27 Jan 2008 (CST)
I believe something that would help is if we knew why ISFDB asks for “Pages” and why it has a section about the roman numeral pages. It seems it’s not to get the number of physical pages otherwise we’d be counting the advertising, etc. after the last numbered page. FWIW - Publishers often upload the number of physical pages to Amazon. Another alternative is that ISFDB wants to know how many pages long the story itself is. I usually enter the starting page number for the story though for novels this gets concealed in the publication display as novels don’t show the title record for novels. The bit about the roman number pages is a curve ball. Template:PublicationFields:Pages says “…where viii is the highest numbered page with a Roman numeral page number. Pages without numbers that fall between the two types of page numbering can be ignored.” Meaning that if there are unnumbered roman pages between the last highest roman numbered that we would not note in ISFDB, and the start of the story, which may be a page numbered 1 or a page numbered 23… Thus I’m confused as to what we are supposed to enter into ISFDB and why. Knowing “why” seems more important as then we will be able to deal with the weird numbering cases more consistently. Marc Kupper (talk) 16:15, 27 Jan 2008 (CST)
WHAT we enter seems mostly up to us: I typically enter less than you about a book, but if you found it important and I'm cloning it I'll continue it. Same with magazine entries - if it's there, I won't delete it. All I'm really interested in is where we can find the SF. Which does seem to mean recording reviews (so we can find the SF being reviewed if we don't have it already), small bits of fiction added to some publications but not necessarily others, etc. I'm actually quite keen to support some artists too, but can't say they're essential to the ISFDB as I first saw it. (If we deleted ALL artists, how much damage would we do? Some ambiguity between editions for the book editors, wailings of anguish from the magazine editors, I suspect.) BLongley 17:08, 27 Jan 2008 (CST)
I agree - WHAT we enter seems mostly up to us. But, If we all are operating under different rules/assumptions when we do enter things then it'll turn into a mess. Thus I'm trying to understand how others are interpreting the Pages field. Marc Kupper (talk) 23:34, 27 Jan 2008 (CST)

Shortfiction vs Essay?

This item ("Professional Dilemma") isn't the only of these I've encountered, but it may be the first one I've dealt with on ISFB. It's listed in the table of contents as a "Special Feature". In this case, the pub was verified (by Alibrarian, who I'm told is MIA), & he entered (or at least verified) it as an Essay. In form it's fiction, centered around an office of patent attorneys. There are others, at least this one ("The Professional Touch"), also entered as an essay; I think I vaguely remember some more. Arguably these are essays on the complexities of patent law, in fictionalized form; but I'd just call them shortfiction if it were up to me. (And I'm not good at classifying lengths, so (again if it were up to me) I'd just leave the length as sf until someone more knowledgeable weighed in.)

At any rate, my question is what to do about these. I'm guessing that all of these are also in the same series: Improbable Profession, The Professional Look, The Magnificent Profession, The Curious Profession, The Lagging Profession, & The Professional Approach, all entered as Essay. (The last is the only other one in an issue I actually own.) Essay or fiction? Thanks for your thoughts.

(I realize that I'm guessing - on the basis of title & their being Essays - that they're all the same series. I'd not be inclined to change any particular one without some verification of that from someone; I can only be sure of two or maybe three.) -- Dave (davecat) 15:34, 3 Feb 2008 (CST)

I'd LOVE a "Fictional Essay" category! E.g. here you can't distinguish between the real essays about the creation of the publication from the essays about the fictional world described. Of course, I'd also like "Fictional Recipe" and "Real Recipe" so I can deal with books like this and this. I don't want to be held responsible for someone trying a "Bloody Stupid Johnson" recipe on my say-so though. ;-/ BLongley 16:12, 3 Feb 2008 (CST)
Um. I think what you're talking about is different. (I'm not familiar with the pub you cite.) These aren't essays about something fictional; they're fiction apparently included for didactic purpose. Which Campbell was quite capable of doing anyway, of course; AFAICS the only reason these aren't just called short stories (or novelettes if long enough) is that they're listed in the table of contents as "Special Feature" not "Short Story" or whatever. I could be wrong, of course. Dave (davecat) 16:26, 3 Feb 2008 (CST)
Try finding the actual stories in this issue of Amazing Stories.--swfritter 16:52, 3 Feb 2008 (CST)
I tried, but kept getting drawn back to the cover-art. :-( Can we find some female editors that won't get as distracted as I do? BLongley 17:19, 3 Feb 2008 (CST)
  • Consider Giant Meteor Impact. This opens with a short piece of fiction, hardly more than a sketch. For the rest of its length, it is a standard essay about meteor impacts and craters.
  • Consider The Pocket Song. This is a fictionalized "Introduction" to the verse/song lyric of the same title, which follows immediately in _Takeoff Too!_. The "intro" purports to explain how the song arose within the "Darkover" world, and how it was translated into English. But it has an in-universe PoV -- i.e. it is written as if the Darkover universe is real.
  • Consider Uncleftish Beholding, a straightforward science fact essay -- from an alternate universe where English never absorbed any significant amount of vocabulary derived from Latin or the romance languages. (The title "translates" roughly as "Atomic Science")
  • Consider the "Interludes" in the original book version of Hammer's Slammers, such as "The Church of the Lord's Universe". They are basically in-universe PoV info-dumps, containing at least some fictional content.
  • For the matter of that, consider the appendices to The Lord of the Rings.
We may well want to consider indicating such "essays" as a separate type, but it would require judgment calls in each case, made by someone who had actually read the work in question -- it is not something that can be reliably determined from a ToC, in my view. -DES Talk 16:54, 3 Feb 2008 (CST)
We have discussed adding more detailed Title Types like "fictional essay" on a number of occasions. The problem with that appears to be twofold. First, it's hard to tell where we will (or should) stop since there are always further sub-categories that could be (arguably) usefully added. Second, it's much easier to add additional distinct types to a reference like the Locus Index, which is displayed as a hyper-linked Web page. If we were to add additional types to the application that displays the ISFDB data, it would require a significant amount of code juggling and even design changes whenever a new type was added. Do we display "fictional essays" together with "essays" or with "short fiction"? Or do we create a new section of the Summary page for them?
These are not insurmountable issues, but given how precious Al's time is, I am afraid that starting down that path would very quickly bog him down. Ahasuerus 22:15, 3 Feb 2008 (CST)
Fair enough. Perhaps we should agree on a more or less standardized notes entry, or perhaps a tag. That could be captured by a script later, when and if such logic is eventually added, and be useful in the meantime. At the least, we should encourge some indication in the notes field, for future reference. -DES Talk 22:57, 3 Feb 2008 (CST)
We could perhaps, also come up with some agreed upon standards as to when to use the "essay" type and when to use the "shortfiction" type in such borderline cases, as guidelines for future entry. -DES Talk 22:59, 3 Feb 2008 (CST)
There are lots of 'fictional essays' (reviews and introductions) in Lem's bibliography and I enter them as SHORTFICTION (well, they are fictional and they are short). But I created a series for them and they were all published in omnibus-like collections (so it is clear which texts belong to the series). There is also some grey area here, like an introduction to a not-yet-written book (it is more an outline of the book, as it nowhere assumes the book already exists) which I would be happy to enter as an ESSAY but mixing essays with fiction in a series doesn't work well. --Roglo 03:15, 4 Feb 2008 (CST)

(unindent) If I'm understanding, most of these (I think all except the Amazing Stories issue swfritter introduced) are about the opposite of my original question. That is, they're about fictional items in the form of essays. That said, I agree that some standardized way of dealing with those would be good. (And I'd agitate for a shortfiction of appropriate length, with a note describing it. Not quite sure what to do about any which are too long for shortfiction, & there must be some, though.) Hmm. And I see that I left this one: "The Present State of Igneous Research" as an essay. Hmm.
I think I'm just going to quietly decide that the ones I originally asked about are shortfiction, misclassified on the basis of being listed in TOC as "Special Feature". And I suspect Campbell did that because they're not, exactly, SF; though he didn't always let that stop him from including non-SF in with the short fiction categories, mind you. But I think these are just pure & simple short stories—plot, character, dialogue as you'd expect—included in Analog because Campbell liked them & they dealt with arcana of a type he found interesting & thought his readers would find interesting. But essays of whatever type they're not.
I'll make a series to include them. I wish I had more of the issues containing these items. Dave (davecat) 11:08, 4 Feb 2008 (CST)

I ran into this problem about a year ago when I was entering some of my 1930s pulps. A couple of popular science essays were thinly fictionalized and after some head-scratching we decided to enter them as short fiction. It can be a really fine line, though, since the amount of fictionalization can vary a great deal. Ahasuerus 15:14, 4 Feb 2008 (CST)
In these cases that I asked about, the story is probably a much greater factor than in those you cite. (The words "thinly fictionalized" wouldn't occur to me.) I've met cases where I'd have left the Essay type in without even thinking to question it, myself, though. Dave (davecat) 08:32, 5 Feb 2008 (CST)
I skimmed through the issues of Astounding/Analog containing the Lockhard stories that you listed above. I believe they should be classified as fiction. The story in the Sept. '62 issue was in fact listed as a short story. I believe if we generate a "fictional essay" category we'll be headed down the slippery slope, as many well known novels would easily fall into that category. Moby Dick and the fifty or so pages on Cetology in the center of the story comes to mind. Mack Reynolds and those interminable political lectures he inserted into almost every novel that he wrote, not to mention Ayn Rand, and Heinlein. Then there are the world builder novels, written by Hal Clement, Poul Anderson and others, containing long lectures on planetary science.--Rkihara 15:29, 4 Feb 2008 (CST)
I agree 100%, as far as this kind of thing goes. (Whether we should have a type for things that are in format essays but are in fact entirely fictitious (which other people have kept bringing up here) seems to me a separate question; & I think that idea has a lot more merit.)

Rkihara, you said you skimmed through those issues. All of them without exception? Can you verify that all of them were this same series, as I understand you to say? I only have issues with (I think) three of them, so I was speculating that at least those with "Profession" or "Professional" in their titles (which I listed) were connected. Thanks (either way)! -- Dave (davecat) 08:32, 5 Feb 2008 (CST)
You have a good point, creating a separate "type" for such works would a) be a good deal of coding work when that is already a bottleneck, and b) lead to significant cans of worms that are perhaps best not opened. However, i would like soem agreement on dealing with the pure, plotless, in-universe PoV infodup, presented under a separate title. This is a work that explains in some way what is going on within a fictional setting, written as if the fictional setting were real. it is most often included in a collection or anthology,or as an afterword or appendix to a novel or long work of shortfiction. I propose as a guideline: in future, is that such works be classed as "shortfiction" (in part because that will cause them to be listed as part of any fictional series to which they belong), and that editors are encouraged to use a tag of "Fictional essay" for such works. What do you think? -DES Talk 15:39, 4 Feb 2008 (CST)
Davecat, I "skimmed" all of the Astounding/Analog issues you didn't have, I didn't check the one issue of Fantastic Universe, as I only have a few copies and they're buried away. I only read enough to determine that I would have classified them as fiction, but didn't pay any attention to plot, setting, or characters. They're pretty short, so I suppose I could sit down for an hour and read them all. If you have more that you would like checked, I'll read them more carefully and get back to you.
DES, I think that the dividing line between "hard" science fiction and a "fictional essay" is pretty vague. How would you differentiate between the two? Consider "Ralph 124C41+" or other utopian novels, "The Cold Equations," or other stories revolving around a scientific or mathematical solution.--Rkihara 12:29, 5 Feb 2008 (CST)
Yes, the dividing line in such matters can be tricky. But there is a specific and IMO recognizable sub-type that i think is worth noting. I am referring to works that are:
  1. Shorter than book length;
  2. Contain no or effectively no plot themselves;
  3. part of a larger series or overall work;
  4. Written from an in-universe point-of-view -- that is written as if the events in the work of fiction were true;
  5. Serve primarily to explain the background of the series or larger work to which they are connected.
Had Asimov given us entire articles from the Encyclopedia Galactica instead of brief excerpts, they would fit the criteria I am listing. I mentioned above the "interludes" in Drake's Hammer's Slammers and they are precisely the kind of thing that I mean. So are the Appendices to The Lord of the Rings considered as separate works. Another possible example is the "history" that serves as an introduction to The Syndic. There is an essay on the development of the Bolos in one of Laumer's Bolo books that is an example. These are all presented as if they were essays or articles, but essays written within the fictional world of a larger fictional world or series. That is the specialized sub-type that I feel is worth noting is some way, probably by way of an agreed-on tag, rather then an entire new type such as "essay" and 'Shortfiction" now are. -DES Talk 13:52, 5 Feb 2008 (CST)
Now things are getting confusing. Davecat was talking about factual essays presented in the form of fiction. If I understand things correctly, you're talking about fictional essays about fictional or speculative subjects. I have to think about this. I'm inclined to think that marking such essays in a larger work is not a good idea, it would add unnecessary complexity. Would the "Prologue to The Stars My Destination" explaining the discovery of Jaunting and it's effects on society be such an essay? Maybe this discussion should be split under the two types to reduce confusion?--Rkihara 14:30, 5 Feb 2008 (CST)
Maybe it should be split, as the two types, while not unrelated are in a sense almost inverses. As to whether to record the sort of things I am speaking of, when the appear as separate entries in a ToC, particularly in a collection (as they do in the Drake work) or when they have separate titles and are long enough that they seem to merit separate entries, and particularly when they might be published separately, or when some editions of a work include them and others don't. In short, i am speaking of things we are already recording as separate works, but with no indication of their special nature. Frankly i don't consider the sort of things Davecat was discussion as anything but fiction that is rather dry, and has a didactic or educational purpose, and that could include a great deal of SF indeed. But Bill's first response in this thread seemed to point to the sort of think I am speaking of: works that are essays or articles in form, contain effectively no fictional plot, but are written as from within a fictional setting, and are separate enough from the fictional work or works to which they relate that a separate listing is justified. -DES Talk 15:07, 5 Feb 2008 (CST)

(unindent)For those unfamiliar with Bill Contento's classification, he has a http://contento.best.vwh.net/abbrev.htm variety of codes] to account for various shades of gray. In this case the most relevant ones appear to be:

  • lk linking material
  • fa facetious article
  • vi vignette (under 4 pages, under 1,000 words) [sometimes used to tag introductions that assume that the described fictional universe is real]

Ahasuerus 23:27, 5 Feb 2008 (CST)

Tracking relationship of magazine cover art to interior story

The first part of this (the 2007 entries) was posted in the Community Portal and then (fairly soon) archived here:

Reposting a question from a new editor for general discussion:

Hi! I'm new. Question for you - are we tracking whether or not publications were illustrated in the original, and especially which stories inside a magazine have associated cover and interior illustrations, by what artists?

I'm particularly interested in the question of tracking printings, reprintings, and online availability of short fiction and am in conversation with some folks about automatically saving TOC - related info for collections that are being digitally scanned. Would love to coordinate efforts.

Netmouse 13:10, 10 Dec 2007 (CST)

(My answer is on Netmouse's Talk page). Ahasuerus 16:10, 10 Dec 2007 (CST)

We do not have a specific field designed to document the story illustrated by the cover of a magazine. As a result the only place to document this information is the magazine notes field which means that the inclusion of such data is not required of the ISFDb editor and not accessible for listings. I always try to include such data.--swfritter 20:20, 10 Dec 2007 (CST)
As I've been entering all these Analogs I've wished many times for this; all (or very, very nearly all) of those covers are intended to illustrate a specific story or article, & it would be good to put this in. I have not, however, been putting this in the notes. (For Analog, at least in the years I've got, in almost all cases, the first story or article after the editorial is the cover feature. And it wouldn't take very much to determine this just from the pub listings, which include the online cover images. Not much need to go back to the original magazines.)
It would really be nice to have a cover-title field in the pub record; I think that would be better than a separate content item, but I could be wrong. Dave (davecat) 11:07, 11 Dec 2007 (CST)
Well, the Coverart Title record is editable after entry, you don't have to leave it as "Cover: Pub Title". In which case making this visible on certain other screens could be the feature request, rather than adding a new field. I've noticed this isn't tightly coupled, as if you correct a title the Coverart record doesn't change. There's a 'title_ttype' of 'BACKCOVERART' allowed for too, which doesn't appear to be used. I wonder what Al intended for that? BLongley 13:21, 11 Dec 2007 (CST)
Aha. I think I'd once observed that there is a coverart record for magazines, but forgotten it. But the format does seem quite standardized. What would be a reasonable replacement? Maybe:
 Cover: Satan's World (Part 1 of 4) (Analog, May 1968)
or something like that? Dave (davecat) 05:46, 12 Dec 2007 (CST)

(unindent & end of first reposting) When this issue later came up on my talk page, Mhhutchins said this:

I honestly feel we shouldn't be changing the title record that's been assigned by the system. Updating the dates, of course, but that's not what we're talking about here. We're talking about changing the title of the title record. If you feel the system's way of assigning a title record is wrong, that's a matter for discussion. And Al von Ruff would have the final say in what changes would be made. I can't imagine the time and effort it would take to go back and change all the cover art titles to match the piece of fiction that suggested them. But that's up to the people who will be making the changes. Perhaps taking this discussion over to the Rules and Standards page would bring more points of view into the mix. Mhhutchins 11:33, 6 Feb 2008 (CST)

To which I replied:

Whereas my own feeling is that, when a magazine's cover art is titled (on the cover!) & intended to illustrate a particular story, the cover art's title should indicate that. And no automatically-generated title will do so.
I do agree that the title should also indicate which issue of which magazine it is a cover for; but we have a canonical form for doing that. Dave (davecat) 13:40, 6 Feb 2008 (CST)
I don't believe that art on magazine covers is titled. Often it illustrates a story that is highlighted on the cover. But one can't assume that the art automatically illustrates the most prominently mentioned story. I've been updating the early 70s Worlds of If. Take a look at this cover. No that doesn't illustrate Simak's story. And this cover illustrates that issue's serial by Colin Kapp, not Pohl and Williamson's "Doomship". What happens when a work of art is used multiple times? We have to give it the name of the pub on which it appears. Mhhutchins 18:48, 6 Feb 2008 (CST)
I have seen on a number of magazine issues, a TOC cover art credit that specifically attributes the art to a particular story, as well as indicating the cover artist. In other cases it is fairly clear what story the cover art illustrates. In yet other cases it probably does not illustrate any story in particular. -DES Talk 19:00, 6 Feb 2008 (CST)
Michael, I never said, & would never have said, that all magazine cover art is titled on the cover is titled, or that it all illustrates specific stories. There are plenty of times when it isn't & doesn't. My comment said "when a magazine's cover art is titled (on the cover!) & intended to illustrate a particular story . . . ".
As it happens, that's almost always true of Analog during the Campbell years, and mostly true of the issues of Astounding I've personally got (which is not very many of those). Sometimes the illustrated story was not the lead story, but the title on the cover went with the story illustrated. (And, yes, in these issues one story or article was listed on the cover. All this changed after Campbell died.) In at least one issue of Astounding January 1959, AFAICS the cover relates not at all to any contents of the magazine, but the title on the cover goes with the cover art not with any contents, so I'd say that in this case it's absolutely an artwork title. -- Dave (davecat) 20:00, 6 Feb 2008 (CST)

For the moment that's the state of the discussion as I know it. (Or rather, it was before more discussion got entered above. -- Dave (davecat) 20:00, 6 Feb 2008 (CST))
(Um. Swfritter & Mhhutchins did also bring in the fact that the Coverart record is invisible from the pub listing. I'm not trying to misrepresent the discussion, but I think that's a side issue except insofar as it means that editors are likely not to think to update a coverart record.) I think it would be good to get a wider range of opinions on this. -- Dave (davecat) 13:57, 6 Feb 2008 (CST)

The only changes I would make to the Coverart records would be to make the titles match the pub titles and to enter the correct date. My preference would be to remove the capacity to edit Coverart records and programmatically insure that they are updated whenever the pub record is updated.--swfritter 18:02, 6 Feb 2008 (CST)
I'm in agreement with swfritter here. If we don't have access to the coverart record in the edit mode, then the least the system can do is update the record's date when the pub's date is changed. And the same thing goes for the inaccessible editor records for mags. Mhhutchins 18:39, 6 Feb 2008 (CST)
The problem with Editor records is that most of the magazine Editor records have been merged - something that makes no sense for Coverart records.--swfritter 18:43, 6 Feb 2008 (CST)
I think it does make sense to merge Coverart records at times: e.g. here some verified covers with identical art (but different blurb) have been merged. It keeps the clutter down, which with frequently reprinted book titles becomes quite useful. Obviously not with Magazines, except maybe ones like Destinies. (Which I see has given us Award problems... but that's another topic.) However, I think either the Coverart Title has to be made visible or it should be dropped, this halfway house is helping nobody. (No point automatically updating them if they're just going to remain as "Cover: <Pub_Title>" or even "Cover: <Title_Title>", generate them on the fly.) A workaround might be to enter Coverart titles as Interiorart entries with a page of 'fc' if needed.
Editor records are another pain entirely. BLongley 13:26, 7 Feb 2008 (CST)
After having added and merged and seriesed over 800 missing Editor records, I agree about that. The fc was also one of my thoughts but another possibility is making a variant title with the story title linked to the Coverart title. Having two records for one cover might get a bit confusing but the linkage between the Coverart title and a variant title might be less so. If we are going to merge artwork we need a policy that requires either physical access to the artwork to be me merged or very reliable images.--swfritter 17:55, 7 Feb 2008 (CST)
I'll leave Editor records to the experts: merging is a clearly good way of shortening the Editor's Author page, but that strikes me as a short-sighted solution causing problems now. Trying a search for a magazine edition from the front page is useless, and I'd like that fixed. Having to know the editor before you can find the magazine is just plain WRONG. BLongley 18:14, 7 Feb 2008 (CST)
With an Editor Series it seems to me to be relatively easy.--swfritter 19:45, 7 Feb 2008 (CST)
Better, but not ideal. A new user should just be able to put the name in the basic title search and find a Magazine record straight away, same as they can enter a Novel and find the Novel straight away (well, so long as the Coverart and Reviews don't swamp the results so much it's not on the first page). The "Editor" records confuse and I only found my first magazine by tracking it down from Coverart, then couldn't do that for another because it had a photograph for a cover! We can definitely improve things further. BLongley 13:18, 8 Feb 2008 (CST)
Coverart entries are a pain we can fix more easily I think - the exact cover IMAGE is recorded at pub level, but when we have some reliable images there we can merge the coverart entries for same artwork. But we need to stop trusting/using Amazon ZZZZZZZ entries first. BLongley 18:14, 7 Feb 2008 (CST)

(unindent) Another branch of this discussion is now on the Community Portal talk page. -- Dave (davecat) 10:59, 14 Feb 2008 (CST)

Discussion moved Community Portal has been moved here.
(Thank you!) -- Dave (davecat) 19:40, 14 Feb 2008 (CST)

Linking Cover illustrations to the stories they illustrate

This sample magazine has a possible solution. I have entered an interior illustration on page 'fc' and made it a variant of the Coverart record. This is the way it looks on the artist's page. The primary downside is that the the Coverart record shows the pub twice which might not be a big deal because it is not likely to be accessed that often. Technical issues? Display issues?--swfritter 19:04, 13 Feb 2008 (CST)

This is a possible solution to what? Marc Kupper (talk) 23:55, 13 Feb 2008 (CST)
See this discussion. We have already discouraged one potential editor because we don't have this capability.--swfritter 09:35, 14 Feb 2008 (CST)
I tried to move that discussion off my talk page to Rules and standards discussions. Unfortunately activity continued, somewhat separately, both places.
swfritter, a problem I see is that it still says "interiorart". But your idea would go some distance toward meeting my own reasons for wanting the coverart record to have a real title.
(But not all. I find it, um, really strange to see dozens and dozens of records that say only that Analog had a cover each month, when those covers are in fact (almost, if not absolutely, without exception) story illustrations; though I guess I tend to overlook the fact that they also record the artist. And granting that other magazines apparently normally have random covers not related to contents.) -- Dave (davecat) 10:52, 14 Feb 2008 (CST)
I tend to agree with Dave (davecat) here. When an illustration in fact illustrates a specific work of fiction, that fact should be tracked here and it would seem msot natural to record that as the "title" of the illustration. Obviously, that it was the cover art for a given magazine is also important and should be tracked. Moreover, when such art is displayed (for example in SF convention galleries) or printed seperately (for example in "collected art of..." books) it often has a title given by the artist separate from the work it illustrates (made up example: "Moonrise -- Cover for The Moon Wizard"). When such titles are known (as often they are not, particularly for older art) shouldn't they be recorded somehow, also? -DES Talk 11:25, 14 Feb 2008 (CST)
I might also note that back cover and end page art are defined as interiorart. Once again, short of software modifications this is the only method I could think of that would accomplish the task. If this is not an immediately acceptable solution I would suggest we drop the issue until an acceptable solution can be implemented.--swfritter 12:09, 14 Feb 2008 (CST)
It looks OK to me - I suggested "fc Interiorart" before, the variant as well doesn't matter too much to me. The question of whether you also want to use this to record "as signed" artist versus "as credited" artist or "canonical" artist remains, but that can be dealt with separately. BLongley 13:13, 14 Feb 2008 (CST)
I picked up the first part of the idea from your suggestion. I think there might be technical issues involved as the implementation of invisible Editor and Coverart records may be modified. If this methodology fits in with those changes it might be a solution and I would suggest we don't implement it without full knowledge of any technical factors.--swfritter 14:57, 14 Feb 2008 (CST)
For future technical factors that may interfere, always ASK Al. For future features we want, TELL Al! ;-). BLongley 16:08, 14 Feb 2008 (CST)

"In Memoriam"

After entering this rather depressing set of entries I found myself wondering once again if we should be recording Obituaries in a special way. They're often quite informative, and may be as useful as "bibliography" essays, but I think I'm leaning more toward a category that puts them up there on the author page along with "interviews". Please discuss. BLongley 18:14, 6 Feb 2008 (CST)

That makes sense to me, provided that the change is not too hard to implement. Come to think of it, there are occasionally explicitly biographical essays that are not obits. If we do create a special category, should these go into it as well? -DES Talk 18:25, 6 Feb 2008 (CST)
There is an outstanding feature request that I filed on 2005-07-25. It asks for "secondary bibliography" support so that we could link books and articles that contain bibliographic information about Authors -- as opposed to Reviews which cover Titles -- to their respective Author records. Perhaps we could have expand the request to cover both bio- and bibliographies, obituaries being a subspecies of biographies? Ahasuerus 20:42, 6 Feb 2008 (CST)
Anything about an author that isn't by the author. Dana Carson 21:06, 6 Feb 2008 (CST)

Current Help on foreign language books

The current Help:How to enter foreign editions page is badly out of date. I suggest we add/modify a few things to summarize our current unwritten policy as follows:

  • English is currently a/the privileged language in the ISFDB in that foreign language translations of English language books do not get their own title records; they are entered as Publications under the English language title instead
  • For works that originally appeared in a foreign language, the master title is the title of the work as written by the author(s) (as opposed to as published to account for Bulmer, Sheffield, et al). English language translations, when available, are entered as Variant Titles. Translations to other ("third party") languages do not get their own title record and are entered as Publications under the main title record instead.
  • Eliminate the section about entering non-Latin (i.e. Japanese, Cyrillic, etc) characters in titles since at this time our software's support for them is very poor and searching is not supported. (This means that we will need to hunt down and transliterate all victims of our unsuccessful attempts to use Cyrillic in early 2007, not that there are that many of them).
  • Eliminate the comment about creating pseudonyms for Anglophone authors whose work has appeared in other languages. We don't want to have Author records for "Исаак Азимов", "אסימוב אייזק", etc just because Asimov's books have been translated worldwide.

Is this a fair summary of what we are currently doing with foreign language titles? Ahasuerus 23:47, 12 Feb 2008 (CST)

Well, what I'm currently doing is avoiding the problem as much as possible... :-)
Still, when I'm just whacking about an unloved but highly-represented-here author's page, I often don't put foreign titles under the English title as I have no idea which came first and I'm never 100% sure that my translation skills have got the right titles matched anyway. In those cases I'll make my best guess a variant title so it still visibly needs work and I can see where to put other editions of it in the meantime. The suggestion for where it ends up is fine by me for the moment but I'm going to want the foreign titles hideable at some point (same as for multiple near-identical English printings - the foreign variants can just be another 'edition' when we sort that concept out).
I don't understand title of the work as written by the author(s) - can you expand on what you mean there?
Yes, please let's get rid of as many non-working characters as we can! I can just about cope with Philip José Farmer and Élisabeth Vonarburg but not much more... BLongley 13:12, 13 Feb 2008 (CST)
The "written by the author" caveat is included to cover situations when a foreign language translation of an English language book appeared first. For example, Charles Sheffield's Convergence, volume 4 in his Heritage Universe, was delayed while he was transitioning from Del Rey to Baen in the mid-1990s. In the meantime, a publisher in Belarus (?) translated the book and published it in Russian in 1995, two years before it appeared in English. This was a recurring question on rec.arts.sf.written in the 1990s and even made the FAQ. If we were to use the language of the first publication to determine the canonical title, then we would have to enter the Russian language title of Convergence and make "Convergence" a variant title, which seems counter-productive.
In addition, there is a number of books and series by English language writers which were dropped by their US/UK publishers in mid-stream but were continued/completed abroad. The two best known examples are Ken Bulmer's Dray Prescot series, dropped by DAW (when Wollheim got sick and his heirs took over) and continued in Germany, and "Ansen Dibell"'s The King of Katmorie books, with 2 of the 5 volumes only available in French and Dutch. We currently list the English language titles of Dibell's books even though they never appeared in English. We don't list any Prescot books post-Warlord of Antares at all, but that's just due to an (unfortunate) gap in our coverage.
The only current exception that I know of is Robert F. Young, whose novella "The Quest of the Holy Grille" was later expanded into a novel, La Quete de la Sainte Grille, which has appeared in French but not in English. We currently list the French title, in part to avoid confusion with the eponymous novella, but we may want to use the English language version instead to be consistent. Ahasuerus 14:29, 13 Feb 2008 (CST)
OK, I'm still confused as to the "English language translations, when available, are entered as Variant Titles" part: is this only for when a title was written in English, published under a foreign title, then printed in English under a different name? It seems a lot to ask people to know what the original unpublished English title was, and certainly no help in my efforts to just plain ORGANISE some neglected authors' titles. BLongley 15:55, 13 Feb 2008 (CST)
I took that to mean that English language titles of translations of works originally published in non-English languages (such as Lem's work) should be entered as variant titles of the original work. Is that correct? -DES Talk 16:06, 13 Feb 2008 (CST)
Looking at the examples mentioned e.g. this one, I guess that the intention is that a work written in English should get an English title even if it first appeared under a foreign name, but I don't see an example of it later being republished under an English title that needs to be a variant of the original one? BLongley 16:30, 13 Feb 2008 (CST)
Perhaps an explanation of Deathworld 4 would help? BLongley 16:32, 13 Feb 2008 (CST)
Look at Podróż dwudziesta druga of which The Twenty-Second Voyage is a varient. -DES Talk 17:15, 13 Feb 2008 (CST)
I rather suspect that Deathworld 4 is an example of how we don't want things done in future, but I may be mistaken. -DES Talk 17:16, 13 Feb 2008 (CST)

(unindent)Bill is quite right that "the intention is that a work written in English should get an English title even if it [is] first appeared under a foreign name". This rule would cover all cases of delayed publication from Sheffield to Bulmer to Dibell. I suspect that we may be running into this issue more often as we go forward: novelizations and tie-ins are usually tied to the release of the underlying intellectual property (usually a movie) in the target market, so even though a novelization may be written in English, it may appear in a foreign language earlier if the movie is released in that market first.

To address the other part of Bill's question, I haven't thought of the class of cases where a text would be "later ... republished under an English title that needs to be a variant of the original [English language] one". And it's a good question too since there is nothing preventing US/UK/Oz editors from changing the title to something completely different if and when they decide to publish the books in English. Besides, the English language titles that we would use under the proposed rules may be literal translations of the known foreign language titles and not what the author originally had in mind. I can't think of any SF examples off the top of my head although some Harry Stephen Keeler mysteries that appeared only in Spanish and Portuguese in the 1950s and were finally published in English in the 2000s may qualify. I guess in a case like that we would have to create VTs for the newly minted English titles.

Another complication to consider is that it's not always immediately clear what the original language of a story/novel/essay may have been. For example, Stanislaw Lem's German was very good, so we enter his German language essays as they appeared since we assume that he wrote them in German. However, it's conceivable that he first wrote them in Polish and then translated them into German, in which case, according to the proposed rules, the first Polish language publication should be used to create the primary Title record. I can't think of any way to safeguard against this eventuality aside from documenting our assumptions in the Notes field.

To address the final point, Deathworld 4 is probably one of the most complicated cases that we have. As we discussed early last year, there were a lot of "sequels-by-other-hands" done in the ex-USSR in the early-to-mid 1990s when many ex-Soviet countries provided no IP protection for foreign works originally published prior to 1973 and copyright violations usually went unpunished anyway. Any remotely popular universe -- from Conan to Edmond Hamilton's Starwolves to Sterling Lanier Hiero (!) -- could (and often did) get a hastily written sequel. We even have some of them in the database.

The Deathworld case, however, was apparently different. As far as I could determine by running Internet searches, the publisher (EKSMO-Press) was one of the bigger players in the Russian market and things were getting more stable in Russia in the late 1990s, so they decided to play by the rules. They hired a reasonably well known writer (who writes SF as "Ant Skalandis") and secured Harrison's permission to use his name and the Deathworld universe to write sequels as by "Harry Harrison and Ant Skalandis". The devil, unfortunately, is in the details and the details are murky. Some sources claim that Harrison approved an English language outline of (at least) the first volume while Skalandis did all of the writing. Other sources suggest that Harrison wasn't involved at all and the only condition that he had was that the sequels should not appear in English (although some have been translated elsewhere, e.g. in Poland). BTW, Skalandis wrote the first three sequels and then another writer (who writes SF as "Mikhail Akhmanov") wrote two more volumes. I suspect that Harrison's input went from "little" to "none" as the series progressed, but that's just a feeling that I get.

Given these uncertainties, should we enter Harrison as a co-author or should we change the attribution to "by Ant Skalandis [as by Ant Skalandis and Harry Harrison]"? Should we make the Russian language title the primary one or, assuming that the English language outline did exist and was reasonably extensive, should we use the English title (which appears on the copyright page of the Russian language edition) instead? I guess it's a corner case and you can't design a bibliographic application to account for all corner cases, so it's Notes time again.

BTW, it looks like Latin-2 and Cyrillic now work with Title searches, although "42" still returns all of our Cyrillic titles in addition to any legitimate matches. Non-Latin-1 Author searches still seem to be broken.

OK, enough masochism for now, I think I need some aspirin... Ahasuerus 23:22, 13 Feb 2008 (CST)

Given that the nature and level of Harrison's input is unknown I'd go by what's stated on the publications. If someone in the know (Harrison, Skalandis, or the publisher) gives an interview that provides hard data then we can document it in the notes. That said:
  • Translate the titles to English but also append something like "(translated title)" to the title and add a title-note explaining the real title and that the English version is a literal translation for ISFDB.
  • Credit the authors as stated if they have Latin alphabet names. If it's a Русский Муравей then you may want to translate that too. Save some of that aspirin for everyone though. Part of my thinking here is I do want the title to show up on Harry_Harrison even if it can be shown that he had zero input other than his name. Marc Kupper (talk) 00:16, 14 Feb 2008 (CST)
OK, here's a test title Doctor Zamenhof. POSSIBLY originally published in Esperanto, Esperanto title used for the English publication, Italian translation reprint. Marc, I suspect you want "(translated title)" added rather than just a Title note. Ahasuerus: have I guessed your intents correctly? Everybody else - HELP! I don't speak Italian so this pub needs treating with a LOT of caution. Blame me, or Blame Google translations - I daren't create variants for the remainder. BLongley 15:05, 14 Feb 2008 (CST)
To address Marc's point, I agree that ghostwriters are a separate can of worms. I suspect that for canonical title purposes it's generally safer to enter ghostwritten stories/books as collaborations between the ghostwriter and the Big Name Writer (Shatner/Goulart, del Rey/Fairman, etc) whether they are published as explicit collaborations or not, in part so that they would appear on both Authors' Summary pages. Except when the Big Name Author is deader than V. C. Andrews, that is. However, that's fodder for another discussion, one that I may start either later today or on Tuesday, after I am done with verifications and return to wandering. Unless someone else beats me to it, of course :)
Re: adding "(translated title)", one downside that immediately comes to mind is that it would break our review links. Perhaps the best bet would be to use the transliterated form of the Russian language title as our canonical title and explain the gory details in Notes the way we do it with Lukyanenko. After all, the only major difference between Skalandis and Lukyanenko is that the latter has been published in English and the former hasn't. And using Cyrillic, Hebrew, etc to enter Author names is out because it would kill our searches. Some non-English characters are OK, e.g. "é" in "Philip José Farmer", but most can't be found using our Search logic, which is why we use an Anglicized version of Stanisław Lem's name.
Re: La Lingua Fantastica, wouldn't most of it fall under Rule of Acquisition 6:
Works of speculative fiction published in a foreign language that haven't been translated into English and whose author's other works have not been translated into English. Arguments for exclusion: avoid duplicating the efforts of foreign language bibliographers in a field where we can't realistically compete with them. (True? False? Revisit if/when we have foreign language editors with extensive expertise in the field who would be willing to merge their biblios into the ISFDB?)
? Up until now we haven't done any serious work on foreign language writers unless their work has appeared in English or they have other ties to English language SF (e.g. see Skalandis above). Do we want to start now?
Finally, it sounds like we are in agreement re: bullet points 1, 2 (except the "as written" part, which we can safely omit for now without affecting 99% of foreign language Titles) and 4 of my original post. Is that right? If so, we may want to give it another day or two and then update the Help page with the bits that we agree on. Then we can continue arguing about our beloved corner cases :) Ahasuerus 23:14, 14 Feb 2008 (CST)
Re: adding "(translated title)" - Yes, it breaks review links. But we don't have a LOT of reviews of translated titles. And when we do have them, it's a toss-up as to whether the review is of the original or the English title... or both, separated with a "^". The review links are very fragile, but after working on them a lot recently to clean up some Authors that only exist due to reviews, I hope people are noticing when there IS a Title match with a Review before updating a Title record? (I hope so, especially now we CAN edit the Review more easily.) In the Long term I'd prefer Review Links to match more solidly and be updated automatically - but then again, I'd also like the option to link reviews to a Publication (or two), or the wonderfully-elusive concept of "Edition". BLongley 14:53, 15 Feb 2008 (CST)
Re: La Lingua Fantastica and Rule 6: Yes, many of the contents probably do qualify for exclusion. Only the Harry Harrison introduction and story seem to count as IN - and the Interiorart by Karel Thole might be of interest, but I don't know where in the pub it is. Would it be better to delete the entries we don't understand, even if they're SF? (We have several Playboys with just English SF entered for "first publication" reasons, so incomplete titles aren't totally unwelcome it seems - but I can't be sure an Italian editor won't turn up tomorrow and tell us we've missed a Foreign pseudonym of an author that IS here - e.g. "William Auld" doesn't sound Italian to me.) BLongley 14:53, 15 Feb 2008 (CST)
And yes, I'm fine with 1, 3 and 4, and 2 probably just needs a little clarification. BLongley 14:53, 15 Feb 2008 (CST)

(unindent)Help:How to enter foreign editions as per the discussion above. Corrections and additions are more than welcome. Help:How to parse data in library catalogs created, although Bill's comments about 17.5 cm UK books made me wonder if the UK part needs to be beefed up. Ahasuerus 00:05, 22 Feb 2008 (CST)

Probably, but I'm not sure how. I checked a few authors where I've a mixture of US and UK paperbacks and there is some truth in the suggestion that UK paperbacks are slightly larger than US ones. E.g. "Retief to the Rescue" (US Pocket edition) is only 17cm but "A Plague of Pythons" (Penguin UK) is 18cm. But "Retief of the CDT" (US Pocket edition) is 17.9cm and bigger than most other US editions. "Shield" (Magnum UK) is 18.1cm, "The Man Who Counts" (Ace, US) 17.4cm. "Robot Adept" (UK) 17.5cm, "Phaze Doubt" (US) 17cm. I don't really want to reorder all my books by year, publisher and height and do conclusive research, but my spot check suggests 18.5cm would be a nice suggested cut-off between pb and tp for British paperbacks. The variations are too wide to separate US and UK pbs though. (Well, the 16.2cm "Bow Down to Nul" (Ace, US) is small enough that I can't imagine it being confused with a British pub.) I think what I'm saying is that the 19cm guideline is likely to be wrong based on my collection - but I wouldn't change it based on my experiences alone, ask some other British pub owners with a wider range that DO happily mix tp and pb more than I do. I also can't offer much help on hardcover size - I follow very few authors at "must have first edition" status and although I can tell you that the Terry Pratchett Discworld books noticeably jumped from 22.2cm to 23.5cm, the juveniles stayed at 22.5cm, like all my Artemis Fowl hardcovers. But that's such a small sample I couldn't say that's a standard for Juvenile versus Adult hardcovers. BLongley 13:58, 22 Feb 2008 (CST)

Canonical Publisher names

The new search on publishers feature has revealed just how non-standard our entries for publisher are. For example, the db contains no less than 40 different strings containing "Ballantine", while this page lists four imprints, and ISFDB:Ballantine has not yet been created.

(There was some discussion of this previously in this thread, and some at ISFDB:Community Portal#Print series?, but it was a little mixed in with other matters, adn it seems to me this part belongs here.-DES Talk 16:33, 18 Feb 2008 (CST))

I suggest that we develop some general standards on how publisher and imprint names are to be represented, and then some specific lists of canonical names for each publisher and imprint.

  1. I suggest that each publisher have a single canonical name, or if a publisher has change the form of it's name significantly, a single canonical form for each era. We shouldn't have "TOR", "Tor", "Tor Books" and the like, one of these should be standard, nd the others perhaps recorded as variants, just as we do for different forms of an author's name.
  2. I suggest that when a publisher and imprint are listed together, the publisher should normally come first, followed by a slash, followed by the imprint name.
  3. In some cases an imprint is well enough known that it may be listed separately. (Del Rey Books comes to mind) We should identify which imprints this seems reasonable for. Pending the creation of publisher biographies, this info should probably be kept in the wiki.
  4. I suggest that, once we agree on standard forms, non-standard publication records ought to be converted. Some of this can probably be done by script. Some will probably have to be done manually, because info that was put in the publisher name may need to be copied to the notes field, or a judgment may need to be made about which canonical publisher or publisher/imprint string should be used.
  5. A separate imprint field may well be a good idea eventually, but if we attempt to follow standards conversion should be automatable when and if such a field is available.

What do others feel about these suggestions? -DES Talk 16:27, 18 Feb 2008 (CST)

I'm just about finished changing all "Del Rey" and its variants to "Ballantine Del Rey", so I have mixed feelings about changing it back to "Del Rey" or "Ballantine/Del Rey". I do agree that we need some standardized rules to follow.Kraang 16:43, 18 Feb 2008 (CST)
Well, we can always debate the proper form, but changing from one specific form to another ought to be something that a script could handle i would think. My reasons for suggesting the slash is that many publisher and imprint names include multiple words, and thus spaces, and a slash will make it very clear where the publisher name ends and the imprint starts, this will make it easier for a script to move the imprint name to a new imprint field, when and if we add such a field. It will also make it easier for automated searches to use our data, i think. -DES Talk 16:55, 18 Feb 2008 (CST)
Sounds like at a minimum we need publisher editing tools that allow the publisher name to be changed without visiting all of the publications (analogous to changing the author's name without visiting all of the titles) which would allow easy switches between "Del Rey" and "Ballantine Del Rey", and tools for merging publishers together. That should lower the amount of manual labor involved. Alvonruff 17:16, 18 Feb 2008 (CST)
That would certianly help a lot. How hard are those going to be to implement? -DES Talk 17:28, 18 Feb 2008 (CST)
Probably too easy. :-( Al, give us a tool that helps us zap Manga or RPG pubs and Titles in one go before you give anybody something as damaging as this could be! BLongley 18:31, 18 Feb 2008 (CST)
I agree that a tool that allows editors to merge publishers could be very dangerous. We have caught enough well-meaning Author merges that would have combined a canonical name with its pseudonyms ("How much difference can there be between Iain Banks and Iain M. Banks??") to make me leery even of Author merges and Publisher merges could be literally orders of magnitude worse. Nothing like destroying hundreds of hours of careful work with one click of a mouse to make us go to back to the backups or see people quit in frustration. Ahasuerus 22:55, 18 Feb 2008 (CST)
Having said that, I can also see value in organizing publisher information better so that we could use the database to answer questions about publisher history. The way things are set up now, a search on "Ballantine" returns so many hits that a casual user will be disoriented, although some subsets, e.g. "Ballantine Adult Fantasy", are obviously useful. I am tempted to propose that we change the name of the current "Publisher" field to "Imprint" and then create a new and separate table that would link these imprints to actual publishers. However, how would we handle "floating imprints" that move from one publisher to another? Would we have to revisit all of their publications and add the publisher's name to the imprint? Ahasuerus 22:55, 18 Feb 2008 (CST)
Anyway, all of this may be premature. First, we have a great deal of simple cleanup work to do using the new tool, e.g. merging Pan / Ballantine, Pan/ Ballantine and Pan/Ballantine. Hopefully, the cleanup process will help us come up with better ideas. Ahasuerus 22:55, 18 Feb 2008 (CST)
I think we can and MUST leave this for a few weeks at least, probably months. I see people are using the new tools and thinking "Doh! I really SHOULD have used THAT name instead!" and they're busy "correcting" their past "mistakes". That's fine by me, it reduces the number of imprints and publishers we'll have to merge/make variants for in the long term. I think we need the "Imprint" and "Publisher" fields established BEFORE we allow ANY automated merges though - that can be done now, just copy all "publisher" field entries to the new "imprint" field and let people fix imprints for a bit, then gather them up and we can see how the "publishers" work. BLongley 18:23, 18 Feb 2008 (CST)
I'm really AGAINST "Publisher" "Slash" "Imprint" standardisation as I feel we should be taking care of the imprints, which change hands more often than people may think. And "Publisher" is arguable - do we want the company quoted on the book, or the parent company that owns them (which might also be quoted on the book), or the Group or Division that owns them (again, this might also be quoted on the book), or the holding company that owns a publishing company or ten... what I want are details of when a Publisher got taken over, and became an imprint of another company that owned multiple publishers, and was then taken over or sold on to another company - if I came back here and discovered all of my Mayflower, Souvenir and NEL books were standardised under one current publisher I'd quit. That's the sort of vandalism of data I expect from Amazon. BLongley 18:23, 18 Feb 2008 (CST)
Yes, we probably want variant publisher/imprint names so we can record what's actually specified on the pub. Yes, we want to reduce the number of names used and have a "canonical" name. Yes, it's going to be difficult. But "NO!" - I don't want any of the changes suggested yet. 1) Define "Publisher". 2) No, create separate fields for separate data. 3). No, start from the lower levels of recorded data and work UP. That's what we have to do with printings. 4) Eventually we can clean up remaining stuff with scripts, but not because of the reasons you suggest, which IMO would have destroyed a lot of data in the meantime. 5) A separate imprint field is a MUST, not a "nice to have" eventually. BLongley 18:23, 18 Feb 2008 (CST)
I see your points, and they have significant merit.
  1. If we were never going to get an imprint field, I think "Publisher / Imprint" would be the best way to handle things. However, I trust that sooner or later we will ahve such a field.
    Preferably sooner, IMO. It'll point editors in the direction we want. (If we do all want both.) BLongley 13:43, 19 Feb 2008 (CST)
  2. I am not in favor of scripted changes which are at all likely to lose significant data. I see no reason why a script couldn't convert "Ballantine Del Rey" to "Ballantine / Del Rey" or "Ballantine Publishing Group" to "Ballantine Books" or "Beagle Books (Ballantine)" to "Ballantine / Beagle" with no loss of data (provided that we agreed on the desired names, these are just examples). What data loss would you see in such cases?
    It rather depends what people have been using the Slash or Brackets for so far. What if somebody has been using the slash to separate Publisher and Imprint already, and your merging a "Ballantine Del Rey" imprint with plain "Del Rey", part of the Ballantine Publishing Group? BLongley 13:43, 19 Feb 2008 (CST)
  3. I think entering new publications with Publisher / Imprint, (and with any available further detail in the notes) would ease conversion when the new field is available.
    A lot of people have and probably will continue to use something like this, but as we've already got such in probably every possible permutation, I don't think it'll make things easier. The script will never know if it was entered according to the new standard or some arbitrary older one. However, introducing the new field now WITH explanations of its intended use will help - the new field can't have been affected by older conventions. Well, maybe if we pre-populate it, but then the fact that we have extra separators in both fields should be a good warning. BLongley 13:43, 19 Feb 2008 (CST)
I am not proposeing wholesale changes be made before they they are throughly discussed and agreed to, and if a new field is likely to be available soon, perhaps any scripted changes to existing data should wait on that field. -DES Talk 18:44, 18 Feb 2008 (CST)
Regarding slash - while it used to bother me these days I'm used to "Publisher name / SFBC" though I'm still bothered that the publications do not state SFBC other than some Doubledays that say "Doubleday Science Fiction" on the jacket though the title page states "Doubleday."
I'm personally in favor of "imprint / publisher" as the imprint name is often more visible in the publications.
Overall, the publisher names are an area where I don't have strong feelings as there are many publications where the name is quite difficult to determine. The Timescape books immediately come to mind but Fawcett gets combined with Crest and later with someone else.
I tend to enter things as stated in the publications meaning I've used "Del Rey / Ballantine" as that's the order the names appear in on the title page. I usually shorten up a name. "DAW Books, Inc." is "DAW Books"
This is a time when I wished ISFDB had automatic notification to verifiers of changes to publications as I can see people going in and making mass changes. I had thought there were some "Del Rey" books without "Ballantine" and that's one of the things I'mn re-checking as I go through my books. Marc Kupper (talk) 22:27, 18 Feb 2008 (CST)
It depends on what you mean by "without". I'm pretty sure there are "Del Rey" imprint books that aren't "Ballantine Del Rey" imprints, but the publisher is still probably listed as Ballantine Books or Publishing Group.
It rather depends on where you're taking the data from - logo, spine, front cover, copyright page, title page - and what you mean by "imprint" and "publisher" anyway. E.g. One of my acquisitions from Sunday says "Panther Granada" on the spine, but with the two words in different fonts, "Panther" before the ISBN on the cover, "Panther Science Fiction" above the title on the cover, "Panther: Granada Publishing" on title page, "Published by Granada in Panther Books" at the top of the copyright page, "A Panther UK Original" two lines later, gives "Granada Publishing Limited"s address, and at the bottom of the Copyright page lists "Granada (R)" and "Granada Publishing (R)" as registered marks. I'm definitely leaning towards "Panther" as the canonical imprint as it's about the only word guaranteed to be in it! And that can move from Publisher "Panther Books Limited" to Publisher "Granada Publishing Limited" quite easily if we separate the fields. But we might want to regularise "Limited" and "Ltd." and "Ltd" in publishers. And where do we stop with the publishers? A publishing company can still be publishing several imprints while being owned by another publishing company, or might be within a Division of a Publishing Group, or even within the publishing division... to keep things simple, I'd record the lowest level of Publisher above the imprint, and try and group ownership of publishers in the Wiki. We haven't got indefinite hierarchies sorted yet (see Series support) so I wouldn't even attempt to put those in the database itself yet. BLongley 13:43, 19 Feb 2008 (CST)
Here's an old discussion about "Del Rey"[1]. As for imprints and publishers two separate fields would be ideal, short of that I would vote for imprint/publisher as the preferred order.Kraang 19:14, 19 Feb 2008 (CST)
"del Rey" is a simple imprint case. Probably a fairly simple Publisher case too - Ballantine Books, Ballantine Publishing Group, maybe they're just "Random house" now? BLongley 15:51, 20 Feb 2008 (CST)
Del Rey Books, an imprint of The Random House Publishing Group, a division of Random House, Inc., New York. (from 2007) --Roglo 16:00, 21 Feb 2008 (CST)
Thanks! Does that mean "Ballantine" is omitted now? if so, only WE can preserve that history! (Ok, dozens of other sites could too, but WE are special, aren't we?) BLongley 17:10, 21 Feb 2008 (CST)
Not quite omitted; there is Ballantine Books on the spine above the DEL REY logo, Ballantine Books, New York on the title page, but it is not mentioned on the verso of the title page where they state 'published by' with description as quoted previously. --Roglo 17:28, 21 Feb 2008 (CST)
Here's how Locus lists publishers and imprints & publishers[2].Kraang 22:21, 19 Feb 2008 (CST)
Good for that year (2004), I see. If there's similar for other years this might be a good source of data for imprint/publisher ownership changes. BLongley 15:51, 20 Feb 2008 (CST)

(unindent) I don't really see that the higher levels of the corporate hierarchy are of much interest, at least not to most of our users. I think if we document the imprint name (where there is one) and the publisher in the DB records, and leave the history of which imprints belonged with which publishers, and which publishers belonged to which parts of which corporate entities in the wiki, we will be providing for most needs. (Del Rey is an easy case, it has never been a separate firm, and it has never belonged to any publisher but Ballantine, and most of the minor variations in how the name is displayed on book covers are not significant of anything -- "Del Rey Fantasy" and "Del Rey Science Fiction" were separate lines with different editors (Lester and Judy-Lynn del Rey) for a time, as I understand things. Other cases are more complex.) For a given publisher or imprint, we should probably regularize, at least where differences are of no significance: for example "TOR", "A TOR Book", "Tor", and "Tor Books" do not indicate any meaningful distinction. Likewise "Putnam", "G.P.Putnam", "G. P. Putnam", and "Putnam Books" probably are not significantly different, but in this case there is enough history that it is possible one of these forms might be a clue to dating. -DES Talk 18:05, 20 Feb 2008 (CST)

Well, with several family businesses like Putnam and Collins there's definitely a chance to separate some generations of the publisher and clarify dates, but the periods are too wide to put an exact year on them (so we need to improve on '0000-00-00') and most of the books are so old and from secondary sources that I'm very suspicious of the Publisher name being entered correctly at the moment. We could definitely regularise "G.P.Putnam" and "G. P. Putnam" though, as we do with authors. BLongley 14:15, 21 Feb 2008 (CST)
One good reason to have SOME link to the higher levels of the corporate hierarchy is that that is where we can find some of the history that the publishers still admit to, even after an imprint has died: and they're actually quite good at mentioning which of their many imprints and subsidiaries and divisions are SF. But I agree, it's not going to be of common interest and I think we can leave this to Publisher's pages (if enough notes and multiple web-links are possible there) or the Wiki. BLongley 14:15, 21 Feb 2008 (CST)
One other data element to consider, strange as it may sound at first, is the book's distributor. There are two scenarios when it's useful to know the distributor. The first one is when a much larger company serves as the distributor for a smaller SF publisher the way Simon & Schuster does it for Baen. This information can help explain overlapping ISBN ranges and also clear up confusion when we import data from places like Amazon.com which sometimes confuse the distributor and the publisher. After all, they got that box in the corner from Simon & Schuster, so S&S must be the publisher, right? :)
This could be handy with US books. Not so useful with British books, which usually don't have the distributor printed on them - and ADDING a known distributor could add to confusion, e.g. most of my newly-acquired US Books from the time I lived in London could ALL be credited to Titan Distribution. Even though they may have only distributed to one store (Forbidden Planet) and may not have had exclusive distribution rights. BLongley 14:15, 21 Feb 2008 (CST)
The second reason why it may be useful to know the distributor is to help our users to find obscure SF books. Some UK based small presses have an official US distributor and providing that distributor's contact data (often a PO box in the middle of nowhere) can be very useful. I would hesitate to put this information in the Wiki because so few passive users ever find the Wiki, but perhaps we could make it a free text field in the Publisher record? Ahasuerus 23:47, 20 Feb 2008 (CST)
That aside, it sounds like there are three data elements that we are interested in: "place of publication", "publisher" and "imprint". One obvious issue with the "place of publication", which we mentioned briefly above, is whether we want to capture it in the Publication record, in the Publisher record or in the Wiki. I can see how capturing the data at the Publisher level would help reduce clutter in Publication records, but on the other hand, if a publisher moved some time during 1986 and published 12 books that year, how can we tell which books were published where? More importantly, if we keep this information at the Publisher level, how do we distinguish between "New York: Random House", "London: Random House, "Toronto: Random House" and "New York-Toronto-Berlin-London-Tokyo: Random House"? Or do we create separate Publisher records for each geographic location? Ahasuerus 23:47, 20 Feb 2008 (CST)
I'd argue over "place of publication" if you mean which city the publisher is based in. The only times it's been of use to me is to distinguish a country - and a currency symbol in the price field is shorter and quicker. Or you need to go to a lower level of detail, e.g. when researching Orbit at the address level I could spot when an imprint/publisher moved in with another publisher, and go look for indications of a buyout or a merger - but despite 8 to 10 "different" addresses (I think they renamed the buildings occasionally, without moving) they would all be "London". At times looking at where a book was PRINTED looked useful, but the same US book reprinted for the English Market by Scottish and German printers didn't make things any clearer. BLongley 14:15, 21 Feb 2008 (CST)
I tend to suspect that "London" or "New York" are probably not of much use, and bill says. Howver "Basking Ridge, NJ" may be of use in tracking a small publisher. i cna see how the street address (when listed, which it often is not) might be of use in tracking merges and other changes of large publishers, which might sometiems be of use. But most peopel won't want to see this level of detail IMO, so if we have a publication-level field for publisher address, we need to have a user option to display or hide such info, I think. -DES Talk 14:37, 21 Feb 2008 (CST)
I don't think we want a Publication-level Publisher-Address. Even when I was feeling masochistic enough to go check, I started from ONE sample from each year, and only checked others when there was an obvious change from one year to the next. If there's some doubt about whether one publisher/imprint is the same as another, then spot-checks like this can help - but this is a case where the pains of recording lower-level data in the short-term don't really give us equal benefits in the long-term. I don't want to destroy overly-detailed entries here in the interest of simplicity IF there's a use for them, but I also don't want to make everyone go back and check EVERYTHING they've ever entered to populate a new field of low value. Still, Publishers like this (same as this one?) might need the detail in the short term - at the moment a clearer address should probably go in the Wiki, but are we beginning to have some ideas about what fields we want on the database Publisher records? "Notes" seems a good start! BLongley 15:09, 21 Feb 2008 (CST)
As far as "publisher" and "imprint" go, how does a naive editor (i.e. one that is not familiar with the history of various SF publishers) know which is which? To pick a book at random, "The Einstein Intersection" (Ace #19681) says "Ace Books, a Division of Charter Communications, Inc." Should our naive editor assume that "Ace Books" was the imprint and Charter Communications the publisher in this case? And if only one data element is present, e.g. if it says "FPCI" on the title page, how is our naive editor to know whether FPCI was an imprint or a publisher? Come to think of it, even I am not sure what its legal relationship with Crawford Publications was and I probably have more books/magazines published by the Crawfords than most editors. Was it a successor? An imprint? A separate entity set up by the Crawfords for legal/tax reasons? Since we currently do not distinguish between imprints and publishers and simply enter as much or as little information as the publication provided, it's not a problem at the moment. However, if and when we create two separate fields for this data, it may become a major headache, so we may want to think about it ahead of time. Ahasuerus 23:47, 20 Feb 2008 (CST)
I think this is the least of our worries. The naive editors are also the moderated editors, and WHEN we can agree on what we're recording, most moderators will either know their imprints from their publishers (and of course some names will be both), OR know where to go look for it. BLongley 14:15, 21 Feb 2008 (CST)
And once we have this sorted out, we will develop a help page or series of help pages, at which editors wanting to become less naive can be pointed. But there is no point tryuing to develop a help page until we decide exactly what info we want to collect and how it should ideally be entered. -DES Talk 14:37, 21 Feb 2008 (CST)
Well, we're not good at updating help even when we DO agree on something. :-/ At the moment I'm happy if we start clearing up the obvious problems (when we understand which ARE obvious, e.g. regularising some names is good IF we start simple with spacing problems, but start checking before we merge all slashed entries with unslashed similar names). We could probably even cope with some ("suggested" for now) "Canonical" Publisher or Imprint names to argue over, just so that we can indicate, for instance, that somebody HAS recognised that "Tor" and "Tor Books" are linked somehow - nothing major, just showing the possible variants for now and if it turns out that Tor was used for one period and "TOR" for another and "Tor Books" and "TOR Books" covered other years then they'll be fairly easy to switch later. Canonical Author Names are a pain to switch, and I already see Canonical Artist names being assigned too early. Indicating the links is fine, and adding data to each is fine, it'll help in the long run. I think I'm beginning to want "Notes" at Publisher level ASAP for this, as we're two clicks away from the Wiki at the moment and some people ARE lazy enough to ignore that as "too difficult". :-/ BLongley 15:26, 21 Feb 2008 (CST)
Going back to the proposed split of the "Publisher" fields in two, one for "imprint" and one for "publisher", I am sure we could easily come up with ways to handle Tor, Corgi, Putnam etc., but there are literally thousands of obscure imprints/publishers that can't be easily pigeonholed as one or the other. Sometimes it may require a fair amount of research to clear the issue up, at which point we would have to go back and repopulate the right fields in previously entered Publication records. Although probably workable in 90%+ of all cases, this would be a departure from our general WYSIWYG approach to publications. Since we create all other unstated-in-the-Publication-record data associations at the Title level, be it variant titles or pseudonym attributions, this would a significant architectural change.
Having said that, I recognize the value of collecting imprint- and publisher-level information, but I am hesitant to record "unstated" data at the publication level. I wonder if we would be better off if we continued collecting stated publisher information in a single field and then linked it to the Publisher/Imprint table in some fashion? Perhaps the filing logic could automatically create a "publication-publisher" link that we could then clean up whenever warranted? Ahasuerus 20:43, 22 Feb 2008 (CST)
OK, I've spent most of today whacking some British Imprints and Publishers around, and STILL can't see any problem with splitting the publications we have into Imprint and Publisher. Not sure if it's an imprint or a publisher? Put the same data in BOTH fields. Small publishers often ARE their own imprint. BLongley 19:02, 23 Feb 2008 (CST)
Sticking with a single field is NOT going to add any value even if we could agree if it's "Publisher / Imprint" or "Imprint / Publisher" or any other convention. We need to slim down the number of imprints and publishers and link them better - or just continue to let "Publisher" be a meaningless free-text field. I've already encountered imprints owned by multiple publishers SIMULTANEOUSLY - and I'm not giving up yet. BLongley 19:02, 23 Feb 2008 (CST)
Clearly, slimming down the number of "publishers" is a laudable goal. If we can come up with reasonably clear guidelines (well, as clear as most things get around here :-) for populating the two proposed fields, I am sure we will go to Al straight away and recommend that he create a new field. I am just not sure how easy it will be to write something simple yet comprehensive.
Let's take a simple example, Kim Harrison's The Outlaw Demon Wails, which is hitting the bookshelves as we speak. It's a hardcover book, volume 6 in a very well known series from a major publisher, so it should be as unambiguous as it is going to get. And yet when I look it up in OCLC, I see that 22 libraries have it cataloged as by "Eos" and 148 as by "Eos/HarperCollins" even though all of them are working off the same OCLC-provided guidelines. If professional bibliographers, who do this stuff for a living, can't get it right more often than 85% of the time, what are the chances that our motley crew of volunteers will do better than 70/30?
No "HarperCollins/Eos" variations though? Great! "Imprint/Publisher" seems to be the way to go if we want to keep stuff in one field. Or we now know how to separate them if we want two, if we're checking with OCLC. But we should have enough data here soon to record Eos as the imprint and HarperCollins the publisher. In fact, I see you mentioned Eos was the imprint when you added it to the Sources of Bibliographic Information. I assume you'll update the Publisher page for Eos shortly then? BLongley 14:22, 25 Feb 2008 (CST)
Yes, I need to create a "Publisher" page for Eos. I have been looking for an overview of all genre publishers/imprints and I think I have finally found something useful. I have created a Publisher:History of SF Imprints page, but it needs better formatting and then we may want to move it some place else, e.g. Help. Ahasuerus 18:37, 25 Feb 2008 (CST)
I suppose even if we decide that we can't expect a high rate of accuracy during the data entry phase, we could still create an "imprint" field, tell our editors to do their best to populate it correctly and then go back and adjust the data based on our knowledge of various imprints/publishers' history. This should improve our accuracy rate to 95%+, but we would be in danger of incorrectly assigning imprints to publishers in borderline cases.
I am not crazy about these scenarios, but on the gripping hand, one could argue that we are not doing a great job of capturing imprint details anyway, so would we really be in worse shape if we went that route? <insert a picture of Hamlet or Buridan's ass here> To quote an old Analog story, there is got to be a better way (hopefully)... Ahasuerus 23:41, 24 Feb 2008 (CST)
I personally use the "if in doubt, do nothing" on individual cases, but add data where I can, it might help later. And we're a specialist case - if we CAN'T improve on what general libraries are doing, why does the ISFDB exist? BLongley 14:54, 25 Feb 2008 (CST)
I could have a little grumble about what some editors are doing - e.g. I see a lot of regularisation of Publishers going on without anyone updating the Wiki page to say what the eventual aim IS. E.g. "Pinnacle/Zebra" or "Zebra/Pinnacle"? I saw a lot of "Pinnacle" becoming "Pinnacle Books" yesterday, for instance, is that going to be reflected in the final pages? Are we going to leave those problems on the Wiki or resolve them in the database structure? I've already seen some of my Wiki pages disappear as the page I was working on for an Imprint got redirected to the one I wanted as A Publisher of the imprint (Yes, there are more I want to link the Imprint to later), and I hope that was just an accident during the namespace move. BLongley 14:54, 25 Feb 2008 (CST)

(unindented)That was me sorting out the names and identifying the Pinnacle/Tor(1981-late'83) titles before Tor became a separate Publisher/Imprint (in late 1983). As fore the Pinnacle name I went with "Pinnacle Books" since this appears to be its most widely used form. Regarding notes in the wiki I felt it best to add them in when I was finished because as you point out if the publishers name changes the notes don't move with it. You have to go and find the old wiki title and move the notes to the new name manually. Most of the publishers I've worked on was only to bring the variant titles (mostly minor order or spelling variants) under one name. Just so everyone knows my preferred order is Publisher/Imprint where it applies.Kraang 19:20, 25 Feb 2008 (CST)

Still, great activity, and I think it's NEEDED, but I'm not sure we're all pulling the same way. A new Imprint field on the publication is what I want, but if I/we don't get that then maybe an Imprint namespace on the Wiki is the way to go. But we really could do with a better aim as to where we WANT to go. BLongley 14:54, 25 Feb 2008 (CST)

Entering Reviews of Ace Doubles

The question has come up as to whether magazine reviews of stories in Ace Doubles are entered individually, or as a whole. It was my understanding that reviews went in individually, but I can find nothing on the help pages on this. I've moved the Discussion here for more input.--Rkihara 10:24, 23 Feb 2008 (CST)

It seems to me that reviews are normally reviews of titles, not of publications, and that ACE doubles are merely publications of titles which have been, or might be, published in other forms as well. If I have understood correctly, there is no reason for our record of a review of a title to be forced to link to the ace double pub, even if that was the pub the reviewer mentions. Rather it should link to the title record, where the ace double pub, and any other pubs, will be listed. Or have I misunderstood the issues here? -DES Talk 15:29, 23 Feb 2008 (CST)
In ISFDB terms, Reviews are of Titles. Ace Doubles are Omnibuses with one ISFDB Title consisting of two other titles separated by " / " (which may or may not appear here separately). So if you review an Ace Double, you have to make sure the review is of the joint title if you want the link. If you enter both reviews separately, you might link to reviews of separate Titles which may not be (and actually, probably aren't) the same versions as the ones in the Ace double. BLongley 16:16, 23 Feb 2008 (CST)
In the REAL world, reviews are often of Publications that we do have here - but it's not uncommon to review a Hardcover and (Trade?) Paperback edition simultaneously. They'd be reviews of an "Edition" I'd say - but we don't have the concept of "Edition" embodied here. It's possibly a bit late to point this out, if people HAVE been entering the separate halves of Ace Doubles, but if you enter a review of the Double's title you're getting closer to matching the actual "Edition". If the review says "both novels have been severely truncated to fit into this slim paperback" it's a bit misleading to point to the original full-length titles. Or in reverse, "this Ace Double edition restores all the text that the original paperback novel versions of the magazine series lost". Even if a review is only of ONE publication, the Title linkage means you don't know which publication or edition was reviewed. Date of Review and Date of Publication usually mean it's fairly clear which edition or publication were reviewed though. BLongley 16:16, 23 Feb 2008 (CST)
Personally, I like the fact that reviews are recorded, as if they review something we haven't got here we can go FIND it. But if you review something separately that only ever existed as half of an Ace Double, you're sending me on a wild goose chase. BLongley 16:16, 23 Feb 2008 (CST)
Basically, reviews as entered HERE are either fillers in other works or useful pointers to titles we do want. Making them point to something that never existed is a bad thing, IMO. I'd quite like REVIEW records to point to the Publication(s) they review, but we've already got enough suggestions that will make everyone go back to square one on their own collection and redo everything! BLongley 16:16, 23 Feb 2008 (CST)
Ace doubles are, or if they aren't, IMO, clearly ought to be, recorded as omnibuses of two titles with two distinct title records, just as any other omnibus is. DES Talk 17:49, 23 Feb 2008 (CST)
Yes, we spent a lot of time making sure of that last year. "Two Novels" seems to work well, there are problems when one half is a collection or an anthology. As usual, discussions petered out before we could get a consensus on how to deal with OTHER doubles. BLongley 18:38, 23 Feb 2008 (CST)
Each of the titles might be, and in many cases has been, published separately. Therefore, IMO, any review of one of the works in an Ace Double should link to the separate title record for that work. DES Talk 17:49, 23 Feb 2008 (CST)
I can only read that as a request to link to the WRONG edition. Please explain! BLongley 18:38, 23 Feb 2008 (CST)
Reviews do not currently link to any edition, they link to a title, and IMO that is what they ought to do, unless we actually come up with an edition-level record. If we do that, much will need to be rethought. -DES Talk 19:49, 23 Feb 2008 (CST)
If both included works are reveiewed, there should be two review records. DES Talk 17:49, 23 Feb 2008 (CST)
Why, if it's one edition? Do you want us to create separate titles for halves of something we are sure about? We have a lot of phantom entries just to deal with pseudonyms already, how does it help to add more? BLongley 18:38, 23 Feb 2008 (CST)
Yes, I do. More accurately, often a single review column will discuss half a dozen or more works, some of them related in one way or another, others not. IMO every single title discussed at any length in such a column merits a separate review record. The titles already merit separate records, and if i under stand your comments above, are already getting them. as to why reviews ought IMO to link to title records, not publication records, it is simply that often, even with the review in front of you, you can't tell accurately which specific pub is being reviewed -- indeed it may well be an ARC being reviewed, so it won't precisely match any possible pub in our DB. And a review of one publication of a title is usually also a review of other publications unless very significant revision has occurred. -DES Talk 19:41, 23 Feb 2008 (CST)
Yes, a review may refer to a particular version of a work, which may match some publications and not others. However, until we have a workable concept of "edition" here, there is no practical way for a review to link to all and only the publications that have the same version of a work as the reviewer reviewed. Furthermore, often a review of one version is none-the-less helpful in considering other versions, and some reviews mention changes between versions. In short, for the present, i think that reviews should link to title records, not to publication records, and that when versions are different enough that the user would want to know this, a mention in a note on the review record (or perhaps the publication that contains the reviews, if need be) is the way to go. -DES Talk 17:49, 23 Feb 2008 (CST)
If people added notes about "what was being reviewed" to reviews it would help a bit. They don't though, so I'm happy to stick with my view that Reviews as we currently have are just guidelines to Publications. It's not as if the reviews we record actually say anything ABOUT the book reviewed - we just say something WAS reviewed. BLongley 18:38, 23 Feb 2008 (CST)
Please note that currently reviews are in fact listed on title display pages, and not on publication display pages. -DES Talk 19:46, 23 Feb 2008 (CST)

(unindent)First, let's make sure that we are all on the same page, folks ;) In ISFDB terms, an Ace Double, just like any other omnibus, consists of a single Publication record which has at least three Title records associated with it. One Title record is of the "Omnibus" type and the other two Title records match the respective type of each half of the Double. The Omnibus Title is entered as a collaboration between the authors of the two halves, so an Ace Double that includes a Bulmer novel and a Hamilton novel will be displayed as "Fugitive of the Stars / Land Beyond the Map (1965) by Edmond Hamilton and Kenneth Bulmer". As Bill pointed out, this works reasonably well when the two halves are both novels, but there are some complications when one or both of them are collections or anthologies. That's fodder for another discussion, though.

The way Author Summary pages currently work, they display both the individual title representing one half of the Double (usually a novel) as well as the Omnibus title since the latter is recorded as a collaborative work. For example, the Bulmer Summary page will display the Omnibus that I cited above as "Fugitive of the Stars / Land Beyond the Map (1965) [O/2N] with Edmond Hamilton".

The way Review records currently work (although we hope that it will be changed in the foreseeable future), they are linked to individual Titles whenever the Title record and the Author record match lexically. This design seems to leave us with three choices:

  1. We can decide to enter Ace Double reviews as reviews of collaborative omnibuses, in which case they will be displayed for that Omnibus Title, but not for the two Titles that comprise the Omnibus. One could argue that this is a problem since our hypothetical "naive user" may well look up "Fugitive of the Stars" and "Land Beyond the Map" under their respective Novel Titles and not under the collaborative Omnibus Title and never find the review.
  2. We can decide to create two Review records for each Ace Double (one for each half), but then the review will not be linked to the Omnibus Title and if a user looks it up there instead of under the individual Titles, he will once again fail to find the review.
  3. Finally, we can decide to create three Review records, one for the Omnibus and then two for the two halves of the Omnibus. This will cover all bases, but will involve additional data entry.

I suspect that the third approach would give us the most bang for the buck, but I am hesitant to propose a complete overhaul of Omnibus reviews since we expect a change in the way reviews are handled Any Time Now (tm).

As an aside, Ace Doubles tend to be more complicated than regular omnibuses since in many cases Wollheim had to cut the original text, sometimes substantially, to make two novels fit into one book, yet he still advertised the cut versions as "complete and unabridged". That's a whole different can of worms, though, and we probably don't want to go there at this point. As long as we stay away from these complications, we can discuss the way we record all Omnibus reviews; Ace Doubles will be just a subclass. Does this make sense? Ahasuerus 20:53, 23 Feb 2008 (CST)

Maybe the novels could be entered separately, but identified as part of an Ace double by the entry of the series number (Ace M-111 for the example above) into the "series" field. The two novels are now linked together, but still listed separately, something we can do now without any programming changes.--Rkihara 22:32, 23 Feb 2008 (CST)
The problem with that is that many works published in Ace Double format are in fact part of real series, and we don't currently support a work being part of more than one series. Also, series data is recorded at the title level, not the publication level, and this method does not take into account that many works initially printed as ace doubles were later reprinted in other forms.
My feeling is that we should take the second approach, or possibly the third. i don't like the 3rd title record, and I would hope that a later revision of how omnibuses are handled would render it unneeded. But for the time being it is here.
The obvious way to handle the length changes is with variant titles -- "(abridged)" or the like -- but determining which ones were significantly abridged and creating all the vt records would be non-trivial. -DES Talk 00:25, 24 Feb 2008 (CST)
You can nest a series, e.g., Ace Doubles/M-series/M-111, and publications can be put in series, consider the anthology series, such as Damon Knight's "Orbit" series. The publisher data will differentiate between different printings of the publications.
Alternately, the Ace doubles could be entered by series number onto a dedicated page, much like the ones we now use for magazines. I believe the doubles disappeared in the late seventies, so the page would probably be smaller than the one we use for F&SF.--Rkihara 00:57, 24 Feb 2008 (CST)
I dislike 2 as you're misdirecting people going from review to publication. I'm not keen on 3 due to the amount of rework involved. I'm not sure how far away we are from 1 at the moment. How many people go from publication or title to review anyway? Yes, it would be nice if everything linked both ways (e.g. every story in a collection could link to a review of the collection, as well as every publication of that collection, and the collection itself). But I for one only get concerned if a review points at something we're MISSING, never the other way round. Sure, it's nice to see how many times something WAS reviewed, it's a good indicator of the expectations people had of a title. But that's a "nice-to-have" presentation thing, reviews of something that doesn't exist is a real concern to me as that's what I use reviews for. So if 1 isn't acceptable, then 3 please. How many Doubles get reviewed anyway? BLongley 06:48, 24 Feb 2008 (CST)
Whereas I, at least sometimes, use the review links displayed with a title to find the actual review, and read it, either to decide about the title, or to compare my view of a work with a reviewer's, particularly if it is a reviewer I know something about. 16:38, 24 Feb 2008 (CST) —The preceding unsigned comment was added by DESiegel60 (talkcontribs) .

(unindent) here is an example: Solar Lottery. It was first published as part of Ace double #D-103, it has been reprinted many times since singly. It has multiple reviews, at least two of which, from the dates, almost surely reviewed the ace double edition. All the reviews link to the individual title record for Solar Lottery; none link to the publication record, and none link to the joint title record. IMO this is exactly the way this should be handled in most if not all cases.

Well, obviously none link to the publication record, the database doesn't work that way yet (and may never do). I'm a bit concerned that none link to the Omnibus - when we are working on MISSING titles based on reviews, there's nothing in the separate review records that could point us at a missing Omnibus, just missing Novels. Still, your example is of one with plenty of data, in which case I can see that I probably want reviews sorted by date (I know, another new Feature Request that Al won't have time for) as it was a mild pain to pick out the relevant reviews that can only have been of the Double (if we are complete in that area and there was no 1955 separate publication). These are the relevant ones, I think?
  1. Infinity Science Fiction, November 1955, (1955), reviewed by Damon Knight
  2. Galaxy Science Fiction, November 1955, (1955), reviewed by Floyd C. Gale
  3. The Magazine of Fantasy and Science Fiction, August 1955, (1955), reviewed by Anthony Boucher
If so, I find myself wondering whether the first Magazine is correct as the other two review the other title as well. Unfortunately that verifier is inactive. BLongley 17:08, 24 Feb 2008 (CST)
It is perfectly possible that it is correct. Knight might very well have reviewed the Dick title and not the title it was bound with, if the second title did not fit into his column or was less interesting in his view. Which also reminds us that a reviewer might not choose to review both halves of a double. But it is surely possible that our listing is incomplete. -DES Talk 17:18, 24 Feb 2008 (CST)
Solar Lottery (D-103, May 1955) was an Ace original while the Brackett half was a reprint of the 1953 serialization, so it's entirely possible that some reviewers concentrated on the new novel and skipped the one that their reader were already familiar with. By the way, this is where Double Your Pleasure: The Ace SF Double by James A. Corrick comes in handy. He has a lot of useful data about uncredited cover art, month of publication, subsequent reprints, etc, etc. It's one of the very few books that I have with me on the road and one of these days I will make sure that his information is reconciled with ours. One of these days... Ahasuerus 18:12, 24 Feb 2008 (CST)
Here's an example of where I'm coming from. This review appears to be of a Double. If it's not - separate review entries please. If it is - well, we've got some more hunting to do. If people entered all reviews as for the individual titles, we'd not even be inspired to go look. BLongley 15:21, 25 Feb 2008 (CST)
Hmm. After Rkihara's comment to me, I did a quick scan of reviews with slash in the title, & concluded that the only ones likely to be Ace doubles were ones I'd entered or "fixed". Obviously my eye skipped right past that one, so possibly some others.
I think I'm pretty much in agreement with Bill on this, but maybe not, so let me stick in my $0.02, especially as I guess I'm the original source of all this debate. DES said, way back, "It seems to me that reviews are normally reviews of titles, not of publications, and that ACE doubles are merely publications of titles which have been, or might be, published in other forms as well.", & followed this up with some similar comments when Bill kept talking about editions. But there's no question but that we're talking about titles; the double is a pub, as is any other edition of the same work, & that pub has a title. The reviewer is working from a specific publication, whose (individual) contents may in fact differ markedly from other editions of the same works.
Moreover, these are recorded as omnibuses; but is there really any fundamental difference in this between an omnibus & a collection? When a reviewer reviews a collection of short stories, and discusses some or all of them individually by name, does this mean that we should be entering a separate review for each story? Even though those stories have their own title records and exist in other pubs as well, conceivably even as chapterbooks on their own?
I haven't myself run into any cases where a reviewer reviews only one half of an omnibus, but others have cited them. In such a case, whether it's entered as a review of the one part or as a review of the omnibus, I'd want a note. But where all the contents are discussed under one common heading for the omnibus, it seems to me really misleading to enter it as two separate reviews. (I have yet to see a case where there was a pub comment saying these reviews were together the reviews of an omnibus; had that been done, I'd probably have followed suit. But as it was, it seemed merely to be a clear case of making the entry reflect the actual contents of the pub.) -- Dave (davecat) 10:27, 26 Feb 2008 (CST)

As to series issues, consider The Beast Master published as part of ace double #D-509, but also part of the Hosteem Storm series. -DES Talk 16:35, 24 Feb 2008 (CST)

Oh, sure, there are many novels that appeared as Ace Doubles and were (or later became) part of a series. However, the proposal, assuming I understood it correctly, was to set up a new Series for Omnibus title records, not for the individual Novel titles. The series would be similar to the way we set up Anthology series like Orbit even though some of the stories published in Orbit may have been part of another series.
Having said that, I am not sure if setting up an "Ace Double Omnibus Series" would help us specifically with reviews. Of course, it would make it easier to track Ace Doubles and if there is enough interest in the subject, some of our users may find it useful. Ahasuerus 18:12, 24 Feb 2008 (CST)
I thought, but perhaps was mistaken, that it was suggested above that the two titles in a double should be put into a series named for the the catalog number of the double involved, to indicated their link. -DES Talk 19:02, 24 Feb 2008 (CST)
Now that I have re-read Ron's proposal, I believe that your interpretation is likely correct. Unfortunately, as you pointed out, the software doesn't let us associate the same Title with more than one Series, which has been known to cause problems, e.g. R. A. Salvatore's Servant of the Shard is volume 3 in his Paths of Darkness series, but also volume 1 in The Sellswords series. Since we don't have software support for this scenario, The Sellswords series currently starts with volume 2 :-(
Also, it's worth pointing out that we have occasionally asked Al to add support for "publication series", which would let us organize Publication records in a series of sorts; it would be used to link publications like the Del Rey "Discovery" series, Ace SF Specials, etc. If and when this level of support is added, it will be easy to create a "publication series" for Ace Doubles. Until then, we use this Wiki to document things like Series:Macmillan's Best of Soviet Science Fiction Series. Note that at one point an editor tried to create a regular Title series for it, but stopped after the first book, presumably because he realized that some of the books (e.g. some of the Strugatsky books) already belonged to a Title series. Ahasuerus 19:36, 24 Feb 2008 (CST)
Speaking of Ace SF Specials, it seems someone has dealt with these by different Publishers: "Ace SF Special [1]", "Ace SF Special [2]", "New Ace SF Special" (and a stray single title under "Ace SF Special"). Does anyone know who did this, and if so can they be informed that the square brackets appear to cause problems with the Publisher Page Wiki Link? BLongley 12:25, 25 Feb 2008 (CST)
I think it was Scott Latham's idea, but I am not 100% sure. Too bad Scott is currently inactive, so we can't ask him. Ahasuerus 13:01, 25 Feb 2008 (CST)

Linking to Author images

Reposting from User talk:DESiegel60:

I have approved the addition of Tom Shippey's picture since it is hosted by his academic site and I doubt that it will cause Saint Louis University bandwidth problems :) Still, if we are going to link to some types of sites without first securing their explicit permission (which I assume we don't have in this case), we probably want to update our policy. I wonder if it is safe to say that all .edu and .gov sites are OK? Ahasuerus 14:32, 24 Feb 2008 (CST)

I assumed that an author's picture, hosted by that author's site, would be normally permissible, as it would be unlikely to cause bandwith problems, and that explicit permission was more important when a site would be the source for multiple images. But if you think better, I won't insert such a url again without seeking explicit permission from the site. -DES Talk 16:01, 24 Feb 2008 (CST)
I think we got burned with Wikipedia images a while back -- we had no idea what the WP policy was and were quite surprised when the issue came up -- so we may have overreacted a bit by requiring that we get explicit permission from all sites that we link to. Let me post this on the Standard board and see where the discussion may lead us. Ahasuerus 20:00, 24 Feb 2008 (CST)

Audio Bindings

In this discussion on the community portal, BLongley pointed out that listing audio books simpl;y as "audio" in the binding field is to lose possibly significant format information. I therefore propose that our standared ought to be "audio (format)" in the binding field for auudio books, with available choices being "Audio (cas)", "Audio (CD)", "Audio (DVD)", and "Audio (ebook)". Other choices can be added as they arise. "Audio" without a format listed should be coirrected whenever possible, and should not be entered for new records unless there is a note explaining that the format is not known. What do you think? -DES Talk 06:56, 1 Mar 2008 (CST)

Well, let's have a look at what we currently have before we go defining exact choices: here's the counts of '%audio%' and '%tape%' and 'CD' bindings from the last backup I loaded: my opinions in third column.
audio			182	Too imprecise
audio CD		143	OK
audio cassette		88	OK
audio MP3 CD		8	OK
audio mp3		7	Not OK - is this a physical CD/DVD or a download?
audioboo		4	Too imprecise
audiotape		1	Change to 'audio cassette'?
1 livre + 1 CD audio	1	Not sure what to do about Book + CD
OverDrive Audio Book	1	Not OK
Digital Audio Cassette	1	Should be 'audio digital cassette' maybe?
audio bo		1	Too imprecise
audio ca		1	Change to 'audio cassette'?
Audio MP3-CD		1	Change to 'audio MP3 CD'?
Audiobook		1	Too imprecise
tape			5	Change to "audio cassette?
CD			19	Too imprecise
CD-ROM			3	Need to check - could be games
MP3 CD			9	Change to 'audio MP3 CD'?
mp3cd			2	Change to 'audio MP3 CD'?
I can't see an example of 'audio MP3 download' or any other non-physical formats. BLongley 09:11, 1 Mar 2008 (CST)
Non-physical formats may be outside our scope - I know I've only spent a few hours on it, but I've not yet found one with a stable record number of any kind. The audio Project Gutenberg items may deserve it when they arrive, but I've not seen examples yet. BLongley 19:12, 1 Mar 2008 (CST)
I also suspect a lot of our more recent 'unk' formats are audio books, and there's a couple of thousand of those to look at. One other thing I'd like people to consider is recording Abridged versus Unabridged versions of Novels, or if they're recordings of Plays rather than the original novel. We should already be recording Narrator/Reader/Performer(s) in notes - and moving them out of the Co-Author status I often see them in. BLongley 09:11, 1 Mar 2008 (CST)
There are a number of audio ebook versions at Project Gutenberg. These are not yet entered, but will be in time. Note that "Digital Audio Tape" in that order, is the proper name of a specific format (often known as "DAT"). But I would in any case use the general format of "Audio (format string)" whatever the "format" string turns out to be -- it might be "DAT" or "MP3-CD" or "MP3-ebook" in proper cases. -DES Talk 12:42, 1 Mar 2008 (CST)
Also see the complete list of binding types as of October 2007 posted above. Ahasuerus 13:18, 1 Mar 2008 (CST)
In there interest of keeping the binding field relativly short and cannonical, of the forms you list above, I would change:
  • audio CD -> audio (CD)
  • audio cassette -> audio (cas)
  • audio MP3 CD -> audio (CD-MP3)
  • audiotape -> audio (cas)
  • Digital Audio Cassette -> audio (DAT)
  • audio ca -> audio (cas)
  • Audio MP3-CD -> audio (CD-MP3)
  • tape -> audio (CD-MP3) (subject to verification that this is accurate)
  • CD -> audio (CD) (subject to verification that this is accurate)
  • MP3 CD -> audio (CD-MP3)
  • mp3cd -> audio (CD-MP3)
The others you list are too imprecise to convert without more data, except that "1 livre + 1 CD audio" I would probably handle like two separate publications forming a "boxed set". -DES Talk 13:26, 1 Mar 2008 (CST)
Well, unless you're volunteering to add brackets to the majority of our current standardized books I'd stick with the top ones for now, and hope Al can convert via a script to the FINAL preferred choice if we eventually reduce Binding down to something we can offer as a standardized list rather than free-form text. It'd be easy to list the irregular ones as a separate project page if people want to work on them, but as I've knocked a few dozen around a bit today that should probably wait till after the next backup. For now, it's easy to combine this with the Publisher standardization - e.g. I've just canonicised "Books on Tape" and there's plenty of publishers/imprints with "Audio" in the name that are a bit of a giveaway as to what they usually publish. ;-) BLongley 15:51, 1 Mar 2008 (CST)
Looks like about 250 edits can handle the ones where the info is there, and the transformation would be straightforward. That shouldn't be overly onerous. But what I most want is standards for going forward. Once those are agreed, I'll be happy to join in the manual conversionb work -- it this point it seems as if waitign for Al on anythign that others can do is a poor idea. -DES Talk 17:13, 1 Mar 2008 (CST)
OK, I'm lazy - I'd rather convert "audio" or "unk" to "audio CD" or "audio cassette" as appropriate (which I think is useful), rather than "audio CD" to "audio (CD)" and "audio cassette" to "audio (cas)" which adds nothing. I believe we can regularize usefully. Use of brackets can wait, and if we regularize then it's easier to convert eventually. The case doesn't matter much, as even if we get searches by binding, by default it will be case-insensitive search. BLongley 19:12, 1 Mar 2008 (CST)
I whacked a few more audio publications around a bit today, and discovered one new oddity. Bookcassette®. I'm open to suggestions on how to record these - see here for an example - I suspect we've classified some as plain cassette versions though, and they obviously aren't if you need specially equipped cassette players. Not a huge problem (as I don't think we actually have any Editors or Mods looking at these) but something to bear in mind. BLongley 15:34, 2 Mar 2008 (CST)
Audible.Com? They have been around a long time; just purchased by Amazon.Com -DRM though. Emusic? Just recently in non-DRM mp3 format. Most downloadable books are derived from a physical counterpart.--swfritter 18:43, 29 Mar 2008 (CDT)

Amazon Blogs

I keep coming across such for active but little-known authors, e.g. Brandon Massey. The URLs don't look especially intuitive so I wonder if they're stable enough for use here? (We can have LOTS of websites for an Author, that's not the problem - Garth Nix has SIX already for instance - but stability is an issue for me.) BLongley 16:48, 29 Mar 2008 (CDT)

Amazon has been attracting more and more writers and editors lately; the attraction is quite understandable since that's where their potential customers are. This can be quite useful, e.g. Lou Anders' Amazon "listmania" page has various lists of books that Pyr has published (or will be publishing). I added a pointer to his page to our Wiki's Pyr page just a few hours ago.
Stability is certainly a concern, especially given Amazon's image handling history, but I would guess that they would be reluctant to upset their "resident" writers by butchering their blog URLs. I could be wrong, of course, but it seems to be worth a try.
Another concern that we had in the past was linking to fan sites -- as opposed to author-authorized sites of living writers -- but it doesn't seem to be a problem with Amazon's blogs. Ahasuerus 05:05, 20 April 2008 (UTC)

Two Variants of something we don't record

It seems we have two novelizations of the same film (Capricorn One) here by Ken Follett/Bernard L. Ross and here by Ron Goulart. Clute and Nicholls imply they're different versions, although I'm not so sure as the UK one refers to the US Fawcett edition: would they do that if it was significantly different? Either way, it's a bit of a mess: if they're the same, then Ron Goulart is a pseudonym of Ken Follett as well as being a real person himself (allegedly): if they're different, then they're both variations of the Screen-Play (which we don't have).

  1. Does anyone have a copy of the US version so we can compare?
  2. If they ARE different, how do we record the relationship?

For once, I'm tempted to create variations BOTH ways round and watch the ISFDB go into a recursive loop. Suggestions welcome! BLongley 14:27, 31 Mar 2008 (CDT)

I have Goulart's Capricorn One in my collection and should be able to check on Saturday. Do you want to post the first sentence of three random chapters (assuming the novel has chapters) here so that I could compare? Ahasuerus 00:57, 17 April 2008 (UTC)
Sure. Chapter 1: If a city is a lady, Houston is a whore. Chapter 6: Charlie Brubaker had a red-white-and-blue trundle bed and red-white-and-blue wallpaper, done in an airplane motif, in his bedroom. Chapter 9: The long black limousine pulled in outside the air terminal. (I don't think those three lines will spoil anything for anyone. Unless they're fond of Houston. ) BLongley 17:59, 17 April 2008 (UTC)
Nope, nothing like what Goulart wrote. My 189pp mass market paperback consists of 35 short chapters with a lot of blank space on half the pages. Goulart was a very busy man in the 1970s, trying to juggle as many projects as he could -- with mostly predictable results. Ahasuerus 05:12, 19 April 2008 (UTC)
OK, clearly different. Quality issues aside - do we want to record a relationship of some kind? Or let people figure it out for themselves? BLongley 23:56, 19 April 2008 (UTC)
I think the safe thing to do would be to add a note to both Title records explaining that there are two different novelizations of this film by different writers and that they have been confirmed to have different texts. Safety first :) Ahasuerus 04:56, 20 April 2008 (UTC)

Published Bibliographies

The Book of Andre Norton contains a bibliography from 1975 of her works, compiled by Helen-Jo Jakusz Hewitt. It references a number of publications not included in the ISFDB, particularly German titles and publishers. The only information beyond title, publisher, year and language is pages for the first publication. Is this worth entering? How to credit? Any volunteers - 11 pages. -- Holmesd 22:17, 3 Apr 2008 (CDT)

I'd certainly enter missing English titles, I personally wouldn't bother with the German ones but then I'm biased against languages I can't understand. ;-) Does it do better than this site though, which quotes that article and might have used it as one of the sources anyway? BLongley 13:42, 4 Apr 2008 (CDT)
The referenced site has the information from the bibliography, but isfdb does not. Sounds like your recommendation is check the publications here and create new ones as appropriate (leaving blank the pages, title, etc.) but add a note for the source. -- Holmesd 12:05, 5 Apr 2008 (CDT)
Yes, that worked for me. Was it OK by you? BLongley

Further question - There is a list of shorter fiction at the end. One sample entry states: "The Boy and the Ogre" Golden Magazine Sept., 1966. This title is not listed currently under Andre Norton. I'd rather avoid doing periodicals - would someone else take these? 21 titles max - the first two titles had all publications already. Holmesd 04:07, 17 April 2008 (UTC)

I'm not particularly interested (so many other things to do!) : but I'd dump them on the Author Page instead of your own talk page if you want them picked up and worked on. From dealing with other titles from the same Bibliography, they look useful. BLongley 21:19, 19 April 2008 (UTC)
Good idea. Done. -- Holmesd 03:35, 20 April 2008 (UTC)

Publishers

I think we've done a lot with the new tools, but we lack direction. I don't actually WANT to be directed, but when someone is actively promoting one variant over another it would help if we knew when we were working against someone else, at least. Mods don't have a view of the most commonly used variations (OK, they may have seen the file I provided of all current variations, with number of pubs attached, but that gets out of date VERY rapidly with the smaller publishers) and there's no way yet to guide or suggest things to Editors apart from direct messaging. BLongley 00:26, 14 April 2008 (UTC)

If you feel we lack direction and yet don't want to be directed yourself then it seems up to you to provide the direction. Marc Kupper (talk) 07:41, 19 April 2008 (UTC)
OK, first directive to everybody: "Go sort out a chosen publisher/imprint you feel fairly expert with, and then post the link here as an example of what YOU want". Once we have some examples we can establish some guidelines as to what will NOT annoy others, which is a start. We may even find some things we all agree on, or at least not disagree with - e.g. I can't see anyone complaining about a Start Date for a publisher or imprint that we can use to spot earlier misclassifications. (WOTC/TSR for example.) Yes, there'll be arguments about what the start date SHOULD be, but that can come next on a case-by-case basis. BLongley 22:32, 19 April 2008 (UTC)
I thought you might. ;-) BLongley 19:42, 22 April 2008 (UTC)

I think what I'd like is that new submissions get checked against the current "master list" of publishers and new publishers have to be approved specially: maybe by making sure the Wiki page exists for new publishers? That should make people stop and think a bit. BLongley 00:26, 14 April 2008 (UTC)

I don't want to do this manually. If you want to add a tag to the publisher record saying "canonical" (publisher or imprint) then the software can take care of warning users if they are entering a non-canonical name and perhaps to suggest the canonical name. Marc Kupper (talk) 07:41, 19 April 2008 (UTC)
I don't want to do it manually either. I think we need some software changes though (as mentioned below) to take this forward - even if we tagged every current publisher for Canonicity or not, with redirects in the Wiki (I've used this a bit for some Imprints) that's far too much work to set up and still requires too much work to use. BLongley 22:32, 19 April 2008 (UTC)

Secondly, I'd like some way of people taking a bit of responsibility for current publishers and suggesting which ones SHOULD be canonicised, even if temporarily. That would need software changes to allow people to mark a "preferred" publisher and also suggest such from the "unpreferred" ones. And vice versa: suggest imprints when somebody wants to use a publisher like "Random House" or "Hachette Livre". We've thousands of "Scholastic" books here that could be better categorized. (Point Horror/SF/Fantasy come to mind.) I'm happy to take a bit of responsibility for a couple of dozen imprints I'm pretty sure I can deal with, and advise on a few score more (yes, they're mostly British ones!) but so far I can't think of a workaround unless every Mod agrees to check publishers against Wiki pages in the meantime? BLongley 00:26, 14 April 2008 (UTC)

I'm not sure what you mean by "a bit of responsibility for current publishers." Can you provide an example of what a responsible person would do? Marc Kupper (talk) 07:41, 19 April 2008 (UTC)
As it stands: Claim a publisher as something YOU'RE working on, on the Wiki page. Work on that publisher - fill in missing ISBNs, missing SF titles, research ownership of the publisher or imprint, establish earliest and last dates the imprint was used, or for the publisher existing: go ask the verifiers of pubs affected as to whether they can update to match the proposed standards or have other suggestions: link to other related imprint's Wiki pages, other publisher's Wiki pages - or clear those if it's a dead end you don't want. Bug Al for software changes to match what you want (eventually - he deserves a rest) but I still want Imprint and Publisher on each and every publication, and Imprint/Publisher hierarchies by date in the long run. In the short term - maybe we have to do that in the Wiki. BLongley 22:32, 19 April 2008 (UTC)
My priority is to move to development. Al posted the current python code and so I'll be updating my server. Marc Kupper (talk) 07:04, 21 April 2008 (UTC)
I also don't know what you mean by "We've thousands of "Scholastic" books here that could be better categorized." Scholastic is a publisher and I'd assume there are many publication records much like there's many for Ace, Tor, Ballantine, DAW, etc. Marc Kupper (talk) 07:41, 19 April 2008 (UTC)
And it seems many "Scholastic" books here are just classified that way rather than recording any useful imprint that would say it's SF or not. "Point Horror" and "Point SF" and "Point Fantasy" are imprints I can support separating out. Ballantine seems to have many that can be merged usefully, but not all. "Ace Books" and "Ace" I'm not sure of the value of separating, but "Ace (5th printing)" and such-like can be merged I'm sure. I think that is now considered a failed experiment? BLongley 22:32, 19 April 2008 (UTC)
I've tended to enter publications as stated and know that publishers are not consistent. I've personally found the "Ace (5th printing)" to be quite useful. Just yesterday I wondered why there were what looked like two identical pubs (same ISBN/price) and clicking revealed one was an 11th printing and the other 12th. Adding the printing # to the db is on the queue and once that's in then yes, the publisher name "Ace (5th printing)" will go away. Marc Kupper (talk) 07:04, 21 April 2008 (UTC)
Are the plans to just add the field or use it for sorting/grouping? I can provide examples of the headaches that the British "Numbering continues across Imprints/Publishers" causes if you want to dive in the deep end, but just recording and displaying it is a good start. BLongley 19:42, 22 April 2008 (UTC)
As it is, the single largest group is the 4930 publisher names with only one publication each. Some like "Macmillan, an imprint of Macmillan Publishers Ltd" and "Tor Books, an imprint of Tom Doherty Associates" seem easy to clean up but apparently whoever entered those thought it was important to note exactly what a publication says or perhaps that's from Amazon. Do you overwrite their work? It's a lot of grunt work to inspect each one to see if it's just a typo that could be corrected or was an attempt to accurately reflect what's stated in a publication record and if this information should be preserved in the notes or some other method. Marc Kupper (talk) 07:41, 19 April 2008 (UTC)
At the moment, my Publisher edits are mostly reshuffling information from Publisher to notes or vice versa. Or just rearranging the order of imprints or Publisher/imprint in the Publisher field. I've questioned some editors that add the "imprint of..." bits but no response so far. I'm definitely conscious of destroying data, and am avoiding that as best as possible (I hate having to record stuff in notes that I think SHOULD be in the database though). BLongley 22:32, 19 April 2008 (UTC)
What order are you using? imprint/publisher or publisher/imprint? The "imprint of..." thing certainly is clear but I agree it looks messy and can result in the publisher name getting truncated. Marc Kupper (talk) 07:04, 21 April 2008 (UTC)
I generally use whichever is most common for now if I'm not working on primary copies. The spacing around the slash can be regularized via scripts if Al is willing to run such, as can swapping publisher/imprint around: it's only on the smaller numbers of variations that I'm willing to regularise manually. However, for the British paperbacks I'm currently looking at the value of having publisher on each pub is often questionable - the imprint alone is valuable, the publisher can be recorded broadly over all pubs for that imprint for several years, and the level of publisher to record is debatable - e.g. Triad books were published by a consortium of three other publishers and the useful data seems to be which TWO imprints are on the pub. BLongley 19:42, 22 April 2008 (UTC)
Oh, and I DO override Amazon Data when I can see it's from there and no human has apparently added anything. If anyone here edits stuff and leaves it looking as if they were Amazon or Dissembler, please speak up. BLongley 22:32, 19 April 2008 (UTC)
When I add a publication record where the source is not the publication itself I always add notes explaining the source(s) for the various fields and also date it with the month-year. If I override something I often leave a note saying the record used to have "x" but the source is unknown (or seems to be Y). For something like the page count or a binding of tp from Amazon I won't bother with noting the override. My thinking that much of the data, even on Amazon, has a valid basis. For example, I held and researched a submission yesterday where someone wanted to change the artist and it turned out the Canadian and USA editions had different covers. Marc Kupper (talk) 07:04, 21 April 2008 (UTC)
The amount I'm willing to do to a publication without adding notes varies according to the reason I'm adding/modifying it, my perceived value of existing data, and my perceived reliability of the source for the changes: e.g. I've added a lot of OCLC data to the Andre Norton submissions made from the 1975 bibliography, and it's no trouble to paste the OCLC record number into those edits which could be useful if people want to go back and add other editions. The dozens of British Horror anthologies I added over the weekend (and am regretting at times due to the sheer boredom of the hundreds of merges they cause) came from specialist Fan sites, cross-checked with Amazon and one or two bookseller sites and sometimes OCLC, with me doing the regularization and over-riding obvious typos - listing all the sources used for each pub would be more work than I'd be willing to take on, and I wouldn't want to influence a verifier by making it look authoritative anyway. The sites proven useful are being recorded at Author, Publisher and Series level as appropriate though, and I'm going back and fixing the debatable parts as I do the merges - e.g. trying to avoid 27 more J. S. Le Fanu variants by sticking with proven ones here already rather than letting every single variation from a fan get through. I'll have to accept that some things are beyond non-primary entry though, series books in particular are notoriously badly classified at Amazon and absent from OCLC. But at least we HAVE some data to work with now - seeing that a series was missing volumes 6-16, easily found elsewhere, started it. Then discovering that a series LOOKED complete but actually didn't have more than a title prompted a bit more. Then spotting where the series was taken over by another editor prompted a bit more. And discovering that some of the authors in the anthologies have been under-represented here (if they even existed at all before) led me to more sites... I'll have to knock that on the head for a bit though, I'm getting very bored with merges, having the submission queue get that long seems to slow the site down (hopefully due to them throttling MY use of the site alone though - or was it slow for everyone when I had 30-40 edits waiting?) and work-pressures mean I might actually stick to JUST "books received" for a bit - some publisher work has led to me purchasing items specifically for research work on publishers. I'm glad I don't have a wife to justify my spending habits to... BLongley 19:42, 22 April 2008 (UTC)
OK, a bit of a waffle there again - sorry! I don't actually LIKE British Horror Fiction reprints much, and am frankly sick of the sight of "Green Tea" yet again. I can't promise that I won't fill in obvious gaps in our coverage when I find a good source, but I really want to concentrate not only on British publishers for a bit, but on Science Fiction in particular. Not Fantasy - not Ghost stories - not Horror stories. The downside of course is that when researching those I BUY books, whereas dealing with Horror let me leave my credit-card unused for a while. BLongley 21:46, 22 April 2008 (UTC)

Doc Savage and house name and numbering

I'm working on the Doc Savage series. Just entered all of mine so I need to start with secondary sources. There are already some of the blackmask.com reprints listed. They have some entered as by Robeson and Dent. Dent was the main writer that used the Robeson house name so having him as a co-author seems wrong. One of the ones is 18 Murder Mirage And Mystery Under The Sea. According to The Doc Savage Fact Page pulp list Murder Mirage is by Laurence Donovan. I think I should change the omnibus to be just Robeson, and then make pseudos for the novels for Donovan and Dent.

Sounds a reasonable plan! Ahasuerus 00:01, 24 April 2008 (UTC)
With 190 books in the series it seems like a totally unreasonable plan. But I've never claimed that my bookaholism is reasonable. At least I'm resisting the urge to fill in my collection with the 140 I'm missing. Dana Carson 07:28, 24 April 2008 (UTC)

Will ask on the verification page if anyone has copies of the blackmask reprints to see what the copyright page actually has.

The number of the series is Bantam numbering. The Doc Savage Magazine number is different. Should I switch to that or leave it? The Bantam series did add new books at the end so it has more than the pulp series did. Dana Carson 09:31, 23 April 2008 (UTC)

Unfortunately, series numbering is not a precise science. Going as far back as the Lensman saga, there has always been tension between the "original publication order", the "revised publication order", the internal chronological order, etc etc. And then there is Moorcock :( We generally try to use publication order, but we will adjust things if there is a good reason to. In this case I would just go with what seems to make the most sense and add a note to the Wiki page associated with the series. Ahasuerus 00:01, 24 April 2008 (UTC)
A simple example would be Narnia - same books, different order. A more complex (as in "Why, Oh Why did I start doing this, Oh my God, it'll never end...") might be the Star Trek Pocket Books. There, it seemed wisest to use the longest series as the canonical numbering to avoid unnumbered extras as best as possible. And there's a few we've left as overlapping such as The Invaders and The Girl from U.N.C.L.E. as the variants are complex enough already or there's little overlap or there are so few it's not that difficult to list both orders. But for BIG series, I would almost always go for the categorization with the longest run as the main one, and reorder the shorter series in the Wiki page. BLongley 00:36, 24 April 2008 (UTC)
Fantasticfiction has the Bantam covers and the pulp numbers. It was a little confusing when the number on the cover image didn't match the number in the caption. Dana Carson 07:28, 24 April 2008 (UTC)
One other thought is that as one version is the Magazine series and so presumably one publication per title, a manual Wiki Publication Series like this might work for that side and use a 'proper' ISFDB series for the other? BLongley 17:53, 24 April 2008 (UTC)

European graphic novels

Are European SciFi and Fantasy graphic novels (e.g. Jean Leloup's Yoko Tsuno or Didier Crisse's Kookaburra) considered to be covered by the database?

According to Definitions in the Rules of Acquisition, comics and graphic novels are excluded. In practice it's a bit different: graphic novel versions of text-only books seem to be mostly left alone as other editions, and certain Big-Name authors like Neil_Gaiman and Alan_Moore seem to have some of their Graphic-only pubs included for completeness. I for one rarely go to the trouble of deleting SF Graphic Novels and Manga when I find they've sneaked in, but I do occasionally have a clean-up of Non-SF ones. I certainly wouldn't recommend adding new ones at the moment though, other editors are a bit more zealous. BLongley 12:43, 3 May 2008 (UTC)
Alright. I'll probably just clean-up the Bernard_Werber stuff (on Graphic Novel version of Empire of the Ants and Exit, a three-volume series he wrote) then. Too bad though. There's a lot of great stuff out there. Circeus 17:05, 3 May 2008 (UTC)
I'm also not sure exactly where the line is drawn on "graphic novels", though. it seems to me that soemthign like the original version of Niven's Patchwork Girl, with one full-page illo to every 5 or 6 pages of text, should be in, but that was marketed as a "graphic novel", IIRC. -DES Talk 00:18, 4 May 2008 (UTC)
Eric and The Last Hero have art on every page and I'd count those in. I guess I'd draw the line at the point where there was no separate text, or where a lot of the sense was in "speech bubbles" or something. BLongley 09:48, 4 May 2008 (UTC)
How about the 2000 graphic novel version of Guards! Guards!? The German translation of it is (oddly enough) there ("Ein Scheibenwelt-Comic" = "A Discworld comic"). Circeus 01:57, 6 May 2008 (UTC)
"Graphic novels" are a rapidly evolving area, so I suspect we'll keep bumping into examples that we never anticipated. The original intent was to exclude comic books since they are really a different breed and not what fiction readers expect when they are looking for "fiction". Besides, we would need to add a number of new fields to support "pencillers", "letterers", "inkers", "colorists", etc. Manga seemed to be pretty close to comics, so it got excluded as well, but now there are illustrated "light novels" and other borderline phenomena that will need to be addressed. Who said the life of a genre bibliographer would be easy?! Ahasuerus 04:38, 6 May 2008 (UTC)
Graphic novel-oriented databases (such as Bédéthèque) and European reference books only use 2 or three of those as relevant (writer, artist, and sometime colorist), so it's not that far from the current cover artist/writer situation. This is much the same that goes in Japanese manga studios: there are often several secondary artists doing minor inking stuff (sometimes even the editor has to do error correction on the art in Japan), but only the primary artist is typically credited. Circeus 01:35, 8 May 2008 (UTC)

subsequent multi-volume "serialisation"

How does one deal with the case where a book gets a subsequent multi-volume publication? Stephen King's It was republished by J'ai Lu in a three-volumes paperback set (coll#s 2892-4, boxed set #6904), as was Asimov's The Robots of Dawn published as two volumes (#1602-3). Denoël republished Asimov's Mysteries (or at least disordered parts of it) in two volumes (#113-4) etc. This is fairly common for these publishers because they stick to a fairly rigid bracket of page numbers. Circeus 22:36, 3 May 2008 (UTC)

I would be inclined to enter those as "It (vol 1 of 3)", "It (vol 2 of 3)", etc. But perhaps there is a better phrasing. In general abridged, revised, expanded, etc works are handled by adding a phrase in parentheses, so this seems like it might follow the same pattern. Other views? -DES Talk 00:15, 4 May 2008 (UTC)
I just found out the situation is even more messed up: the original Albin Michel French publications of It and The Stand were in two volumes, as were the republications by LGF. Re-publications by J'ai Lu, however, were in three volumes, eventually collected in boxed sets (I own the It set, a friend owns the Stand one). How the *** am I supposed to deal with that? Circeus 03:35, 6 May 2008 (UTC)
It might be best to wait a few days before entering any of the titles. I'll give it some thought and if no one else has a solution than we can enter one of the two volume sets and the reprinted three volume set and I'll play around with them.Kraang 03:54, 6 May 2008 (UTC)
Unfortunately, our support for Big Fat Books (BFBs), which later get split into 2+ volumes, has always been weak regardless of whether there is more than one language involved :-( The only discussed approach that could conceivably work within the current framework would be to retroactively convert the original BFBs into Omnibuses once they appear as 2+ volumes. We could then create a Series, which would include the original BFB and its offspring. Unfortunately, the cure is probably worse than the disease in that it changes legitimate novels into omnibuses, does it retroactively and unexpectedly, and, besides, doesn't help when the same book has been broken down in more than one way -- see Circeus' example above.
For now, we have been mostly stuffing these multi-volume reprints as Publications under the parent Title -- see Brian Jacques' Martin the Warrior, which includes 2 volumes in German and 2 volumes (out of 3) in French. If you can think of a better system -- short of rewriting the application -- please speak up, because the current one is admittedly suboptimal! Ahasuerus 04:54, 6 May 2008 (UTC)
Or there is the non-solution used for Cyteen where the parts are simply recorded as separae novels, with no clear indication of the relationship. i don't advocate this. -DES Talk 05:01, 6 May 2008 (UTC)
Cyteen looks particularly horrible, with FOUR Number 1s in the Series, not even showing in the right order. Even the notes are misleading it seems: "split into three sections for its inital [sic] publication" doesn't match our data. BLongley 19:34, 6 May 2008 (UTC)
The problem is that there are really two novels in that series, it just happens that one of them exists in three sets of covers for soem cases (was it for the inital MMPB pub?) I know that CJC was reported to have sworn that it would never be reprinted except between a single set of covers. I own copies of the split version, BTW. -DES Talk 20:50, 6 May 2008 (UTC)
On the other hand, it is pretty well arguable that thsoe two novels dont really form a series, anyway. They are in the same settign, occur at about the same time in internal chronology, and there are about 3 paragraphs in Cyteen mentioning the events of 40,00 in Gehanna. There are no common characters, and the events and thems also incluence other works set in that universe. -DES Talk 20:50, 6 May 2008 (UTC)
Indeed I wouldn't know how else to proceed. *shakes head*. And I'm not intending to add these particular publications for the time being. I intend to first get through Werber's stuff, than my French Asimov, Pratchett, Sheckley and Silverberg stuff. Circeus 05:05, 6 May 2008 (UTC)

just fictionally speculating . . .

Just to satisfy my overgrown curiosity, does Reading Oprah: How Oprah's Book Club Changed the Way America Reads really have a place in ISFDB?

I don't think so, and i have submitted a delete publication edit. I reveiwd the description provided by amazon, and this is a non-fiction book that is not SF, not about SF, not by an author noted for SF, nor about such an author. It is relevant only in the sense that SF books are also books, and any book about how readers read is in some slight degree relevant. -DES Talk 21:35, 5 May 2008 (UTC)
Although the Rules of Acquisition are still questionable, I have to say that DES has got it right. The only way I'd include it is if we had a REVIEW of it that the person providing the review was unwilling to change to "essay". And only then if we made it clear that it's only here because of an SF pub review, and we should not encourage further entries by the same author, same publisher, etc. We can accommodate SOME non-fiction - for instance, bibliographies and SF encyclopedias are handy to know about - and some science works are here as sources sited for SF works. But we do need to draw the line somewhere. BLongley 23:31, 5 May 2008 (UTC)
I am willing to include critical studies of SF. For example, IMO Heinlein in Dimension belongs in, and so do at least the more well-known book-lenght studies of Tolkien. But this isn't such a study, either. -DES Talk 23:44, 5 May 2008 (UTC)
I approved the deletion, maybe some other data base(recycle bin) will find it a home. I wonder how may forests were chopped down to produce it?Kraang 00:45, 6 May 2008 (UTC)
I thought it was a subtle dystopian alternate history. Alvonruff 17:29, 6 May 2008 (UTC)
I think it is just straight-up fantasy. Fantasy becomes reality when enough people believe in it.--swfritter 18:41, 6 May 2008 (UTC)

Magazines reprinted as Anthologies

Take a look at this reprint of Captain Future, Summer 1942, which is currently listed as an anthology edited by Edmond Hamilton. It will be easy to change the editorship to Oscar J. Friend, but the underlying problem is that this is really a reprint magazine in disguise. And, as we know from our past struggles with trans-Atlantic reprints of various magazines, we don't support multiple versions of the same magazine issue particularly well. Any ideas? Leave it as an anthology just like we did with this 1981 reprint of Astounding Science Fiction, July 1939 ? Ahasuerus 05:19, 7 May 2008 (UTC)

I believe there is also a PDF version Amazing Stories v1n1 and physical reprints of others. Anthology makes most sense to me for right now although it would be nice if we could somehow document the existence of the reprint pub on the mag wiki - perhaps a special section?--swfritter 20:09, 7 May 2008 (UTC)
Sure, we could add notes to each affected magazine's Wiki page. We could also update the Note fields in the Magazine pub and the Anthology pub to point to each other. Ahasuerus 23:16, 7 May 2008 (UTC)

Publisher names again

With the advent of publisher editing, and the plans for publisher merging in future, i think we need to resume our discussion of how to handle publisher names, and this time bring it to at least a partial conclusion. I propose the following steps:

  1. We agree on standard forms for which I might call appendages to publisher names: for example should we use "& Co", "and Co." "and Company" or just leave the company off altogether. Should it be "X Books" or just "X". And so on. For these we ought to be able to come up with standard answers that can be applied to most if not all publishers.
  2. Work through publisher by publisher.
    1. For each, those who know most about the publisher will help determine what info is significant. For example does the imprint matter? Does it matter when the publisher became a division of MegaCorp, so that we record "Jones Books" and "Jones Books, div of MegaCorp"? Does it matter when "Smith & Son" became "Smith Books, inc"? etc.
    2. Based on this, we devise a set of canonical names for the publisher and whatever lines or imprints or other variations seem worth recording. (the long list of publishers now in the DB will help with this)
  3. Then we consider which existing forms can be safely converted en masse to one of the agreed canonical forms (using the publisher edit tool) and which will require an editor to do individual research before making changes.
  4. Change en masse the forms that can be safely changed (and merge them once this is available)
  5. Start working through the list of remaining forms doing required research on individual publications.
  6. Meanwhile, once a canonical name or set of names is agreed to for a given publisher, new data should be expected to use only one of the canonical names.

Does this strike people as a worthwhile program? -DES Talk 20:52, 7 May 2008 (UTC)

I like the way Locus lists each publisher's books every year -- see, e.g., their list of all books published by Gryphon Books in 1999 -- and I believe that implementing a similar approach within the ISFDB framework would be useful as long as we can implement it well.
A number of our editors have already done a lot of work in this area and discussed it extensively, but there are two areas that I haven't seen addressed (perhaps I just missed the discussions) and I am not sure how we can handle them.
The first area has to do with ensuring, as David wrote above, that "new data should be expected to use only one of the canonical names". How would the average editor know the canonical name(s) that he should enter for each edition? In most cases we follow a simple "what you see is what you enter" philosophy when dealing with Publications and using canonical publisher names instead seems to be a significant departure.
The second area has to do with merging publisher records. I am sure that 90%+ of publisher merges will deal with things like "Dell" vs. "Dell Books" and other reasonably simple issues, but there will also be more complicated and/or obscure areas where a significant amount of preliminary research will be required, e.g. Australian imprints of similarly sounding UK/US companies and such. How will we ensure that we always contact our expert(s) in these fields before we accidentally mess up months of their careful work? For example, I happen to know that User:Clarkmci has massaged hundreds of Australian publications to get the publisher just right, but he rarely participates in Wiki-based policy discussions and I am not sure whether he knows about these developments. We may be able to drag him here, but what about other specialist editors who have been quietly working in their chosen niches? Ahasuerus 00:35, 8 May 2008 (UTC)
As to your first concern, it is going to have to be a matter of informign individual editors. There will ned to be a page or pages (probably in the wiki) with lists of cannonical publisher names, and thsoe lists will need to be linked to from the various editing help pages. Then the moderators will need to check and inform people when non-cannonical names are used, just as when prices are given without currency symbols, or dates not in yy--mm-dd form. Ideally, there would eventually be a warning when submit is clciked "Publisher name is non-cannonical, do you wish to submit anyway?" But that will have to wait for both programmer time and a moderatly stable set of cannonical names.
(Actually i would love a confirm display after submit, like the "preview" button on wiki edits, anyway, at elast as an option.)
As to merging of publishers, in any case where thinks aren't obviously harmless (merging "Baen Books" with "Baen" or "TOR" with "Tor", say) I would expect the merge to be proposed on the wiki and wait some days for comment. Then if anyoen knows about the xpecialist editor and his work, the comment 'Contact Joe, he knows this area well" can and should be made. Not perfect, but I see no better way.
And remember this is a fairly small community. Currently there are only 50 editors who have more than 50 edits. we can find ways to inform people if we must. And not all will get the word, but many/most will. -DES Talk 00:54, 8 May 2008 (UTC)
At a minimum merging should be a two step process, change name and then merge, this way if a name gets accidentally change to an existing one their not automatically merged. Changing of names at the publication level can be left as it is to general editors, but changing of names to all publications and merging should be moderators only, with no access to this function for non moderators.Kraang 01:00, 8 May 2008 (UTC)

Uncredited vs. Anonymous vs. N/A plus Reviews

As per Help:Screen:NewPub:

  • Anonymous or uncredited works. If a work is credited to "Anonymous", then put "Anonymous" in the author field. The same applies for any obviously similar pseudonym, such as "Noname". If the work is not credited at all, use "uncredited", with a lower case "u". This applies to editorship of anthologies that are not credited. Omnibuses, including dos-a-dos publications such as Ace Doubles, should be given an author of "N/A".

At this point we have a grand total of 2 Essays attributed to "N/A". Is it fair to say that the last sentence quoted above is out of synch with our prevailing practices and we may want to change it?

Also, although we generally use "uncredited" when the anthology editor is not known -- as per the Help recommendation above -- a number of magazine records use "Anonymous" when reviewing anonymously edited anthologies, e.g. see the review section of Science-Fiction Plus, June 1953. Is that by design or do we wan to change all reviews to use "uncredited"? Ahasuerus 00:51, 8 May 2008 (UTC)

Collections in Omnibuses

This section has been copied (with deletions) from the thread at ISFDB:Community Portal#Collections in Omnibuses, because this is really the relevant location, IMO -DES Talk 21:17, 12 May 2008 (UTC).

I was under the impression that for an omnibus that included both one or more novels and a collection we typically did not list the individual shortfiction works in the collection, particularly not if the collection had previously been published separately and has that publication listed with contents. But I am not sure if we have ever really codified this policy one way or the other. (if an omnibus is a collection of collections I suspect we would just treat it as a different, larger, collection). -DES Talk 22:42, 10 May 2008 (UTC)

I was thinking about this the other day. I see two possible problems with using Collection Titles -- as opposed to using their constituent stories -- in Omnibuses. First, if you are looking at an Omnibus page and it contains a Collection, you need to click 2 times to get to the Collection's contents, once on the Collection's Title link and once on a Collection Publication link. Worse, once you click on the Collection Title link, you are then faced with anywhere from 0 to dozens of Publication records. And, of course, each Collection Publication may have somewhat different contents -- essays, forewords, afterwords, etc -- so how do you know what exactly the Omnibus contains? Second, if you are looking at the Title page of a story published in a Collection that later appeared in an Omnibus, you will not see that Omnibus publication, which defeats the purpose of listing all Publications for that story. It seems safer by far to list the constituent stories and avoid these problems. Ahasuerus 23:12, 10 May 2008 (UTC)
I see. A problen is that there is no concept of hierarchy to the contents of a publication, including an omnibus. If one could list both the collection title and its contents in such a way as to make clear what belongs to the collection, it would be much better. There are other cases where this would be helepful, too. For example it is not uncommon for a collection to contain groups of stories with a common group title, as well as individual titles. "X: Three tales" consistign of "A", "B", and "C". without a "group title" element, one is tempted to reder these as "X:A", "X:B", and "X:C". Sometimes this can be done with a series, as in Distant Friends, but this doesnt always work well. In any case, we have often entered omnibus pubs that contain collections without the short fiction -- Indeed I rather thought it a rule that no title in an omnibus could be of type short fiction. For example, see Mountain Magic which contains the collection Old Nathan or 3xT which contains two collections and a novel, but lists no shortfiction (and if it did, how would the user know which short works were part of which collection?) or Annals of the Time Patrol which contians two collections, and many in simialr state. If we decide that the individual works of short fiction should be enterd, i will do that. Help:Screen:NewPub is not very helpful on this point at present. -DES Talk 00:05, 11 May 2008 (UTC)
I think we have to enter the short-fiction. :-( More work, I know. Reasons: 1) We'll never be sure WHICH version of the collection is included, so which variant titles etc are in it. Or even contents: some collections are different in different countries due to having to split them up into separate volumes, or because one or two stories aren't free to be published in that country. (I have MANY books that say "For copyright reasons, this book is not available in the USA" - although Canada normally gets permitted.) 2) People use the pagination for Length-determination at times - personally, I'm not that bothered, but you can't get it from Magazine versions. E.g Story X may be entered as starting on page 22, another story starts on page 25, so it's 3 pages long and therefore Short Story? Not if it's "Continued on pages 125-148" - which those lazy Magazine editors never mention. ;-) It's rare (but not unknown) that a Book publication splits contents that way. ("Destinies" comes to mind though, as a Magazine in paperback format that we've had to deal with as "Anthology" - not sure if it ever did such magazine content-splitting though.) I think Ace Doubles of a pair of Collections have been the worst to deal with so far - workarounds such as Roman Numerals for one side and Arabic for the other have been suggested, but as usual never made official. BLongley 22:58, 11 May 2008 (UTC)
I see your point, and perhaps it is the best way, but it seems to me that at the least the collection titles ought to be entered in addition to the shortfiction titles. If full page numbers are entered, this will make it at least fairly clear what stories belong to what collection title. But note, lots of omnibuses are not entered in this way, even ones that are verified, and if this is to be the standard, the help should be changed. I am going to raise this on the rules and standards page, I think. -DES Talk 15:33, 12 May 2008 (UTC)
I'd favor entering both the collection & its contents. A problem with doing this, which sometimes shows up in other places (book review columns, say) is that we have no way of forcing two titles with the same page number to display in a particular order. But that's relatively minor.
The ideal solution would be a whole slew of programming changes, with some database changes involved, which would always expand collections in searches & displays. I kind of shudder at the changes needed, though. -- Dave (davecat) 16:47, 12 May 2008 (UTC)
I agree about the ideal, but we need a rule for the present.
  1. One way is to go forward as at least some editors have been working, entering the collection title but not its contents in any omnibus publications. This makes the pub displays shorter and simpler, but has all the problems Ahasuerus and BLongley warned of above.
  2. The second is to clearly adopt the standard that when an element of an omnibus is a collection (or, I suppose, an anthology) both the overall collection title and the individual shortfiction (or essay or whatever) contents should be entered in the contents list of the omnibus. If we agree on this, the help should be changed to document this as the standard, and we should be aware that many existing pubs, including verified ones, will not meet the standard.
  3. The other option, I suppose, would be to enter the individual contents of the collection without its overall title. IMO this is the least acceptable option, but maybe other editors will favor it.
Well, what do others thing on this issue? -DES Talk 21:17, 12 May 2008 (UTC)
I've been practicing (2) on my own publications for some time. Yes, we do get caught out when a Novel becomes a Collection, and much rework ensues and more data (pagination and exact titles) is required: it's still quite a bit of work to go the other way but at least you can do that without help (and not completing the task doesn't seem to cause problems anyway, leaving the constituent short-fiction in a Novel doesn't throw up any significant warnings in the database, although some clean-up projects would flag it.) This is not a common problem though, it's less than once a month in my experience. BLongley 22:17, 12 May 2008 (UTC)
There's probably a few pubs still where someone has verified "yes, this is an omnibus containing that collection", and that was possibly OK at the time or we didn't make the current standards clear - many of us new editors verified happily, it was easy when we started, and the questions about what verification actually MEANT came later. (I'm still not entirely sure what it means, if it doesn't "fix" the publication in that state then any edit or outside change like Amazon changing the default coverart-image should probably unverify it "a bit". But that's another discussion, that we've had many times already, and probably needs new blood before it can be resolved.) BLongley 22:17, 12 May 2008 (UTC)
Anyway, my vote is for (2). If someone can find an Anthology example it might help people decide, but I can't think of one. I know I've fixed at least one (1) example recently but can't recall that either. (3) looks really undesirable. BLongley 22:17, 12 May 2008 (UTC)
I verified Mountain Magic quite recently, following the rules I thought applied. if we go with (2), I can easily add the additional contents. I will ned to think of an example of an omnibus containing an anthology, if I can. (2) would be very acceptable to me, does everyone else here prefer it? -DES Talk 23:00, 12 May 2008 (UTC)
(2) seems to be the intuitive choice, so I am inclined to go with it as well. If we find problems with this approach, we can always revisit the issue. Ahasuerus 01:11, 13 May 2008 (UTC)
I vote for (2). (But still the problem what is an omnibus remains, e.g.: Isaac Asimov's Science Fiction Treasury.)--Roglo 08:27, 13 May 2008 (UTC)
I would call that an omnibus, and if we are going to follow (2) converting it to one would be trivial. -DES Talk 14:11, 13 May 2008 (UTC)
My vote would be for (2), it's what I've been doing when I catalog my books that fit this case on Librarything.CoachPaul 01:20, 14 May 2008 (UTC)
Given that in a week no one has supported any other alternative, and several experienced editors have supported (2) above, I will taken this as consensus, and edit the help page appropriately. -DES Talk 17:01, 19 May 2008 (UTC)

I have made a change to Template:TitleFields:Title to record this decision. Please see if this is acceptable to all. -DES Talk 18:24, 19 May 2008 (UTC)

Looks fine to me. Of course, the problems caused by page-numbers restarting for each component will still be an issue, but that can be left for another discussion. BLongley 18:59, 19 May 2008 (UTC)

Merging Audio Dramatizations with Stories?

Someone else may have a better idea but this, I expect, might be an audio collection of dramatizations of stories rather than actual readings of the stories. Matheson's "Death Ship", for one, has been merged with print versions.--swfritter 16:39, 19 May 2008 (UTC)

The alibris description says "Adapted for audio from original scripts written for the television show created by Rod Serling that aired in the 1960s, these radio dramas feature music and sound effects and a distinguished lineup of guest stars". Based on that, i don't think this belongs here at all, any more than DVDs of TZ episodes based on written SF stories. -DES Talk 16:59, 19 May 2008 (UTC)
To add to what DES says if we allow these, then we'll need to add old X-1 and Dimension X episodes which used stories from the 40's and 50's to the db too. I do think that we can show some of this information though, by adding a Tag to these stories showing that they were TZRD episodes as I have done with Dimension X stories here.
I disagree that we'd NEED to add such. At the moment it's a case of "these sneaked in, they've been organized a bit as deleting is a multi-step pain". There's no obligation to ADD anything - e.g. I cleaned up Daisy Meadows a bit over the weekend, but felt no need to go add the information that "Sunny The Yellow Fairy" is actually a variant of "Saffron The Yellow Fairy". This is the slippery slope started when audiobooks were permitted - e.g. we have Doctor Who Audiobooks with no paper editions known - yet. At the moment Video seems right out still, but I know I've listened to "Radio Dramatizations" of short stories here that were single-reader performances, and others that were multi-reader performances, both close to the original text that I recalled. (BBC 7, for those that can receive it - some are rather good.) But demanding a dead-tree version isn't going to help here if there WAS one initially: can we agree a "two-steps removed is too much" rule? Or that Music and Sound Effects rule some audio out? Or do we just deal with them if they arrive and don't encourage more? BLongley 18:55, 19 May 2008 (UTC)
I'm not so sure that we can make generalizations here, and probably for now should go with a case-by-case basis. For instance, these TZRD cd's are filled with radio plays formatted to fit into a "radio hour", in much the same way as old time radio shows were. "The Castle in the Attic" has an audio version, that is complete with sound effects, and each part is voiced by a different voice actor, but the reading is exactly that, a reading of the book, and as such is another edition of the novel and should be here. (I believe the Redwall audiobooks are the same.) Therefore if we say that it shouldn't be here because it has multiple readers, and sound effects it shouldn't be here, we are missing out on editions that IMHO should be here. My thoughts are that if dialog from the story is changed and story elements are either left out or paraphrased, then it shouldn't be here. If it a reading of the story, no matter how many voices and special effects are used, then it should. Partially, if you hear someone utter the words "He said", (feel free to change "he" with any pronoun or proper noun that may fit,) then it is most likely a reading and not an audio play and should be included.CoachPaul 22:08, 19 May 2008 (UTC)
Maybe I should organize my thoughts better, I started the above by saying we shouldn't make generalizations...and ended with one.CoachPaul 22:10, 19 May 2008 (UTC)
I lean towards general acceptance of these (but not encouragement to enter more) when they are purported to be the same story. Otherwise we might have to delete all the "Abridged" audiobook CDs, cassettes, etc. If it's retitled ONLY for an audio publication then I think we can leave it out, they're not important enough to us to create variant titles. I'd discourage inclusion of non-SF material in the same work though - I'm happy with partial-contents entries for fringe publications: e.g. many SF works were published in non-genre magazines, and it's nice to have a reasoned entry for the first publication: we can cope with a "Saturday Evening Post" or "Collier's Magazine" or "Strand Magazine" with SF content only listed. We have "Playboy" first publications for some stories too, but I've never seen one verified yet... perhaps editors get too distracted? ;-) BLongley 22:33, 19 May 2008 (UTC)
Treating them as variant titles and adding "(Audio Dramatization)" to the variant title might be an idea. Abridged stories are sometimes treated the same way. These are actually entries that I find valuable. Movies and TV shows are well documented in The IMDb but this is data that is not as easy to find. I guess we would have to categorize them as shortfiction. If there is no acceptable option for including these then we should remove them. Keeping them is likely to result in setting a standard of inclusion.--swfritter 00:10, 20 May 2008 (UTC)
I would be inclined to include them as well for all the reasons listed above. As far as verifying old Playboys goes, some of them are collectible items, but not necessarily the kinds of collectibles that SF collectors would be interested in, which is somewhat unfortunate since Playboy has published a number of stories by big name SF writers over the years. After all, they had more money to throw around... Ahasuerus 02:25, 20 May 2008 (UTC)

OCLC

Mike Hutchins showed me a way to link to OCLC/Worldcat records like this, which has proven stable for a few weeks now:

 <a href="http://worldcat.org/oclc/30288510">OCLC record 30288510</a>

Given the number of times we're using OCLC/Worldcat to justify dates, explain stub entries for pubs we don't have but think exist, etc, should we be raising the profile of this resource a bit more? If it's good, then it could be a new source to verify against: if only so-so, then guidelines for such links would be appropriate: if VERY good, then maybe we should ask for OCLC references to be added to pub-level entries in the same way that we have ISBN/Catalog numbers. Comments please! BLongley 21:32, 27 May 2008 (UTC)

I now use worldcat a lot -- I have entered many titles (several dozen at least) and probably hundreds of pubs based mainly or solely on data from this source. I almost always go in through the "First search" interface, which seems to give more complete records than the worldcat.org site does. Still, the link is probably a good idea. I would favor adding this as a verification source, with some caveats.
  • Worldcat seems to often shorten publisher names, so the reported publisher name may well not be the form on the actual book.
  • When a book is published as by a pesud known to the worldcat system (or perhaps to the entering librarian) the "author" is listed as their canonical name, and you need to check the "responsibility" line to see the actual name on the printed work.
  • Often an illustrator is listed as a co-author, again you need to check "responsibility" where this is usually spelled out.
  • Dates are rarely given closer than a year.
  • When an illustrator is listed, there is often no way to tell if the credit is for interior art or cover art, or both.
  • Publication formats must be determined from size in cm, although paperbacks are often noted as such. But this includes both trade and mass-market formats, so size must be checked.
  • Often multiple dates are listed, with the earlier being a copyright date, or the date of an earlier edition, particularly if the later ed is a facsimile.
  • Anthologies and collections sometimes have contents listed, and sometimes not.
  • Older works, printed in multiple volumes, often give no page counts but only volume counts.
but with all the above mentioned, I agree that we should raise the profile of this source a bit, and at least consider adding it to the verification matrix. -DES Talk 22:04, 27 May 2008 (UTC)
I agree that OCLC is a very important source, but all of Dave's caveats must be taken into account. And Bill, I think the OCLC number is as stable as they come. (Now that I've said that, it will all fall apart in a few days.) It would be nice to have it as one of our verification sources, which should be easy to set up. Any moderator can do it. But before that happens, maybe Al can create an entry field so that when OCLC is used for verification the OCLC number must be entered and an automatic link be created. MHHutchins 00:02, 28 May 2008 (UTC)

<edit conflict with Michael's last comment>

Some other things to keep in mind:
  • Some editions have multiple records in WorldCat and you have to pull them up side by side to see if there is additional information to be derived from the interplay of different fields. The need to use multiple OCLC records to create an ISFDB publication record could make it difficult to create a single "OCLC number" field even if we decided to go that route.
  • Anthologies/collections which have contents level data often use abbreviated author names like "R. Heinlein" or "R. A. Heinlein" and there is no easy way of telling how the authors were actually credited. Punctuation, capitalization and even spelling are also often mangled.
  • Prices are usually reflective of whatever Baker and Taylor charged the last time the book was available to libraries, which may be significantly higher than the original price of the book. Also, library editions are sometimes priced differently.
  • If we link to the Web version of WorldCat but use the FirstSearch record to enter data, we can confuse our users who may not be able to access FirstSearch and wonder where our data had come from.
  • Some authors are much better covered in WorldCat because the cataloging librarian had access to a specialized collection, e.g. Andre Norton's books are exceptionally well covered (although the data is not easy to find).
  • WorldCat rarely lists printing numbers and its subject headings leave much to be desired, but you can always troll other online catalogs for this data (which opens a whole different can of worms).
Having said that, WorldCat has a ton of useful (and searchable!) information about Authors, Titles and Publishers. As long as we keep its limitations in mind, it can be a wonderful resource.
P.S. I wouldn't refer to WorldCat as "OCLC" in Notes to avoid confusion when new users access our data. OCLC owns and maintains dozens of other catalogs, which we generally have little interest in. Ahasuerus 00:27, 28 May 2008 (UTC)
When I base a record on it, i alswas call it "OCLC/Worldcat", but I have been calling the record number "OCLC:" because i understand that the record number is unique across all of OCLC. -DES Talk 13:51, 28 May 2008 (UTC)
I admit I use "OCLC: nnnnnnn" as that's the simplest copy'n'paste from the interface I'm currently using. But I'm willing to change (obviously, or I wouldn't mention the opportunity to do even more typing!). Thanks for all the comments: I think a lot of these warnings can usefully be added to the relevant help - I hadn't even considered using price info from there though! I'd certainly like to see 'stub' or 'place-holder' publications entered from there refer back to the source(s), (point taken about the multiple records used, Ahasuerus, although I usually find one better than all the others) and I see no harm in adding it as an extra confirmation of existing data. I guess the big question is whether anyone wants to do any coding changes for this, or should we just agree on how to enter such information in notes? BLongley 19:12, 28 May 2008 (UTC)
One other comment, about the weakness of the publisher data there: I really want to preserve Imprint data here, so please do read the notes. Entries like "London: Lane" are too vague (is it "John Lane" or "Allen Lane"?) and sometimes inaccurate ("London: Panther" for a company based in St. Albans instead of London - and I think I've seen "London:" prefix used for SCOTTISH companies!). BLongley 19:12, 28 May 2008 (UTC)
I'm late again but another warning: content listings in WorldCat sometimes have initial articles missing, if you copy and paste without double-checking you might create a dozen of non-existent title variants (I almost did). On the bright side, they have nice 'search by isbn' feature: URL like
http://www.worldcat.org/isbn/1852424761
so it could be added to the left panel with Amazons etc. to use where we do not have OCLC. And my latest 'standard' to record this number was: OCLC: 69022631 (<a href="http://www.worldcat.org/oclc/69022631?tab=details">WorldCat</a>). --Roglo 20:01, 28 May 2008 (UTC)

(unindent) The warnings above are IMO too large for Sources of Bibliographic Information#Aggregate Library Catalogs and Search Engines -- there is already a lot of info on that page. So i have created Help:Using Worldcat data, and linked to the new page from the existing one. (Please take a look at my new page and improve it if you can.) I don't see an urgent need for coding changes to create a worldcat number field in the db -- The notes should do IMO. A sidebar link would be nice, but is not urgent. I think we should add worldcat to the verification matrix without waiting for any new coding, but I don't want to do that unless others agree that it would be a good idea. -DES Talk 21:02, 28 May 2008 (UTC)

Sorry about the comment above re: not referring to the OCLC numbers as Worldcat numbers. I am not sure what I was thinking since I have been using the "Data from OCLC record NNNNNNN" format for well over a year. I think I'll go back to bed now... Ahasuerus 23:16, 28 May 2008 (UTC)

(unindent and restart) 12 days ago, I said above "I think we should add worldcat to the verification matrix without waiting for any new coding". There has been no substantive comment on this since. Does anyone agree? Does anyone object? BLongley, MHHutchins, and perhaps BLongley seemed to support adding Worldcat as a verification source in the discussion above, but ther was not clear-cut agreement on a specific proposal. So, my proposal is to add Worldcat to the verification matrix, with a corresponding line in Help:How to verify data, without waiting for any coding changes in the DB that might add further support for Worldcat. -DES Talk 04:54, 10 June 2008 (UTC)

It can be useful, but (unlike Tuck et al) we'll still need to identify the OCLC record number(s) in the Notes field since there can be multiple OCLC records for a Publication, right? As long as non-primary sources of verification are not displayed on the Title page, I am not sure if the additional functionality will justify the extra work needed to mark pubs as OCLC-verified. Ahasuerus 06:01, 10 June 2008 (UTC)
Yes, we would need to include the record number in the notes field, unless/until Al added a special field (which would probably need to be multiple, like authors and webpages). But we are doing that anyway, at least I am, if I consult Worldcat in creating or checking a pub. Currently, i normally add a note of the form "(DES Month YYYY)" if I add data to a pub record from worldcat, having it in the verification matrix would avoid the need for that. It would be a way of saying "Info in this pub was checked against whatever info was available in the listed worldcat record, and matched except for any data noted not to match". We could just put in a note about Tuck, also, or Locus, as they don't show up on the title screen either. It seems to me that the utility is very similar. Note that having an OCLC record number is not the same as certifying that all available info matched, or noting any diffs. An editor could always choose not to Worldcat-verify -- I don't see what extra work this is imposing. -DES Talk 14:30, 10 June 2008 (UTC)
Ah, I think I see where the difference in our usage patterns is. You add notes like "(DES Month YYYY)" to the Notes field while I generally don't. Now that I am thinking about it, it does seem useful to be able to ask secondary verifiers questions like "Hm, I don't see this data element in OCLC's record NNNNN, how did you derive it? Or has the OCLC record changed since you verified it in October 2007?" With that in mind, moving this information to the verification matrix sounds like a good idea. Ahasuerus 00:47, 11 June 2008 (UTC)

Killdozer by Theodore Sturgeon

I'm thinking that this Title and this other Title, either need to be merged, or the pubs that refer to the revised story from the Notes need to be removed from the first Title and added to the second. The first option is the easiest, the second would take a bit of work. If it were up to me, I would merge them, and then get back to my other editing. Are there any other opinions on how this should be handled?CoachPaul 00:51, 30 May 2008 (UTC)

Since there is a documented difference in the texts, they should not be merged. I'll look into it. -DES Talk 04:46, 30 May 2008 (UTC)
I think the pubs are now all properly linked. -DES Talk 05:06, 30 May 2008 (UTC)
I don't buy it. First off, if I'd just wanted one Mod to make a decision and forgo the whole discussion of the matter, I'd have just made one myself. If you notice I asked for OPINIONS, and not a unilateral decision. Secondly, there are pubs with the same name, but different printing dates in both records. I think that the likelihood that they reverted back to the original draft in subsequent printings unlikely, though I don't believe it to be impossible. Third, if you follow the link to the Lotus Index, it doesn't even mention Selected Stories there at all. Who put in these notes, where else did they get their information from, and why didn't they change it to begin with? I think more discussion is needed on this matter.CoachPaul 15:29, 30 May 2008 (UTC)
The pagination seems to have been lost in this title. I think the correct thing to do is identify HOW they differ: ideally with someone with both versions comparing them. Failing that, a Verification request on all printings to see what notes there are about printing history. When we have that information, we should hopefully have enough to decide on whether the variations are significant enough to keep separate, and at least some of them should be definitely under their correct titles: and we can have a guideline on the "probable" title for the unknown ones. I don't think the word "revised" is a big enough distinction in itself, we need to explain why it's SIGNIFICANTLY revised. BLongley 18:03, 30 May 2008 (UTC)
I apologize if people feel that I acted too hastily. I took "If it were up to me, I would merge them.... Are there any other opinions on how this should be handled?" to mean "If no one objects, this is what I will do. If anyone objects, be prepared to do the extra work involved". I did, I was, and so i did. But if more discussion is needed, fine. Nothing I did removed any info, it can easily be undone.
As to "explaining why" I think i disagree, if I have understood Bill correctly. More preciusely, if in fact there are two version, whose differences are more than trivial, IMO it is enough to call one "Killdozer!" and the other "Killdozer! (revised)" If someone does a side-by-side comparison, I would want the results, or significant portions, in the bibliographic notes in the wiki, for the primary purpose of allowing future publications to be correctly classified. I don't see a need for such notesd in the title records themselves.
They're needed in the notes as that's all that people check (and sometimes not even then). When you try and merge two titles, you can quickly look at each and see whether there is a "DO NOT MERGE!" warning there, go back in your browser, check the other, then submit the merge. Wiki Notes do NOT get looked at at that point, in my experience. There seem to be enough people that look at "Title X" and "Title X (revised)", disagree about the revision on the basis of maybe one sample of each, and attempt to merge, that the reasons against such need to be very visible. I want all "Revised" suffixes justified very visibly, or I can't really blame people for attempting to merge them, or mods for approving such. BLongley 21:59, 30 May 2008 (UTC)
I originally read the story in the "Spectrum" anthology, more than 30 years ago, and I recall it vividly. That was a library copy, but I may own a copy now. I read the "Complete Sturgeon" collections when they came out, but I don't recall the differences in the story. I don't know if I have copies that I can check side by side, but I will see. -DES Talk 18:56, 30 May 2008 (UTC)
The second link to Locus index is in the Selected Stories notes. Killdozer! was on page 5 in Aliens 4, and page 104 in Killdozer! (data from May 19 backup). Page numbers were probably lost when unmerging the stories (SF.net tracker).
The notes are mine (I've added content to Selected Stories). Generally I would prefer to see where different versions of stories were published but I see it may be difficult with stories that were published in dozens of collections so I won't insist. --Roglo 20:19, 30 May 2008 (UTC)
Two apologies here. First, I didn't make it sound like I wasn't willing to do the work. I found the problem, I was, and am willing to fix it. If I wanted to take the easy way out, I could have just merged them and been done with it. Second, as I read back over my response to DES, I was a bit harsh, so I apologize for that too.
I have three copies of this story in my collection that I know of. All three are shown on the non-Revised page, and they all seem to be the same after taking about 20 minutes to spot check all three stories at about 30 or so various places in the story. I agree that if there is a significant difference in the stories, then we need to differentiate between the two. Either way, we will need to put good notes into the note field so that future editors of these stories can make good judgments about where to place which story. I think that there was at least one MOD who had VERIFIED copies of both, maybe they could tell us how big a difference there is in the stories.CoachPaul 21:05, 30 May 2008 (UTC)
Mike Hutchins has this verified as well as the one with the lost pagination under the other title, so yes, we probably have someone that can do a comparison. (If he's willing, of course: it's a very boring task.) I can check mine if people give me something to look for: I think I only have one version though, but I could be wrong as I often depend on me having verified something here to be sure I have it - and whenever someone else has verified it first I lose track of the other editions I own. (How many of us are rooting for Multiple Verifications to be allowed?) BLongley 21:59, 30 May 2008 (UTC)
Paul Williams' story notes in the Collected Stories edition of "Killdozer!" explains the text revision. It's basically the last 8 paragraphs (roughly one and a quarter pages). If the story ends "So they let him." you have the revised version. Williams also reprints the original ending which last line is "That's just the sort of thing they'll expect from him!" That being said, the version in Wondermakers is also the revised version. I'll go back and restore the page numbers which were dropped due to the merge. I discovered this particular bug when I was trying to remove pseudonymously credited stories in single-author collections. That's one of the reasons I stopped trying to make corrections in these. It was too much trouble trying to keep up with the page numbers when doing unmerges, and the work-around involved so many steps it didn't seemed to be worth the effort. MHHutchins 11:49, 31 May 2008 (UTC)
Again my apologies if my edits caused loss of page data, I had no idea that unmerging would do that, and I don't see why it should, but if it does I will be more careful with that in future. Also my apologies to CoachPaul, I guess that I badly misunderstood your initial msg. I suppose that in part I was thinking "If I want the option that takes more work, i ought to do the work". Regardless, we will get things into the proper state.
If you (any of you) found a title XYZ and another title XYZ (revised) would you actually merge them without checking the wiki links? or indeed without asking publicly? Since this isn't the sort of indication that could be from amazon or other such unreliable sources, I had rather assumed that such an indication would be a large red flag. That said, assuming we find and record detailed differences, would it be sufficient to record in the title notes "significant differences found in side-by-side comparison, see wiki for details"? When i noted that a version of RAH's Puppet Masters was "expanded", I did not list any description of the differences -- should I have done so? (I wasn't a mod then, and the approver never suggested such a note that I recall) -DES Talk 22:33, 30 May 2008 (UTC)
I personally never merge without some checks, but the Wiki pages are probably the LAST place I'd think to look. Given the workload for Mods now I doubt anyone else really does make the effort either, unless they've recently come from Wikipedia editing and assume that the Wiki is as important as the database itself. ISFDB predates Wikipedia by some years, and the useful Pub level data in the Wiki is almost negligible. The Wiki only really gets used for things that can't be fitted into the database structure easily: alternate Series orders, some Biographic and Bibliographic Notes, etc. When in doubt as to where it should go, we ask Al to change the database structure and the Wiki stuff is considered almost temporary data - not part of ISFDB itself. Well, that's my opinion anyway - if this becomes ISFWiki rather than ISFDB, then the STRUCTURING of data is lost and the usefulness to me is lost. Even "Put it in Notes" is a short-term cop-out to me: yes, the notes are available in the database downloads, but free-form text like that is a pain to work with. I try, but only so I can point out where we can improve the database. BLongley 23:08, 30 May 2008 (UTC)
Your suggestion that title notes record "significant differences found in side-by-side comparison, see wiki for details" would be enough of a "large red flag" to me though. Why it needed to be recorded in the Wiki rather than the database would still be questionable, but it would certainly be a good start. BLongley 23:08, 30 May 2008 (UTC)
My concern is that excessive detail on an issue like that (and I tend to think of anything beyond "differences exist, see <note>" as excessive) might well be distracting to users who are not editors, and it should therefore be tucked away, available but not in the face. if others disagree with such an approach, then we won't do it that way -- that is how i would do it if I were the dictator here, which i am glad not to be. Also, as I understand it, there is a length limit on notes in the DB, and the results of a true comparison might be rather long. -DES Talk 23:53, 30 May 2008 (UTC)
I don't think Notes tend to bother editors that much: and we need Notes differences to flag dangerous merges - those HAVE to be fairly "in your face" for Approvers. They don't have to be LONG notes - but I'd at least want to see which was the longer version, "revised" is not enough for me: is it revised upwards or downwards? Have I got the Author's intended text or an Editor's? BLongley 22:37, 31 May 2008 (UTC)
A Wiki difference is undetectable by the ISFDB software, and rightly so IMO: the Wiki side is NOT part of ISFDB. A lot of people use the database - there's several other websites based on our data - the Wiki has to (again, IMO) cope with the exceptions that we can't fit into the database yet, and chats like this about how to improve the database content. But all the Wiki stuff is transient - you don't get it if you download the database, so important bibliographical stuff has to be forced into the database somehow, all the arguments, comments and other stuff is left out. BLongley 22:37, 31 May 2008 (UTC)

(unindent) Boy, you spend most of the day up in the air and when you land, you find 13 edits on just one page! :) I am glad the misunderstandings have been resolved, but it just goes to show how easy it is to misunderstand the "tone" on the internet, so we may want to be doubly careful. For example, I recall reading DES's often used comment, "Fine", for the first time and wondering whether it meant "That's fine by me" or "Fine, have it your way". It quickly became obvious that DES had the former usage in mind, but I can see how a simple comment like that could easily lead to a misunderstanding. I am not sure what we can do about this except try to re-read our comments before posting and always give other posters the benefit of the doubt.

Re: using the Wiki for publication level notes, one problem that we have had with this approach is that Wiki URLs are not linked to our submission approval software. When a pub record is deleted, its associated Wiki page becomes an orphan and over time we can end up with a bunch of orphan records. I asked Al about this issue over a year ago and he was going to look into it at some point, but I doubt it's high on the list. If a collection has a convoluted publication history, I try to add notes to the Title record as well as to the affected Publications, since I figure you can never be too careful :) Ahasuerus 02:47, 31 May 2008 (UTC)

Since the differences between the two versions are rather minor (as I noted above), I don't think there should be two records for the story. And until we establish a relational link between titles, the notes field in the title record will have to suffice. I've never used the Wiki for notes, since it's not actually a part of the database. MHHutchins 12:01, 31 May 2008 (UTC)
There are Wiki Pages here?? OK, I really know that there are, but I quit going to them long ago, because everyone that I ever went to was blank. I have now found, thanks to the information supplied by Mike, that my copy of Spectrum 3 (a 1968 pb version) has the revised story, but the two Isaac Asimov Presents books have the other version. I agree with Mikes reasoning of having only one TITLE record for the story with a good and proper note in the proper field, but would also propose placing notes in all pubs that we know contain the revised story that it is slightly different from the others.CoachPaul 13:29, 31 May 2008 (UTC)
I agree. I added my Mammoth Book of Golden Age Science Fiction. It has So they let him. as noted but no mention of any revisions and the 1959 edition is not stated anywhere. I thought I have the original text. --Roglo 21:38, 31 May 2008 (UTC)
Interesting: we have more facts, and yet draw different conclusions. My Spectrum III copy is the So they let him. version. So now I want the original ending with That's just the sort of thing they'll expect from him! - if I'm reading Mike's comments correctly, I'm missing an extra page as the revised version is shorter? And people want to hide that away in pub notes, and merge the variants? I'd have thought that two different ENDINGS was one of the main reasons to keep them separate! I don't want to hunt the longer version down by reading every single publication record in case it has notes, I'd like to be able to pick the most easily available copy of the longer variant. BLongley 22:15, 31 May 2008 (UTC)
Actually the difference between my two variants isn't length, it's just that the last 7 paragraphs are different. Now that I've read Bill's comments, I have to change my mind, again I think, and agree with him. If different endings aren't enough to have two title records, what is? Now my next big question is, in which TITLE record do we put the Stories when we're not sure where they should go, and do we assume until proven differently, that if one copy of say Spectrum 3 has the story ending in "So they let him.", do we assume that they all do?CoachPaul 22:41, 31 May 2008 (UTC)
Re: titles. We could abuse a little the title-parent relation and create variants. Something like:
  • Killdozer! - exclamatory title, unknown text variant
    • Variant Title: Killdozer - non-exclamatory title, unknown text variant
    • Variant Title: Killdozer! (Astounding, 1944) - original text
    • Variant Title: Killdozer! (revised, 1959) - revised version
so we could see all publications together for the parent title, or specific variant if needed. Unknown variants would be merged as usual, with Killdozer! and if someone verifies pub with the parent title, we (after spotting it) could ask the verifier to use one of the specific variant titles. --Roglo 22:01, 1 June 2008 (UTC)
We're doing that to some extent with Authors already - putting a suffix after each one and any that arrive without a suffix need sorting. "David Alexander" seems to have lost canonical example status, "Steve Jackson" maybe still is - but now I look at the Steve Jacksons, they've been mercilessly trimmed down from the last time I looked! One I saw today seems to have sneaked through though: this pub has a note that says "This is most likely not the John Hill that is the Koontz Pseudonym." but has been allowed in under the Koontz pseudonym anyway. I don't think that's good moderating, but I'm up late, tired, suffering from hayfever and can't honestly say I didn't approve that myself. Maybe we need a temporary demoderatorizing policy for my Hayfever season? BLongley 23:49, 1 June 2008 (UTC)
I've made the author John_Hill_(screenwriter) and linked to ImDb. We need something like author disambiguation pages. Once I found three numbered authors of the same name in ISFDB. --Roglo 22:23, 4 June 2008 (UTC)
Thanks for the adjustment, agreed, and which numbered authors? And was it me that let the other John Hill through? BLongley 22:41, 4 June 2008 (UTC)
You're welcome, we have three Kate Thompsons (and wiki doesn't like the square brackets), I think someone was splitting their biblios, and I have no idea who was working on John Hill. --Roglo 23:01, 4 June 2008 (UTC)

"Digest" sized books

Please see User talk:Dragoondelight#Encounter in Space and User talk:Dragoondelight#The Devouring Fire for context. much of what follows will be quoted from those discussion.

Harry (User:Dragoondelight) has attempted to enter records on some books that are not standard Mass-market paperbacks, nor standard "trade" pbs. He wants to call these "digest" format. But our current practice, as documented in Help:Screen:EditPub currently reserves that term for magazines. He says:

"I am defining digest as a Reader's Digest size, not book sized nor trade paperback. It contains two staples with the cover page glued to the pack. The binding is not similar to paperback. I say digest to keep people alert to what they would see. It is a definition that pertains mostly to the binding. Though the binding seem the staples make it look just like a slightly smaller Reader's digest. You can change it and it will make the buyer of one surprised to find the anomaly. I am trying to be correct it what I see. Digest is used because it fits. Is it a book or a magazine. It is a magazine more than a book in it's appearance."
"The first example would be the standard magazine printed in digest format with multiple stories. The next step was the use of full length novel with some old material, such as editorials, reviews and letters to the editor. This becomes less as more space is needed by larger novels. You then evolve to Galaxy Science Fiction Novel model which no longer pretends at other elements of a magazine, but retains the digest size and softcover, not heavy paper front page usually with the staple crimped front to back as opposed to the more usual Magazine center fold staple style. The paperback was current I believe before WWII and was liked by those who could tuck into a small space and not destroy it. It was used by many American troops, and maybe not at all by the British. There are even examples of stories even reduced further than we see today. The digest style was used more in Britain, Europe and American cities where the newstands could readily display them. The size has a certain advantage in that it was harder to steal and hide."
"[M]y copy is a perfect example of a digest not a paperpack. The staples are actually alternately protruding in the upper top side and indented in the lower one. I know it is hard to change an idea, but this looks more magazine than paperback and if you had wanted a paperback or tradeback you would be disappointed. "

What shall we call such a book format? Shall we just use "pb" or "tp" with a note in the notes section describing the unusual format? Shall we use the term "digest", at the risk of possible confusion with the rather different magazine digest and large-digest formats? Or shall we use some different term, and if so what? I am not an expert on the history of publishing, and I am not sure what term to use. I am sure that if we call both what User:Dragoondelight and a recent IASFM by the term "digest" there will be confusion unless we somehow annotate the possible meanings of the term. -DES Talk 21:07, 25 June 2008 (UTC)

Note this entry which lists a Galaxy Novel as a digest as is the case in for least three other novels. If that's valid it seems like the opposite is true.--swfritter 22:16, 25 June 2008 (UTC)
I am still inclined to think that using "digest" for such different formats is confusing. but if that is the accepted name for each, and inventing a new term would not be helpful to users (which is perfectly possible), then IMO we ought to clearly document the multiple meanings of the term as it is used on this site. Wikipedia:Digest size seems to say that it is defiened entirely in terms of size, without regard to the actual method of binding. I don't really care what term or terms we use, but I do want our terms reasonably well defined. -DES Talk 22:32, 25 June 2008 (UTC)
I didn't realize that according to Help we should not be using "digest" when describing book bindings. I think that we would be much better off using the same binding codes for books and for magazines as long as their physical dimensions are the same. Moreover, it's not always clear whether a publication is a book or a magazine -- see the previously mentioned Galaxy Novels (good to see that one of my old Wikipedia article has survived essentially intact!) or some UK publications which were ostensibly magazines but were actually single author collections written under a variety of pseudonyms. If we were to follow what Help currently tells us, then we would have to change binding codes every time we reclassified books to magazines or vice versa.
As far as staples go, they were widely used in the 1950s during the digest revolution, but eventually technological advances made it possible to do away with them, so they are no longer a distinguishing characteristic. Also, now may be a good time to mention that last summer I posted a list of all Binding Types that we are currently using. At some point we'll presumably want to clean it up and/or add clauses to Help to account for frequently encountered gray areas. Ahasuerus 03:42, 26 June 2008 (UTC)
Our current definitions are based more on size than they are on actual binding. There are at least three digest bindings - non-stapled (F&SF), saddle-stapled, and side stapled. And just think about all the possible binding type for fanzines.--swfritter 14:51, 26 June 2008 (UTC)
Oh yes, fanzines are particularly bad, which will make it hard to standardize their definitions :( Ahasuerus 16:23, 26 June 2008 (UTC)

(unindent) Meanwhile i am holding two submissions, in significant part on this issue. Should I, or should I not, approve submissions of apparently non-magazine publications with a binding field of "digest"? —The preceding unsigned comment was added by DESiegel60 (talkcontribs) .

Well, we could approve these two submissions for now, but if User:Dragoondelight plans to continue working in this area we will likely see this issue come up again. I'll post a note on the Community Portal to see if we have consensus re: changing Help. Ahasuerus 16:23, 26 June 2008 (UTC)
Why have we stated in the Help files that Digest can only be used as a magazine type? Did we not know about the Digest sized books at the time, or were we ignoring them? If we have books that are Digest sized, then they should be listed as Digest. I would say that we should change the Help if the Help is wrong. Either that, or somebody hop into their time machine, go back to the 30's or 40's whenever it was, and change the format of the books. I could live with either of these two options. Let me be serious here for a second though if I can. A "digest" sized novel , would not fit into any of the definitions that we currently have for pb, nor tp, so if we had to come up with a new format name, it should be real intuitive to use, something like "Digest Book". We would also need to document this in the Help pages. IMHO, it's easier, follows a more common sense rule, and seems more correct just to change the existing Help files and call a Digest sized book a Digest.CoachPaul 19:30, 26 June 2008 (UTC)
Very well, I will deal with the submissions on hold on the assumption that "Digest" is an acceptable term for a book binding of this sort, and that the help page will be changes to more clearly indicate what "Digest" means in the case of books soon. If we were to decide to use some variant term such as "Digest book" we could change the current submissions to match easily enough. -DES Talk 19:43, 26 June 2008 (UTC)
If the Magazine editors get to use "pb", I have no trouble using one of their categories for books. The help does need updating, as the examples are meaningless to me: "Bedsheet" and "Tabloid" are Newspaper sizes to me, and I don't own early issues of Amazing, or the 1942-43 version of Astounding, or even the British Science Fiction Monthly. The Interzone example is particularly annoying as I recall that most are a nice standard A4: including A4 and A5 as recommended sizes would solve a lot of your British Magazine and Fanzine categorization problems as those cover almost all the amateur printings, although I guess a few might have indulged in novelties like foolscap. BLongley 23:00, 26 June 2008 (UTC)
I agree but we should abbreviate "digest book" to "db" to be consistent with the other descriptions.Kraang 23:59, 26 June 2008 (UTC)
I would respectfully disagree as "db" is way too often used here for "data base" and that would be too confusing.CoachPaul 00:04, 27 June 2008 (UTC)
I don't think anyone will confuse "db" for "data base" in the bindings section.Kraang 01:10, 27 June 2008 (UTC)
It sounds like we all agree that Help should allow us to accurately enter digest sized books, we just need to decide what to call them. I think that simply "digest" would be better than a variation on "digest book" since that way the binding field will be unrelated to the whole separate issue of whether a given publication is a book or a magazine. Ahasuerus 01:14, 27 June 2008 (UTC)
"digest" is fine with me for both magazines and books so long as it's clearly defined what it means, and what the limits are, and we don't get people mangling sizes together TOO misleadingly. If people want to save some typing and move it to "dig" or "dgst" or "dg" that's fine by me too so long as Al does a mass update. I don't really want to separate digest books from digest magazines in the binding type alone if they're the same thing really - although I'm quite happy to support pairs/triples of alternatives for the same thing if they're clearly explained as being similar enough. We should probably look at the UK sizes versus US sizes a bit - I see the older Ace paperbacks are noted as significantly smaller but not worthy of noting, I personally DO note it: but if we DO decide that Bedsheet and Tabloid and such are generally useful, we should explain the nearest equivalents in ISO and other formats. I don't want to end up with everyone measuring every book or magazine they own to the nearest millimetre, but we should agree a few categories and tolerances. Does nearest half-an-inch in each dimension work for you, for instance? Or should it be quarter-inch? I don't want to proliferate bindings too much, but would like better explanations of which bindings are equivalent enough. (Although I do get annoyed when one issue of a magazine doesn't fit into the standard box.) BLongley 01:55, 27 June 2008 (UTC)
"digest" is fine with me. In the context of books it suggests a cheaply bound edition.--swfritter 15:47, 27 June 2008 (UTC)
"digest" is fine with me too and in fact I recently added a chapter-book. In doing so I discovered that ISFDB's support for these is weak and entered the story as a novel (both title and pub record) with a "chapbook" binding, and the contents being the "novel" and a short-fiction title record. Hopefully the publication notes explained the non-standard entry well enough. Marc Kupper (talk) 18:24, 29 June 2008 (UTC)
I think the above actions were a mistake, in that they created a misleading "novel" record. Anyone using the biblio page for a list of novels by Bradbury (a very common use of the ISFDB, I think) will get inaccurate information -- note that such users typically do NOT look at the notes on individual title or pub records. If I had found this record when just looking through the db (say while entering another work by the author) I would have automatically and without discussion changed it to a chapterbook and deleted the "novel" that isn't. We do need better display support for chapterbooks, but adding intentionally inaccurate information to the db is not the way to go. -DES Talk 18:30, 29 June 2008 (UTC)
Even I, that hates books published as Novels getting called chapbooks just because there's newer definitions of Novel-length, can't support a 7 page Novel. I do want to be able to separate individual titles that are individual publications I can buy from titles that are only available in Anthologies and Collections though, and am open to suggestions to how to accomplish this. Today I've worked a lot on some 1950s British paperbacks that look at risk of being converted to chap(ter)books or even magazines (I see "Authentic Science Fiction" has already gone that way) so it DOES need attention soon. Well, if I win the lottery - I can't find a single one of those shilling and sixpence books going for less than at least 60 times that now. 1200 percent modifiers are not unknown. (How does that compare to the stock market?) BLongley 20:20, 29 June 2008 (UTC)
1200 percent or 1200 times? :-) But yes, certain old magazines and paperbacks turned out to be very good investments. On the other hand, for every 20 Million Miles to Earth, which has appreciated some 1,000 times in 50 years, there are dozens of "Spicy Railroad Romances", which may not have even kept up with inflation, so I am afraid that on average you would have lost. Going back to the "7 page novel" issue, I too think that it would be better off as a short story and I am glad to see that it has already been changed. Ahasuerus 22:26, 30 June 2008 (UTC)
For now there seems to be no support for CHAPBOOK at the title level. I agree that NOVEL is misleading. It seems like there's a desire for CHAPBOOKS and so I'll add it as a feature request. Marc Kupper (talk) 05:38, 2 July 2008 (UTC)
Actually, we did have support for Chap(ter)books at the Title level at one point, but it was removed once we realized that a Title published as a chapbook may have also been published in a magazine, a single author collection or an anthology. What Bill is proposing (as far as I can tell) is that we display all Titles that have been published in Chap(ter)books twice, once in the Shortfiction section and the again in a separate "Chap(ter)books" section. I see where he is coming from since many genre encyclopedias display any short fiction pieces that have been published separately in a special way, but it may not be easy to implement. If the Author page logic has to check every Publication for every Title at display time, then the Author page for Asimov will take a very long time to appear. Alternatively, we could add a new field to each Title record that would indicate whether this Title has appeared in a Chap(ter)book, but that's not particularly easy to do either :-( Ahasuerus 07:28, 2 July 2008 (UTC)
My simple requirement is separating "things I can buy" from "things I can buy only as part of something else". If that's a separate display issue - fine, it's a display issue. In the meantime, hiding things I can buy by taking them out of a category I normally look at (Novel, Anthology, Collection, possibly Omnibus although I tend to avoid those as too BIG) is annoying. BLongley 22:05, 2 July 2008 (UTC)
FWIW, Template:PublicationFields:PubFormat distinguishes between pb and tp solely on the size. I've generally also considered the paper quality and binding when deciding how to classify a publication. For example, I have tp size publications with pb quality paper and binding and pb size publications with tp quality paper and binding. With those I'll enter the ISFDB "Pub Format" based on the size and add a note. The purpose of the note is to help explain possible confusion as to why someone may see listings of the publication as a pb in one place and tp in another when in fact both referred to a single publication. I recently untangled DAW Books one-and-only attempt at publishing a book that was 7.9 inches high. They apparently had great plans as they created entirely new catalog # and ISBN # series for this but abandoned these after one publication. As expected, bookseller listings for this are all over the map as the book does not fit into any standard categories.
Looping that back into this thread, I'm in favor of adopting/using new binding types or "pub formats" but we should also keep an eye on the list and to try and regularize it as much as possible. As the binding, other than hard vs soft, is not a common search or distinguishing criteria, I don't want to get too carried away with descriptive language in the pub format field. I believe the goal should be that someone should be able to look at ISFDB and from the publication listings to understand how one publication is different than other publications of the same title. A title that's published only once, a bookazine for example, does not need a lot of fancy language to describe it even though it may be "digest" size. If you have, or are seeking a tile, then it should be sufficient to know that that it's available in softcover and that various ambiguous seller and fan listings of that title all probably refer to the same publication. Marc Kupper (talk) 18:24, 29 June 2008 (UTC)

(unindent) Looks like we have a consensus re: changing Help to allow "digest" books and an overwhelming majority in favor of calling the binding simply "digest". I plan to make the change on Tuesday morning unless there are other comments posted here. Ahasuerus 22:28, 30 June 2008 (UTC)

Fine by me. But can we please have some better definitions? I.E. NOT in terms of other examples we don't have. I can convert inches to millimetres or centimetres, but if people want to make stapling or paper quality important then we need better definitions. I'm currently entering a load of "digests" from a secondary image source which is clear on hc versus pb, and I think I'm right in adding them as "pb" in the meantime till someone adds a physical verification of this new classification. Verifiers can always adjust it later to the new sub-classification. BLongley 22:05, 2 July 2008 (UTC)

Letters to the Editor

Entry is currently defined as debatable but common usage has resulted in a de facto standard. I think it's fairly obvious that they should be included in fanzines and it does not seem to make sense to have a separate standard for fanzines and magazines. If we are beyond that debate? then we need a standard for entering them that will make it easy to find and group them. Proposed update to pub Help:

Letters should be entered with the following format: Letter (publication name, publication date). Example. If the letter has a title the editor has the option to append it by adding a colon and the text of the title to the letter entry. Example. The editor may add letters selectively. If all of the letters in a publication are added, a notation should be made in the publication notes. In order to group letters on an author's bibliography page they can be added to a series with the following format - Letters: Author Name. Example.

Instead of the series format we might even want to use something as simple as "Letters" although that would get to be a very long series. This is a flexible standard which allows the editor a wide range of options. It also allows a subsequent editor to expand upon existing data.--swfritter 15:12, 12 June 2008 (UTC)

I seem to recall a convention somewhere that such letters should be entered only if the letter writer already had an ISFDB entry for non-letter works. I looked for the wiki page where I read this, but I can't find it. never the less, i propose the following standard for letters:
  1. IN -- Letters published by a magazine or other publication that is IN on other grounds, which are signed by individuals that have an ISFDB author record for works other than letters.
  2. OUT -- Letters signed by individuals that do not have an ISFDB entry for works other than letters.
  3. OUT -- Any letters that are unsigned (except a collective entry for an entire letter column).
  4. OUT -- Any letters published in a non-genre publication, or any publication whose contents would not otherwise be recorded in the ISFDB.
I see no reason to record the individual writers of letters who are not otherwise significant. I could see an argument if a person who was very famous in another field wrote a letter that should be recorded, if say a Nobel-Prize winner or a Head of State had a letter published in an SF magazine, perhaps we should record that fact. But otherwise, known authors and artists only, i think. -DES Talk 15:33, 12 June 2008 (UTC)
I have no objection to your proposal for how letters should be formatted and organized, when included. -DES Talk 15:36, 12 June 2008 (UTC)
What you were reading was another interminable discussion which came to no conclusion. What about fanzines?--swfritter 15:39, 12 June 2008 (UTC)
I would apply the same rule, just an entry for "Letter column" unless an individual letter writer also has records for fiction or art.
In fact, if I had been writing the fanzine rules, I would probably have said that we should only record fanzine contents for people with professionally published items on file. And the same for self-publications. But I guess that decision is already made. -DES Talk 16:32, 12 June 2008 (UTC)
I had been generally supportive of (and using) the de facto standard as articulated by DES and I think it worked well for a while, but lately I have run into a couple of issues with it. First, there are cases when a letter is signed by a person who is currently not in the database but who gets added later on, perhaps when we add some of the more obscure magazines. In cases like that (and I think I have run into them twice now) we have to go back and retroactively add a bunch of letters, which may not be easily identifiable after the fact. Second, what do we do when the letter writer's only Titles are locs ("letters of comment") in prozines and Essays in fanzines? As we continue cataloging fanzines, we are likely to start seeing more and more of these cases, which will make keeping track of who is "in" and who is "out" a major headache.
I am not entirely sure what to do about this messy area, but one thing that comes to mind is that it may be easier to create a new Title Type for letters than to try to keep them organized in Series. Something to ask Al about, perhaps? Ahasuerus 00:02, 13 June 2008 (UTC)
I do see your point -- people who are going to be in, but aren't yet in, are a problem. I would have chosen to reduce this problem by not cataloging anything by anyone who didn't have at least one non-letter publication in a pro or semi-pro zine. But i gather that the consensus is otherwise. I would say, if the person entering the data is reasonably sure that the letter writer will eventually be included for non-letter content, insert. I really think that entering every letter will seriously decrease the utility of contents displays in most cases, and complicate author searches by returning far too many false matches. Sooner than that, I would go with a more restrictive rule: No individual letters except from people with a significant body of professionally published SF art or fiction -- say at least two novels or 10 works of short fiction. Most people who would qualify under that rule are already in the DB, so decisions will be much easier. It would also make letters rare enough that a separate type becomes of little value, i suspect.-DES Talk 01:10, 13 June 2008 (UTC)
I saw a comment on a user's talk page that if any letters are done, all should be. If those are the choices, i would prefer none, including deleting ones already entered. -DES Talk 04:18, 13 June 2008 (UTC)
I think all letters should be included if we start to enter them, since a letter column is like a round table discussion. They've had "flame wars" as long as there have been letter columns, practical jokes, and hoaxes (I can think of a few). Imagine trying to follow the progress of a flame war, when you can't easily find the responses, like listening to half of a debate. You might also miss the pseudonymous letters, which were written by well known authors.
Part of the problem of dealing with letters is that we have no software support for it. If we could enter letters in the same manner as the book reviews, that would compartmentalize them. Especially if the "concise listing" tool was modified to expand or hide specific types of entries.--Rkihara 05:16, 13 June 2008 (UTC)
I think that if we enter all -- or anything close to all -- letters, we will have far too many entries in the Authors table, and someone searching for an actual author based on a partial name will get far too many hits, thus reducing the usefulness of the db significantly. Rather than this, I would much prefer no individual letters at all (and I'd be happy to help delete the ones already entered). If we added separate software support for letters, it would, IMO, have to include a separate letters_authors table, so that people whose only records were for letters were not returned on a basic author search. Also, there would need to be a display option to suppress letter-level detail in pubs. If, and only if, those changes were made, i could support entering all letters. -DES Talk 05:50, 13 June 2008 (UTC)
Frankly the flame wars, jokes, and the like are of no great interest to me, and I think they are of less interest to most users. The only letters that seem to me worth recording are those in which a noted author (or artist) discusses his or her work, or the work of others, or those in which a person who was later to be a noted author makes an early appearance, which have historical interest. I would expect perhaps one individually noted letter a year from most monthly prozines. -DES Talk 05:50, 13 June 2008 (UTC)
I guess I wasn't clear. The flame wars are usually between author and reader, so if you just record the author, half of the dialog is missing. The hoaxes and jokes were usually perpetuated by authors/editors (think of Ben Franklin's "Dogood letters." From my work on Amazing, maybe 2-3 letters a month from pros. Then there are the "pro fans."--Rkihara 06:07, 13 June 2008 (UTC)
P.S., I know of the flame wars, jokes, and hoaxes because they are showing up in histories of SF. They are definitely of historical interest.--06:14, 13 June 2008 (UTC) —The preceding unsigned comment was added by Rkihara (talkcontribs) .
Ah, I had misunderstood a bit. I still think that these would clutter the author list excessively, particularly if you do the later mags where the vast majority of letters are simply comments on recently published stories by readers. (I had one or two of those published myself, in the early IASFM). I would wait until/unless we have software changes to exclude these from basic searches and displays. -DES Talk 06:20, 13 June 2008 (UTC)
I will be on the road for the next 24 hours, so I won't be able to comment at length, but I just wanted to point out that we have always had editors with slightly different ideas about the desired scope of the project. Some editors see the database as a research tool and/or tend to be inclusionists (notably Al) while others are more worried about making the database accessible to the average user. Some of these issues can be addressed programmatically by adding new fields or making the search algorithm smarter, e.g. by splitting the current "Title" search into "Titles (all)" and "Titles (fiction)" so that our fiction-oriented users could easily filter Reviews and Essays. There are other things that we could do, but I have to run. Unless Friday the 13th proves to be particularly unlucky, I should expand on these matters on Saturday :) Ahasuerus 07:03, 13 June 2008 (UTC)
The addition of fanzines to the mix is going to result in a huge number of names being added to the mix no matter what - so that argument is moot. Search and display issues are a secondary consideration - there is substantial room for improvement and the potential for other people to develop alternate methods using the Creative Commons data. Our rules for data acquisition should not be based upon current search and display limitations. Like any other data the editor has no obligation to enter any more data than they want but they should have the option to enter as much as they want. I might also note that only a couple of people wanted Project Gutenberg entries and some were adamantly opposed but they are in there anyway.--swfritter 17:04, 13 June 2008 (UTC)
I'm not too worried about the Project Gutenberg stuff if someone else deals with it, I'm just not going to for now. The Display problems are getting bad though and do need some work soon: I see Translator support is on the list, and I can see a good reason to separate artists from authors and reviewers and interviewers, and possibly from letter-writers in the future: and Editors (the two types entered here rather than working here) need clarification as well: but that's perfectly possible without destroying current data. BLongley 21:46, 13 June 2008 (UTC)

(Unindent) I'm a bit worried about the proposed DELETION of letters, when I've already seen quite a lot of effort to get them included in certain circumstances. I propose a more "live and let live" policy - if somebody wants to add letters, let them, so long as they keep to the usual regularization rules for title and do the pseudonym set-ups. If people want to encourage entry of such for certain types of publication, that's fine too, but every editor should be able to say when that's more than they want to deal with, and the people wanting such entries can go do it themselves. The magazine editors seem quite happy to go into more detail than I would on a magazine, and if me just doing the SF contents doesn't help then they can ask me to leave them alone so they don't get confused by my verifications, or just accept that a verification by me doesn't mean the same as a verification by one of them and more work is required to bring it up to their standards. Similarly, I'm not asking people to find coverart images for books they've verified, or for details of every other price noted on it. I may ASK if the cover I found is right, or ask if the price is the main one or just the additional one added for their country, but there's no requirement to go do this additional work or even answer. But the more we add to the "desired data" list, the more relaxed I think we have to be about whether it's actually required before it's an acceptable entry or not: and when we start talking about DELETING data just because not everybody wants it in the pubs they're interested in, we really ARE going to start putting editors off. BLongley 21:39, 13 June 2008 (UTC)

I guess I'd tend toward the inclusionist end of things; I do think of ISFDB as a research tool. But I'm very much with Bill on the live-&-let-live end of things, too. It's good to suggest to an editor, particularly a new one, that we're interested in including something (artwork, book reviews, whatever, & maybe letters); I'm not ready to say we should be demanding these things, or taking a don't-bother-entering-unless-you'll-enter-it-all approach. Less work for someone else later, as long as what is entered is up to standards. -- Dave (davecat) 22:15, 13 June 2008 (UTC)
Mike did the suggest-but-don't-demand thing very nicely in mentoring me, BTW. -- Dave (davecat) 22:34, 13 June 2008 (UTC)
On a personal practical level: for the magazines I have only been entering letters by authors whose names I recognize and have no intention in the near future of doing any differently. The fanzines are different. It seems almost pointless to me to do them if all the letters are not entered - although that should still be at the editor's discretion. A very large percentage of fanzine letter writers would very likely pass any tests concerning non-letter contributions - but we won't know that until all the fanzines ever published are entered. The chances of anyone in the near future actually entering all the letters from the prozines is actually quite small.--swfritter 19:23, 14 June 2008 (UTC)

(Unindent)From what I can gather the general consensus is that letters should be included although we should not encourage mass entry until there is some software support . I would suggest using the wording above but removing the section about placing them in a series in hopes that software support will be forthcoming at some future date.--swfritter 18:08, 19 June 2008 (UTC)

I have made some very cautious changes to Policy concerning letters. Letters are now "In With Reservations". There is also a section in New Pub Help.--swfritter 17:30, 1 July 2008 (UTC)

Novels that were later split into multiple volumes

One of the major outstanding issues with the current database layout is that we have no real support for novels which first appeared in one (usually long) book, but were later reprinted in multiple volumes. The last time we discussed this issue, we didn't want to retroactively change the original novel to an omnibus since it really wasn't an omnibus. And we didn't want to add parenthetical descriptions to the titles of subsequent reprints because we were afraid that it would break the "lexical match" logic which was used by Reviews at the time. The ultimate solution appeared to be the long requested addition of support for a new field for "relationships" between Titles, but that's easier said than done. Thus, a stalemate.

Well, now that Reviews no longer use the infamous "lexical match" <three cheers>, I wonder if we can revisit the issue. (Granted, Serials still use the lexical match logic, but they are unlikely to be involved when dealing with novels split in multiple books and, besides, they are next on Al's list of things to liberate from "lexical match".) We already use parentheses to disambiguate letters, introductions, abridged versions, etc and I wonder if we can press them into service in this case as well. Here is what Hugh Cook's Chronicles of an Age of Darkness looked like a week ago:

   * Chronicles of an Age of Darkness
         + Wizard War Chronicles: Lords of the Sword (1991) <- should have been listed as #4
         + 1 The Wizards and the Warriors (1986)
              o Variant Title: Wizard War (1986)
         + 2 The Wordsmiths and the Warguild (1987)
         + 2 The Questing Hero (1988)
         + 2 The Hero's Return (1988)
         + 3 The Women and the Warlords (1987)
              o Variant Title: The Oracle (1987)
         + 4 The Walrus and the Warwolf (1988)
         + 5 The Wicked and the Witless (1989)
         + 6 The Wishstone and the Wonderworkers (1990)
         + 7 The Wazir and the Witch (1990)
         + 8 The Werewolf and the Wormlord (1991)
         + 9 The Worshippers and the Way (1992)
         + 10 The Witchlord and the Weaponmaster (1992) 

Here is what it looks like now after some tweaking:

   *  Chronicles of an Age of Darkness
         o 1 The Wizards and the Warriors (1986)
               + Variant Title: Wizard War (1986) 
         o 2 The Wordsmiths and the Warguild (1987)
               + Variant Title: The Questing Hero (first half) (1988)
               + Variant Title: The Hero's Return (second half) (1988) 
         o 3 The Women and the Warlords (1987)
               + Variant Title: The Oracle (1987) 
         o 4 The Walrus and the Warwolf (1988)
               + Variant Title: Lords of the Sword (first third) (1991) 
         o 5 The Wicked and the Witless (1989)
         o 6 The Wishstone and the Wonderworkers (1990)
         o 7 The Wazir and the Witch (1990)
         o 8 The Werewolf and the Wormlord (1991)
         o 9 The Worshippers and the Way (1992)
         o 10 The Witchlord and the Weaponmaster (1992) 

The initial reaction on Usenet, where this question was first raised, was very positive, but, of course, our users are not always in a position to see what's going on in the kitchen. Am I missing any potential problems with this approach or does it look promising? Ahasuerus 23:01, 30 June 2008 (UTC)

I had thought that, up until now, Variant Title were used only when the texts were essentially identical. i agree that the above improves the series page, but on the other hand the shorter titles now disappear completely from the alphabetical display. I would have been inclined to handle this with a sub-series so that The Questing Hero and The Hero's Return were both part of a series "The Wordsmiths and the Warguild", and perhaps a parenthetical qualifier (First half of The Wordsmiths and the Warguild) for the first of these. There are advantages to the above method, but it does complicate getting a list of the different works published. Normally we exclude variant titles from such lists (the alphabetical and chronological bibliographies) because they really aren't separate works, merely different names for the same work. Here that is not true. I think the ultimate solution will be the enhancement discussed in ISFDB:Proposed Design Changes#Based on, but that is no doubt some time away. -DES Talk 15:22, 1 July 2008 (UTC)
Putting my "new user" mode on and searching for "The Questing Hero" finds "The Questing Hero (first half)" alright, but then that makes me think it's the first half of "The Questing Hero" and I'm searching for the whole thing. "The Questing Hero (first half of The Wordsmiths and the Warguild)" would have been clearer. That would obviously clutter up THIS display though. However, THIS display also makes me wonder where the other two-thirds of "Lords of the Sword" are - or is it two-thirds of "The Walrus and the Warwolf" that's missing? The first third is still misplaced in Series display too:
   *  Chronicles of an Age of Darkness
         o Lords of the Sword (first third) (1991) by Hugh Cook
         o 1 The Wizards and the Warriors (1986) by Hugh Cook
         o 2 The Wordsmiths and the Warguild (1987) by Hugh Cook
         o 2 The Questing Hero (first half) (1988) by Hugh Cook
         o 2 The Hero's Return (second half) (1988) by Hugh Cook
         o 3 The Women and the Warlords (1987) by Hugh Cook
         o 4 The Walrus and the Warwolf (1988) by Hugh Cook
         o 5 The Wicked and the Witless (1989) by Hugh Cook
         o 6 The Wishstone and the Wonderworkers (1990) by Hugh Cook
         o 7 The Wazir and the Witch (1990) by Hugh Cook
         o 8 The Werewolf and the Wormlord (1991) by Hugh Cook
         o 9 The Worshippers and the Way (1992) by Hugh Cook
         o 10 The Witchlord and the Weaponmaster (1992) by Hugh Cook 
It's slightly better, but the change of use of variant titles doesn't really help the new user much and needs explaining to those of us that do use it only for "mostly the same text" currently. BLongley 19:06, 1 July 2008 (UTC)

Variant titles: for identical texts only?

Until and unless we have support for a "nature of the relationship" field in a vt link, I am inclined to think that it is a mistake to use a vt to indicate an expanded or revised edition with a significantly different test. And if, as I have suggested, we instead add a "based on" relationship, similar to but distinct from a variant title (see ISFDB:Proposed Design Changes#Based on) all these vts would need to be redone.

For the time being, I suggest that VTs be restricted to situations in which the author or title or both have changed, but there has been no significant textual change, as far as we know.

An example: When I added the record for The Languages of Tolkien's Middle-Earth I very carefully did not mark it as a variant of The Languages of Middle-Earth, because it is a significant expansion. Instead I added a note indicating the relationship. This note would have been ripe for expansion by an editor who actually had physical possession of both editions. Another editor has now made this a variant, and in the process the note has been reduced to the single word "expanded" in the title. One problem with this, it can hid on the author's biblio page, the dates when the expanded or revised version was created, if the vt is given the same date as the original (as the help seems to call for, although it was not done that way in this case. Another problem, in the Alphabetical bibliography page, the expanded version is hidden completely. These problems are all larger in the case of a more prolific author.

Now maybe others want to use vts in this way. If we do, then we need to clarify in the help just how this is to be done, IMO. But I think we need to decide one way or the other. -DES Talk 01:41, 5 July 2008 (UTC)

Unfortunately, this thorny issue has been discussed a few times over the last year or so but no consensus has emerged so far. I was going to post something about in the section immediately above this one, but the Noel issue came up first.
I think we all agree that a new "relationship" field would be the answer to our prayers -- see Feature:90155 Add an optional "nature of the relationship" field to the Make Variant screen for details -- but it's not clear what to do in the meantime. The other problem here is that the very fact of creating a variant title can wipe out Notes and other fields in a Title record, something that I had forgotten about (oops, my bad!).
The main advantage, as I see it, of linking two non-identical titles as "variant titles" and documenting the difference in parentheses is that it shows our users the nature of the relationship on the main page. I think this is hugely important since, based on rec.arts.sf.written feedback, even experienced ISFDB users tend to use the main Author page exclusively and ignore any Title and/or Publication notes entirely when viewing an author's bibliography. And it's hard to blame them when many authors are responsible for dozens or even hundreds of Titles. For example, we already had the nature of the relationship between the different reprint/split editions of Hugh Cook's series document at the title level, but the user who brought up this issue never found that level of detail. When I made the changes that I posted above, he said "Much, *much* better. Makes full sense now", which is what prompted me to post my proposal above.
The main disadvantage of this approach is that it means that we would be no longer recording titles "as stated", at least not at the Title level. I don't think this is a fatal flaw when dealing with novels, where subtitles and series prefixes often make the exact title hard to determine, but it's a bigger issue with short fiction. Also, we used to have the infamous "lexical match" problem to worry about, but with the recent changes to Reviews it's less of an issue. There may be other issues that I haven't thought of, though, so please post all the pluses and minuses here.
BTW, on the plus side, we have only a couple hundred Title record with "(abridged)", "(revised)", etc in the title, so changing them wouldn't be too hard. Ahasuerus 02:09, 5 July 2008 (UTC)
I see your point. But don't forget that it can hide detail even at the author display level, if a parenthetical modifier is not used, and that it always hides data on the author's alphabetical display. I still think that "based on" is the best long-term solution, but who knows when such a thing might be implemented, even if everyone agrees. -DES Talk 02:19, 5 July 2008 (UTC)
If we are going to do this, though, we do need to change the help, IMO. -DES Talk 02:19, 5 July 2008 (UTC)
As far as Alphabetical and Chronological pages go, their display logic lags behind the logic used by the Summary page. About two years ago Al made a conscious decision to concentrate on the Summary page and then, once it was stable, go back and upgrade the logic used by all other pages -- including, notably, the Series page logic -- to work the same way. I don't think it has happened yet, but perhaps now that the Summary page logic is reasonably stable, we could review the behavior of all other pages and get them to behave the way we want.
Re: updating Help, perhaps it will be easier to get a consensus this time around because the "lexical match" issue is not as prominent as it once was. Ahasuerus 17:56, 5 July 2008 (UTC)

NONFICTION about Manga/Anime

Lorenzr has been very active today, and as usual we should have a lot fewer comics and non-SF manga after his edits: but he's on the borderline in some cases, IMO, and I'm particularly concerned that we may be crossing the line from "Manga and comics must go" to "Anything about Manga and Comics must go". Some of the proposed-for-deletion pubs and titles are actually reference works: some by people we'd never hear of if it wasn't for Manga, but some are by or about Authors we do have, in sufficient quantity or quality that we might invoke either of these rules:

  • In - Works about speculative fiction published in the English language and their foreign language translations.
  • In - Works (both fiction and non-fiction) that are not related to speculative fiction, but were produced by authors who have otherwise published works either of or about speculative fiction over a certain threshold

I personally have nothing against SF Manga so don't delete it, but don't add it either: NONFICTION about it I tend to favour inclusion if we'd include the author as a Critic or Bibliographer anyway. I DO delete non-SF Manga on sight. However, we're getting to the point were we're being asked to delete NONFICTION by Fred Patten, and NONFICTION that includes essays and interviews with Otto Binder, Peter David and others well-represented here. Please have a look at the Submissions I've held and the discussions I've started on Ray's page. I may just be being over-fussy tonight (it's been a bad week) but some of these deletions are ones I'd actually be tempted to expand on instead. BLongley 20:17, 18 July 2008 (UTC)

I tend to agree with Bill here. Works by or containing interviews with well-known SF authors such as Otto Binder and Peter David are clearly IN, and should not be deleted, IMO. As to whether books about Manga, even SF Manga, are works about SF, when we don't so classify the manga itself, that is harder. IMO if it only discusses manga and how to write it, it is out as long as manga itself is out. But if it compares Manga to more traditional written SF, or is in some way also about written SF, then it should be IN, or that is my view. -DES Talk 22:45, 18 July 2008 (UTC)
I'm okay with deleting almost all of the comics/manga/anime, even the SF stuff, although I read and collect a lot these. There's just too much of this stuff to include in the ISFDB. Anyone who's been through a manga store in Tokyo would probably agree. There is probably more manga published every month, than the total of SF magazines in the whole history of genre. I think the only exceptions are those that are linked to reviews, and major influences such as Neil Gaiman. Non-fiction works about manga are okay.--Rkihara 23:00, 18 July 2008 (UTC)
I agree with the above. Just too much for here. Ray 17:41, 26 August 2008 (UTC)
Since Non-fiction by "reasonably prolific" SF people -- which we don't define very well, but Patten, Binder and Davis clearly belong to this very exclusive club -- is in, I don't think we should be deleting their books about manga, comics and other dubious things. Ahasuerus 01:08, 19 July 2008 (UTC)
Yes does prolific mean x number of entries. or x number of published significant worked entries. Some authors are better know outside of SF. Example Carl Sagan. So which side of line for prolific? Ray 17:41, 26 August 2008 (UTC)
Yes the interviews are in. But reviews and materials not relating to SF even from prolific authors is a bit much. Watching Anime, Reading Manga : 25 Years of Essays and Reviews by Fred Patten. Yes a prolific reviewer but about material not in the DB. (Oh he is wrote about Furry too. Not exactly liked in the anime/manga world. Sorry had to laugh). This is just too much material to cover to do a good job of. Isaac Asimov in worldcat has about 400+ nonfiction English book entires while the DB has about 86. I think the ones included related the areas of SF than the missing ones. My point is it isn't something that is done well compared the are of specialization of SF and it is hard enough to do just that. (Scope creep) Ray 17:41, 26 August 2008 (UTC)
I'm not too worried about scope creep. While I will probably not start a project to get the other 300+ NF Asimov works in (and AI surely don't think we need every publication of each) I might do bunches from time to time, and i definitely wouldn't object to anyone else putting them in. If Patten is considered "significant enough" or "prolific enough" (and I agree those terms could perhaps be defined a bit better) for any NF or non-genre works of his to be in, then IMO there is no bar to any of them. No one is required to enter any such items, but IMO there is no ground to object to anyone else doing so. -DES Talk 20:19, 26 August 2008 (UTC)
If somebody wanted to add CONTENTS to "Watching Anime, Reading Manga : 25 Years of Essays and Reviews" I'd probably remind them that a NONFICTION work isn't really intended to have contents: we can cope with a few essays already present here, and if those are reprints from something SF already it might be a nice "here we draw the line" moment where we include those and nothing else. But there's no point adding any known non-SF content. One title isn't a botheration and we can live with it for "other works". If somebody adds the missing Asimovs, I'm not too bothered either. They're often so low on my radar that I've actually turned down copies of such for free as they're not of interest - if it reprints something from an SF magazine I would probably have accepted, recorded THAT BIT, and discarded such. This sort of borderline doesn't really inspire me to delete or support though. If someone wants to add such, then they'll need some sort of Mod approval. If they want to delete, similar. But I'm just going to ignore Furry-content additions and complete book-deletions. BLongley 21:47, 26 August 2008 (UTC)
My point with Asimov is even for a top 50 author the DB is missing a large set of non fiction books. But it looks like not a priority to add. Without looking I would expect (trust?) the speculative fiction works entries are complete, in depth, and accurate. (Almost would need one of the corporate mission statements or goals) Ray 17:42, 27 August 2008 (UTC)
Here I disagree a bit. I recently purchased a copy of Le Guin's Dancing at the Edge of the World, and went to enter its contents, only to find that someone else had already done so. Now this is a collection of essays (one might arguably be called a story, but only one, and only arguably) and reviews by a noted SF author. A few of them (4 essays out of 33, and 5 reviews out of 12, by my count) are in some degree about SF, and a few more are bout writing in general. Others are personal history, or anecdotes, or essays about philosophical, or ethical, or political issues. Still, I had planned to enter all of them. Now I don't say that it is essential that the ISFDB contain all of these, but I see no good reason why it shouldn't. The case for the contents of "Watching Anime, Reading Manga" is probably weaker. But I probably wouldn't reject such an edit or approve the deletion of such contents with an "out-of-scope" reason unless an argument was made that the author was not "significant" enough to have any non-fiction or non-genre works recorded at all. Once an author's non-genre works are IN, I think they should all be at least potentially in (with the exception, already in the RoA, of works that never had a book-form publication). -DES Talk 22:15, 26 August 2008 (UTC)

Other non-fiction works

Here is another instance that turns on some of the same points as the one above. Some time ago, Lorenzr submitted a delete for Exploring Psi in the Ganzfeld by Carl Sargent on the grounds that it was not SF. (His deletion not actually said "medical reference journal???"). I felt that this might be a work of non-fiction properly included under RoA #8, and queried the deletion, and put the edit on hold. Recently i have been trying to clear up and edits i had held, and came to this one. I posted again on his user talk page (see User talk:Lorenzr#Exploring Psi in the Ganzfeld for the full exchange).

His response contained the statement "I don't think the DB should include unrelated works non fiction from the author not about the author or other fiction works. Since it isn't actively maintained.". I disagree with this in principle. However, in this case i am not at all sure that the author is above the "threshold" that RoA #8 mentions (but does not define). I would appreciate comments on whether this particular author is in fact over the "threshold" or not, and perhaps whether we can define the "threshold" a little better. (If this work is not to be deleted, it needs to be corrected from NOVEL to NONFICTION. But that is a trivial matter.) -DES Talk 22:48, 26 August 2008 (UTC)

As RoI #8 states, the original idea behind the "certain threshold" standard was to include the non-genre/non-fiction stuff written by "genre writers" -- something that many of them had requested -- without including everything written by prolific (and often reprinted) non-genre writers like Robert Louis Stevenson, Rudyard Kipling and Jack London whose oeuvre also included some speculative fiction. Another consideration -- which was still nebulous at the time, so it didn't make it in the RoI -- was to clarify the genre status of certain books. For example, suppose that Author X published 3 books in his lifetime and 1 of them was SF while two were non-genre or non-fiction. In a case like that, it may make sense to list the non-SF books so that when our users find them in another catalog or database, they don't have to scratch their heads and wonder whether the other 2 books are non-SF or whether we haven't entered them yet.
Al and I discussed this issue in May 2006, but we couldn't come up with an objective standard, so we left the "threshold" undefined. I guess now that we have many more editors, we need more specific guidelines or else we will have one group of editors diligently deleting the stuff that another group of editors is adding.
After briefly playing with different scenarios, the following approach seems to cover the most common permutations: "An author's non-genre and non-fiction book length Titles are included if at least one of the following three conditions applies:
  1. Over 50% of the author's novel length output has been speculative in nature [the default scenario]
  2. There are 5 or fewer non-SF Titles [see the "nebulous" case above]
  3. The author has published at least 10 book length works of speculative fiction [eliminates E. Phillips Oppenheim and other insanely prolific non-genre writers with a few borderline SF books]
  4. The author is a prominent writer of speculative fiction as measured by the number and caliber of award nominations [needed to preserve, e.g., John Myers Myers's non-fiction]"
Naturally, this is just a rough draft. Thinking some more about it, this approach may not work too well for some living writers who may be in and out of the gray area throughout their career. Ahasuerus 00:08, 27 August 2008 (UTC)
5 (Proposed) The non-Fiction that is inherently about the same topics as hard science fiction (Or it was at the time of publication, or it is Non-Fiction that has become Science Fiction). Examples include Zubrin's Non-Fiction, NonSF, Gerard K. Oneil's work. NASA-SP-413 Space Settlements, NASA-SP-428 Space Resources and SPace Settlements, and NASA-SP-509 Space Resources (I planned to submit these three in the near future - Great Hard SF Non Fiction from 1978 through 1992), Lewis's 'Mining the Sky' etc etc. All of these are great reading, but Non-fiction, not-'about'-Speculative Fiction, but of interest to SF fans, and... as they age... equivalent in many ways to SF that has been passed by in history. All of these authors are more well known for non-fiction o are non-fiction only, though Zubrin at least wrote some fiction for getting a toe in.Kevin 05:32, 27 August 2008 (UTC)
I could approve Ahasuerus's draft above, or something like it. The precise numbers could be debated, but I don't care greatly one way or the other. As for Kevin's proposal, If I have understood it correctly, i don't like it. This would mean that any good science popularization, at least about certain topics, would be IN. If we actually tried to cover all of that, it would be a major distraction at best. Nor do i see that it would serve any of the goals of the ISFDB as I understand it. Perhaps i have misunderstood this proposal. -DES Talk 15:14, 27 August 2008 (UTC)
I like number 2. I was worried about old works not being included. I could agree to any version items 1-4. Not sure about how item 1 applies to the prolific count. Does it mean a novel work is considered to be speculative if it is over 50% for counting in items 2 or 3? I brought up speculative works above due to the reviews, essays, etc. The Watching Anime Reading Manga book above was not from speculative fiction author and contain reviews for material not included in the DB. Editors I think should be included in the authors group. I guess the question is does this apply to which certain people types? (authors, editors, artist, reviews, interviewers) Ray 17:28, 27 August 2008 (UTC)
I thinmk the idea was to have all 4 together, and any author to whom any one (or more) of the 4 applied would have his or her non-SF and non-fiction works IN (all of them). Number 1 means to me that if an author wrote, say 14 novels, and at least 8 of them are SF, then that author's non-SF works are IN. Judging whether a work is SF or not is a different question, but [[ISFDB:Policy#[edit] Definitions]] already covers that, fairly well I think. -DES Talk 19:12, 27 August 2008 (UTC)
I think I don't want to make a definitive rule for Inclusion. If it's not SF but somebody wants it in, and someone else agrees enough to approve the addition, that's fine by me so long as it doesn't mess up the SF works - i.e. NONFICTION or NONGENRE is clearly marked as early as possible. So even if we bypass the "TWO people want it" guide by having a Mod enter and approve something only THEY want, it shouldn't mess up the rest of us. We may need to tell Dissembler to stop adding more of the same though. And nobody should feel obligated to add NONGENRE or NONFICTION stuff just because we have some of it already. BLongley 22:01, 27 August 2008 (UTC)
For Deletion cases, again I'd resist deletion of stuff by an obviously-proven SF author/critic/artist, even if it's only ONE proven case. Moving the non-SF to NONFICTION or NONGENRE should keep the database usable. Most of the data on Amazon is crap, but if you stick to the indexed and organised works you can still buy a book. We're a bit more specialised and you should be able to find an edition or even a printing here that you couldn't on Amazon, but we've not become unusable just because we've got some extra records that wouldn't be in a perfect database. ("ISBN" "999" prefix records for instance?) BLongley 22:01, 27 August 2008 (UTC)
I've seen an author's works get decimated (Clive Cussler, John Grisham) and some grow to total unmanageable status Robert Louis Stevenson but in neither case do I feel it worthwhile expending any major effort. The Non-SF usually gets sidelined easily enough, and I'd recommend we just continually sideline dodgy stuff for anyone that is IN, deletion is a major pain and really doesn't benefit anybody that is coming here for SF as they won't see it. I'd personally like Jane Austen deleted from the ISFDB entirely (long story, I'll spare you it): but if we have to have her here, then it shouldn't take more than a couple of people actively working on her to keep her works out of my sight. She's been reviewed and anthologised and that gets her in, but being in doesn't mean it's got to annoy people. BLongley 22:01, 27 August 2008 (UTC)

Del Rey Discovery - Imprint or Publisher Series?

I just verified McLendon's Syndrome with the existing publisher "Ballantine Del Rey" as it seems someone had put it into the "Del Rey Discoveries" series already, and internally it is definitely a Del Rey book published by Ballantine. However, when I thought I'd check the other books in the series, there aren't any, and the first I thought of adding, Demon Drums turned out to be in another series already, so ISFDB series won't work. I've dealt with multiple series occasionally by creating lists on the publisher page, but Ballantine Del Rey is a publisher so big we could end up with a massive page with lots of lists of "Publisher Series" on. So I'm tempted to make "Del Rey Discovery" an imprint of its own, obviously a child imprint of Ballantine or Del Rey or both - and I see we already have Ballantine Del Rey Discovery. So: does anyone object to this plan, and if not what is the preferred imprint? BLongley 20:12, 22 July 2008 (UTC)

Structurally, it's just another "publication series" just like "Ace SF Special" and the rest of the lot. I don't think there is any harm in making it (or other publication series) a separate imprint, but I wonder how many editors will remember/know to use it when entering other Discovery publications instead of just using the publisher's name? Ahasuerus 20:22, 22 July 2008 (UTC)
First pass done. The "Del Rey Discovery" name seems to be the preferred option on the web, and at least matches what the logo says. Whether people will use it - I don't know. But it's an obscure enough set that I haven't had to mess with any verified pubs but mine so far. We probably ought to update help to at least suggest people don't create new publishers/imprints where an existing one will do, I don't think we're in a position to point people at "definitive" publishers yet by a long way. BLongley 19:43, 23 July 2008 (UTC)
Ace SF Specials used to have their own problem with the [1] and [2] suffixes which meant you couldn't look them up, but I see that's sorted now. Imprints still have their own problems - e.g. I can't search for all Ace books published in 1976 and get the Ace SF Specials included. And the Ace SF Special [1] list looks FAR shorter than I remember it - so the Publisher Regularizers are probably working against each other. :-( BLongley 21:01, 22 July 2008 (UTC)
I just had a look at both Ace Special [1} & [2} and they look fine. I did some work with these about six months ago when I bought a large run of them and the list has not shrunk.Kraang 01:37, 23 July 2008 (UTC)
OK: it's probably a bit longer than that since I last looked at them, so I guess I was looking at the untidied version. BLongley 18:01, 23 July 2008 (UTC)

Project Moonbase and Others: The Scripts of Robert A. Heinlein, Volume One

OK, I didn't put it on the front page, but as it was there I thought I'd add the contents (got to make a good first impression, I think). But it's a book of scripts! I've qualified all the titles so they don't get merged blindly, but maybe this is a book we shouldn't be pushing so much? BLongley 20:50, 25 July 2008 (UTC)

Given the author, I think things are justified here that might not be in most cases. There are few better known authors of SF, after all. And Destination Moon in particular, was very influential. -DES Talk 21:06, 25 July 2008 (UTC)
True - Destination Moon is a classic even in film form, and may even be in future volumes of scripts. I'm not advocating deleting the book, but the contents are going to cause several "is it the same work?" discussions which are probably better held behind closed doors. Of which we have none, of course. (This conversation will no doubt be Google-able in hours at most.) BLongley 22:57, 25 July 2008 (UTC)
I am not sure how our current algorithm determines what goes on the front page every week. Is it fully automated or does Al have to do something to pick the lucky winners? Ahasuerus 21:35, 25 July 2008 (UTC)
I think you'll have to ask Al. Things like "author_annualviews - This column holds the number of times this author's bibliography has been viewed in the current calendar year" makes me think there's some data-based logic behind it all, I just go clear up our most-visible pubs occasionally when I step back and put my "newbie" hat on.
I'd encourage ALL editors to step back occasionally and see what the home-page offers, what the simple searches deliver, etc. Frankly, I have to explain FAR too much to people I introduce to ISFDB - why on earth (or planet of your choice) do we offer COVERART and INTERIORART results for a TITLE search? Who uses YEAR searches? TAG should surely be advanced search only? I'm sure we're improving our data all the time, but we're not doing well on making this place more USEFUL to anyone but us. (I used to think MAGAZINE searches were seriously broken ((and I still think they are)), but frankly we're making BOOK searches a pain too.) BLongley 22:57, 25 July 2008 (UTC)

tp for title page?

I just approved SHDWF1988 where the submitter put an illustration on "Title Page." As expected this was truncated to "Title Pa" but it made me realize there's no code listed on Template:PubContentFields:Page for the title page. It seems that some of ISFDB's codes are already unique such as fc for "front cover" and so I'm wondering it "tp" for "title page" makes sense. Marc Kupper (talk) 22:20, 26 July 2008 (UTC)

Fine with me, although the title page is logically part of "bp". -DES Talk 00:32, 27 July 2008 (UTC)
I'd agree that the title page is nearly always part of bp but in this case someone wanted to refer to an illustration that was on the title page. I'm hoping someone could cite bibliographic standards rather than us inventing a new standard such as "tp." For example, rather than a code such as "tp" I've seen people indicate unnumbered pages with either parentheses or brackets such as (3) or [1-4]5-180. Sometimes when you count backwards you'll find the page numbering goes negative. I have a reprint where it looks like the publisher added a four page preface but also kept the original pagination for the story. The resulting construction could be said to start on page minus 3 with the preface running from page minus 1 to unnumbered page 2. I could convert the pages before unnumbered Arab number 1 into Roman numbers meaning it's [i-iv,1-3]4-860 with the preface spanning (iii) to (2), and the story starts on (3), and the first paginated page is 4... Marc Kupper (talk) 19:39, 3 August 2008 (UTC)

First publication date

It's been my understanding that we record the first book publication date. Template:TitleFields:Date is vague on this point. I'm entering/verifying a collection where a story first appeared in Analog, May 1973 and this year had its first book publication. Choices

  • Do nothing - we have one title record dated 1973
  • Move the existing title a type SERIAL record dated 1973 and add a new title dated 2008-03-00.

Either way, the rules should get clarified. Marc Kupper (talk) 23:14, 26 July 2008 (UTC)

The template you cite (which, I think, isn't actually part of the help pages) appears to be quite clear: "Date - The original publication date of this title", regardless of the source, either magazine or book. It doesn't state that the first book appearance should be considered the title's date of publication. If that were true, most of the ISFDB shortfiction titles would have the wrong date. The work you refer to should have a publication date of May 1973. MHHutchins 00:30, 27 July 2008 (UTC)
Another thing: that page (Template:TitleFields:Date) appears to be relatively recent (October 2007). I'm not very familiar with templates, but I thought they're used to make short-cuts in the creation of other Wiki pages. If so, I don't see the purpose of this particular page. Any important information on it (especially the part about magazine serializations) should be part of the help pages. IMHO. MHHutchins 00:37, 27 July 2008 (UTC)
(after edit conflict) The template is actually used as part of the following help pages:
  • Help:Screen:NewPub
  • Help:Screen:EditPub
  • Help:Screen:AddVariant
  • Help:Screen:EditTitle
  • Help:Screen:MakeVariant
This can be seen with the 'what links here" feature from the template. It does say "record the first book publication date" for works later published as books. Perhaps it is not quite clear that this does not apply to shortfiction published in magazines. I will add a line to clarify that. -DES Talk 00:40, 27 July 2008 (UTC)
You're right, it only refers to serials. This doesn't apply to the situation that Marc is referring to. I can see how a new person might be confused, but Marc, buddy, you've been around the block more than a few times. :) MHHutchins 01:34, 27 July 2008 (UTC)
Templates can be and are used for many different purposes, not just shortcuts. One common use is when there is a piece of test common to multiple pages. By putting that text into a template, there is only one copy, and changing that changes all the pages where it is used together. We have quite a number of templates used in this way. -DES Talk 00:42, 27 July 2008 (UTC)
Thanks for the clarification about templates. I see now how that particular section has been duplicated in the pages that link to it. What you might call a "shortcut". :) MHHutchins 01:34, 27 July 2008 (UTC)
I suppose. Anyway, templates can be used for that, or for standardized links (like Template:A) or for bolierplate with parameters (like Template:Unsigned or more complexly, Template:Cover Image Data) or for standard images (like Template:GoLive-ToDo), or in lots of other ways. Most of the complex "InfoBoxes" and standard citation formats on Wikipedia, for example, are implementd via templates. -DES Talk 01:52, 27 July 2008 (UTC)
FWIW - I called it part of the help as I went to edit-title, clicked Help:Screen:EditTitle, and as the template contained the specific bit of text I wanted to comment on I linked to it. It's not really a "shortcut" but rather transclusion to toss around the eleven dollar words. :-) Marc Kupper (talk) 04:08, 27 July 2008 (UTC)

(unindent) Date - The original publication date of this title is unclear as many bibliographers define "publication" as published in books. Magazines are ephemeral and publishing in them apparently does not count when it comes to defining when something was "published." Aren't you supposed to throw away a magazine the instant you receive the "current" issue?

On ISFDB we have been lax about this and there are many title records where the date is the magazine date and it's possible there is an informal consensus that publication in magazine counts as "published." If instead we decide the date really should be the first book publication date then an idea that just came to me would be to use a single title record for the magazine and book publications but to date it the date of the first book appearance. Anyway, I'd personally prefer that the title date for a story be the first date regardless on if it's in a magazine or book form. Serialized novels though would use the first publication date of that novel as a whole. The reason I brought this up though is that I believe there is a rule out there about magazines though of course I can't find a reference that supports this... Marc Kupper (talk) 04:08, 27 July 2008 (UTC)

In response to the earlier discussion here I added the following text to the template mentioned above, and so to the various help pages that transclude it: "Shortfiction: Works of short fiction are always given the date of first publication, whether that is in an anthology, collection, magazine, or other form. Only if a later version is significantly changed should it have a different title record with a different date." Marc promptly reverted this change. I am not going to edit-war about this, but I th9ink that the above states what the proper practice should be, and what I have always understood the IFSDB practice to have been. It sounds from the above note as if Marc actually agrees with this rule ("I'd personally prefer that the title date for a story be the first date regardless on if it's in a magazine or book form.") But perhaps i have misunderstood him. Is there anyone who actually disagrees with my text above? If not, perhaps it can be restored? -DES Talk 04:46, 27 July 2008 (UTC)
As Michael pointed out above, we have always used the first publication date -- magazine or otherwise -- to date Shortfiction records, which is what most genre bibliographers do. Serialized Novels are treated differently as per "Serial installments of a work are always given the date of the magazine in which they appear even if the work has been published previously in book or serial form", although now that I am thinking about it, there have been a few serialized Novellas too and I don't think we have decided yet whether they should follow the Serial rules or the Shortfiction rules for dates. Ahasuerus 06:00, 27 July 2008 (UTC)
Serialized Novellas - I think the same logic applies since, if I remember correctly, one of the reasons for treating serials separately is that the texts are quite commonly substantially different in subsequent publications.--swfritter 17:44, 27 July 2008 (UTC)
For a novel, we consider any magazine appearance to be "serialized" even when it is contained withing a single issue. Are you proposing that any novella published in a magazine should have two title records? Or only if a novella is actually published in multiple parts (as is sometimes, but not often done) I understand the statement that novels are often different for serial publication is in many, perhaps most cases due to the need to cut for serial publication. I presume that no such need applies to most novellas -- i have certainly never heard such a thing from anyone who ought to know. Certainly in relatively recent times, most novellas where i have read both the magazine version and a version later printed in a collection or anthology are the same text. Indeed the only exceptions I can think of were accompanied by sufficiently drastic expansion that we ought to list them as "revised version" anyway. Is there any source for the suggestion that most or even many are revised at this stage? In the absence of such a source, i would not favor a separate title record, nor would i ignore the magazine publication date for novellas. (Novellas later published as stand-alone books even if less that the "novel" word count may be a special case, particularly in the 1940s-70s.) -DES Talk 20:27, 27 July 2008 (UTC)
I am not proposing any change at all, am in fact suggesting that the current standard is acceptable. When a novella is published in multi-part serial format it is lexically matched to the novella title. I certainly am not suggesting that a novella that appears in one issue of a magazine be categorized as a "Serial" and appended with "(Complete Novella)".--swfritter 00:32, 28 July 2008 (UTC)
I thought that sounded very odd indeed -- I am glad to learn that i misunderstood you. I agree that the current standard does call for using a SERIAL record when a work is actually published in multiple parts, no matter how long or short the work is. I think it was the phrase "Serialized Novellas - I think the same logic applies" that lead me to misunderstand you. -DES Talk 01:45, 28 July 2008 (UTC)
I might also note that another reason for our current handling of serials is to ensure that the book version of a novella is the primary version - that would, of course, not be the case if all subsequent printings of a novella printed as a multi-part serial are printed in collections or anthologies but not in book (or chapbook) form. In any case, multi-part serial novellas are quite rare.--swfritter 17:20, 28 July 2008 (UTC)
I assume that the idea that the "book version" of a novella is primary apples only when 1) the novella was published in multiple parts, and 2) the "book" publication is between separate covers, or perhaps in an Ace double. Unless the principle extends to novellas that were not published in multiple parts, it would, as you say above, be quite rare. In the case of a novella first published in a single part in a magazine, and later reprinted in collections or anthologies (which i think is the modal case) no assumption as to which is "primary" need be made, and I correct? And even in the case of a novella published in a single part and later reprinted as a "short novel" or in an Ace double or the like, is such an assumption really justified? I understand that in many cases such a work was re-edited for the Ace publication with little or no input from the author, and the magazine version is more likely the intended text. (I base this in particular on the woks of Jack Vance, where the VIE people have gone to huge effort to compare the various texts, and have indicated the results in some detail. How typical the results are, for other authors, i can't say, but at least it raises the issue.) So do we really want to automatically regard the "book version" as "primary" in all such cases? -DES Talk 17:50, 28 July 2008 (UTC)
I'm not too worried which is "primary", the issue is mainly the ability to record both (all) dates and show the connection. Having said that, has anyone found an example yet? I suspect it's such a minor problem that notes will cover it easily. (And is preferable to workarounds like including all the serial records in book publications.)BLongley 18:05, 28 July 2008 (UTC)
Very rare and usually in minor (often semi-pro) magazines. An infamous example is "The Titan" by P. Schuyler Miller. It first appeared in the semi-pro "Marvel Tales" in 1934. That magazine failed before the last part was published and the manuscript was lost which meant that Miller actually had to re-create it when it was published in a collection in 1951. It's a minor issue right now but I don't know how Al is going to deal with this when serials are db linked rather than lexically linked. I presume they will be linked by title id which should also work.--swfritter 18:34, 28 July 2008 (UTC)
What had me concerned is there there seemed to be a suggestion that novellas (and perhaps other short fiction) would have as the date of their title records the date of their first book publications, even if such works had been published in a magazine in the normal single-part manner. I am now not at all sure if anyone really intended to suggest this, but I thought it was being suggested. I agree that shortfiction that is actually published in multiple parts is sufficiently rare that we don't need to change the db design for such a case, a note will do for such odd situations, just as a note was the proper solution for the "split intro" by John D. Clark (or was it Pohl:)) for Uller Uprising. -DES Talk 21:07, 28 July 2008 (UTC)
I think the "published in parts" as opposed to "published as a whole" deserve different dates, and we more or less automatically get that for novels now we've clarified serial dates. For novellas - I can't think of an example off-hand, but I'd like to see some before offering a view. It does strike me that "The Trouble with Tycho" issue is relevant again here, as if we DO get the single text all nicely merged we'll only have a title date for the original magazine publication and people would have to derive the first BOOK date from publication lists alone. I'm not sure how people would feel about "The Trouble with Tycho (Complete Novel)" and "The Trouble with Tycho" as separate titles. BLongley 22:08, 27 July 2008 (UTC)
I would strongly oppose having any such thing. I can, just barely, see that if such a work is actually published in parts, there needs to be special handling to link the parts and record a separate date for each part, and that therefore it can be argued that that none of these publication dates is the "real" date. But for the ordinary novella, which is published in a magazine in a single issue (most genre magazines include one or more per issue), and is later reprinted in an anthology, or a collection, or both, with no significant changes (as I think is by far the more common case when a novella is reprinted) I see no valid reason to have more than one title record, and very little reason to use any publication date but the first. -DES Talk 23:47, 27 July 2008 (UTC)

(unindent) I have actually found an example of a serialized novella: ...And Now You Don't by Asimov, the story that became the second half of Second Foundation. Apparently it was never published separately aside from the original magazine publication, and it must be near the boundry of novel length. Surely not a common situation. -DES Talk 15:42, 30 July 2008 (UTC)

I've run across serialized novelettes and short stories while working on magazines (Amazing?). Not sure what the point of doing so was. I didn't note them, so I don't remember which stories they are, but one which was serialized in three parts totaled about fifteen pages.--Rkihara 17:47, 30 July 2008 (UTC)
Good find! And now I see it, I have no qualms about letting the magazine date stand and the first publication date in a book go un-noted. But I'm a weirdo that doesn't really care about owning first edition books, and am happy to listen to anyone that actually buys books for first edition SHORTFICTION. (I've bought books for the "ONLY" SHORTFICTION publication in a book though, but once it's anthologised/collected once it's not normally long before it comes round again.) -BLongley 19:52, 30 July 2008 (UTC)
I'm back for a moment (end of month meaning I'll be tied up for a few days). I'd need to dig through the help but believed we already had it that if a story was serialized into two or more issues that we would use separate SERIAL records and if that story is ever printed as a whole (regardless on if it's as a novel or shortfiction) that we'd have another title record dated the first time the story appears in whole. IIRC, the example in the help mentions a story where part 1 and 2 were printed ~years~ apart and I have a story where part 1 was in a magazine (and it's title says it's part 1) and years later part 2 was printed in an anthology. I'm fine with how this works as far as dating ISFDB records goes.
It looks like there's a consensus where if a story is printed in a magazine as a "Complete Novel" that we will have one title record dated using the magazine's date. I suspect I can unrevert my last reversion but have one question. There are 414 title records of type SERIAL that also contain "(Complete Novel)" in the title. Why do these exist rather than calling these stories SHORTFICTION? I'm assuming each of these was a magazine splashing "Complete Novel!!!" on its cover in relation to the story to indicate they are novella or larger works. Marc Kupper (talk) 05:38, 1 August 2008 (UTC)
As to your first point "if a story was serialized into two or more issues that we would use separate SERIAL records and if that story is ever printed as a whole (regardless on if it's as a novel or shortfiction) that we'd have another title record dated the first time the story appears in whole." The current help could be read as saying that we do that only if the work is a novel, but I can agree with doing that for any story that is printed in multiple parts, whatever its length. if we agree on that, we should make that explicit in the help.
As to SERIAL records with "(Complete Novel)" appended to the title, it was my understanding that these should only be used when a work is in fact of novel length (40k words or more). I think some people have used them if the work was later published in separate covers or as part of an Ace double, even if under 40k words. I don't think any statements on the magazine's cover have been or ought to be a determining factor, although the term is doubtless derived from such magazine cover statements. Again, we should come to a clear agreement on when such records are to be used, and make that agreement very explicit in the help, IMO. -DES Talk 15:21, 1 August 2008 (UTC)
The 40k limit applies - such creatures have been very rare since the 50's and most occur in the large-sized pulps. The magazine definitions of story length are notoriously inconsistent. Also, if I remember correctly a lot of thought went into using the book publication as the first date - this is based upon common bibliographic standards in the wider world.--swfritter 16:59, 1 August 2008 (UTC)
Before any significant changes to Help are made I would suggest starting a new section and summarizing the changes with very precise, condensed, and balanced arguments so a hands-up, hands-down vote can be taken. Those involved in discussions should perhaps consider themselves a sub-committee who are bringing a report to the main body.--swfritter 16:59, 1 August 2008 (UTC)
Swfritter, you wrote this is based upon common bibliographic standards in the wider world. Can you cite these standards? I've looked and was unable to find any standards nor anything that's for or against how works should be dated. Obviously ISFDB can publish its own standards and to change them as we learn more but it would be great if the ISFDB standards were at least in alignment with common bibliographic standards in the wider world. For example, I don't know of the standards related to this but we used to date magazines, and their articles/stories, using just the year as the physical date of release often times had only a casual relationship to the issue date printed on the magazine. More recently the "ISFDB standard" seems to be to use the stated date unless it can be shown it is wrong and that we use the publication notes to document things. Marc Kupper (talk) 18:26, 3 August 2008 (UTC)

Chapbooks again

I've played about with some Chapbooks today, and it seems we CAN get them to display the way I want, e.g. here but the first edit to it will destroy it as the software demands that the CHAPTERBOOK entry must change. It's also impossible to create variant titles for things like this as you can't create a CHAPTERBOOK variant. I think we've spotted a hole in the ISFDB design as we can (with difficulty) create the right display (leaving my issues with title-type descriptions aside) and have the BOOKS appear above everything else, but the ban on changing or adding CHAPTERBOOK entries leads to my frustrations. If we can have CHAPTERBOOK variants and Content additions I can live with the fact that they're misnamed, so long as the books display like that. (We can discuss SHORTNOVEL or such names later.) I think CHAPTERBOOK is now in the same category as EDITOR records - we need them for reasons that it isn't easy to explain, or justify. But people can make EDITOR records "work" to some extent, whereas CHAPTERBOOK ones are a one-shot "do it right and hope nobody ever wants to change it" case at the moment. BLongley 22:08, 27 July 2008 (UTC)

It looks to me as if you have created a title record of type CHAPTERBOOK, which i would have expected to be impossible, and now that I know it is possible, I think this is a bug. IMO such a type ought to be used for publication records only (whichever set of publications we decide it should be used for). Specifically, I don't think this solution in any way helps if a work has both been published separately and as part of a collection and/or an anthology. I suspect this will wind up with having two distinct title records for the same work, which is the one thing I wish to avoid at all costs. I think we are left exactly where we were on the issue of chapterbooks, novellas, short novels, and the like. I think you've spotted a hole that should be plugged -- it should be impossible to create title records of type CHAPTERBOOK by any method at all. -DES Talk 23:39, 27 July 2008 (UTC)
At one point (some 2 years ago, I believe) ISFDB supported a separate CHAPTERBOOK Title type along with a CHAPTERBOOK Publication type for the very reasons that Bill outlined above. After thinking about it, we decided that this Title type would cause various problems. For example, any short fiction Title that has been published in a Chap(ter)book and also (re)printed in a collection, magazine or anthology would appear in the latter as a CHAPTERBOOK, presumably confusing our users. At that point Al put a few safeguards in place to prevent CHAPTERBOOK Titles from being created, but he didn't remove the logic that displays all CHAPTERBOOK Titles separately and there are a few holes in the edit forms that let you create them -- as Bill demonstrated above.
Once we decide what we want to do about this whole area, we can either completely remove the CHAPTERBOOK display logic and plug the holes in the application that let you create them or, conversely, we can re-enable their creation and clean up the display logic so that it would display them in a user-friendly fashion when CHAPTERBOOK Titles appear in collections/etc.
Fundamentally, I believe the issue here is whether we want Chap(ter)books to appear on the Summary Bibliography page the way Bill wants them to appear. If we do, then we need a way to tell our display logic that a Title has appeared in a Chap(ter)book Publication and handle it differently vis a vis regular Short Fiction Titles. Ideally, this would be done dynamically at the time the Summary page is generated so that editors wouldn't have to worry about these issues at all. However, that would require that the Summary page logic parse all Publications for the Author, which would kill our performance. The obvious way around this problem would be to create a "This Short Fiction Title Has Appeared as a Chapterbook" flag which would tell the display logic to handle the Title differently. It could be either maintained automatically when Chapterbook Publications for the Title are created/edited/deleted or it could be done via a cleaned up version of the current (half-enabled) CHAPTERBOOK Title implementation. Ahasuerus 16:41, 28 July 2008 (UTC)
My feeling would be that the best way to implement such a display, assuming we want it (and i would be in favor, in accord with Bill's arguments) would be to have the "Chapterbook flag" as you mention above, maintained automatically. I would thing the easiest way to maintain it automatically would be to have it internally be a count, not a flag ("This Short Fiction Title Has Appeared as a Chapterbook NN times"), so that an edit that creates a chapterbook would increment the count, and one that deletes it or changes a chapterbook to some other type would decrement the count. But however the flag is implemented, I think that would be the way to go. Obviously Al's opinion counts rather more than mine does on the technical issues. -DES Talk 17:37, 28 July 2008 (UTC)
No, you're going the wrong way - the chapbook record is a separate record from the shortfiction contents. Did you look at the second example? It's a CHAPBOOK collection which shows the extra record. (Among other oddities.) I don't think you can combine them into the shortfiction record - you need separate CHAPTERBOOK pub and title types because a CHAPTERBOOK publication can also contain INTERIORART, maybe even an introductory ESSAY or POEM. This does satisfy DES's requirement that there is only one title record for the SHORTFICTION: the CHAPTERBOOK record is NOT the text, it's like the COLLECTION or ANTHOLOGY records and if done right will NOT mess up collections or anthologies because they will contain the record for the SHORTFICTION only. However, we're still some way off solving the display issues as you can see: but they've mostly been solved for other types already. BLongley 17:55, 28 July 2008 (UTC)
Ah I think i see what you are doing. You are creating a Title record of type CHAPTERBOOK. To this you are attaching a Publication record that is also of type chapterbook, and that in turn has contents: at least one work of shortfiction, possibly more, and possibly works of interior art, poems, or essays. You are, as you say above, treating the title record as similar to an anthology or collection record. This could work, and maybe it is the way to go, but it won't work well without at least some of the software changes that Ahasuerus mentions above. i don't think it would be my first choice if I were implementing things, but maybe it would be easier, and if Al prefers this route, i would have no objections.
Since software changes will be needed in any case, I would think that changing the NAME of the type would be trivial, and I would be inclined to favor such changes, to "Chapbook" or perhaps "Shortbook" for the publication record, and to something else, I'm not quite sure what, for the title record. But I don't feel strongly about that. In any case, whatever technical solution that Al approves which retains, as much as possible, the "one text/one title record" principle, and allows short works published in separate covers to be clearly listed on the author's biblio page is fine with me. -DES Talk 20:57, 28 July 2008 (UTC)
Yay, go ME! I finally explained something right! :-) I shall now go to bed happy and hope that Ahasuerus understands what I'm getting at too. We'll need software changes, yes, but I think Al has already started most (or completed them, then removed some parts). BLongley 22:27, 28 July 2008 (UTC)
And not only that, we appear to be reaching violent agreement! Who'da thunk it? :)
I think there are actually several ways to implement the changes desired, and while I have views on which one might be chosen, I'll accept whatever makes the programming easier and more reliable. The principles that I think are important in this area are (more or less in order of importance):
  1. As much as possible, one text shall have one title record. Variants to allow for differences in published title, published author name, or works published under a pseudonym are an exception.
  2. Works published under separate covers should be noticeable as such on the main biblio page for an author, whatever the terms used on the display may be.
  3. Works of less than novel length should be distinguishable from works of novel length in most or all displays.
  4. Correct data entry shall be made as easy as possible.
  5. It is undesirable to have to manually change record types retroactively because a new publication was found or created. However if internal flags or data elements can be automatically changed in such cases to facilitate proper display, that is fine.
In addition, if the label "chapterbook" can be changed to something more accurate, that would be nice.
Are we in agreement here? -DES Talk 22:47, 28 July 2008 (UTC)
OK, I see it now. CHAPTERBOOK as a container Title the way COLLECTION and ANTHOLOGY Titles are essentially container Titles since they have no text which is not also entered separately in the Contents section. Kind of what Marc was trying to do, but in a more integrated fashion. Hm...
After thinking about it for a little bit, it seems to be a better solution than what I was driving at above: it lets Short Fiction titles appear on the Summary page twice, once (or 2+ times, I suppose) in the Chapt(er)books section and once in the Short Fiction section, which seems to be better from the presentation standpoint. It will be likely a little easier to implement (no flags required!), but perhaps it will be a little harder to wrap your brain around as a new editor. Overall, I like it better than anything else that we have been able to come up with so far.
As fas as David's principles go, they look solid enough. The Devil, as usual, is in the details :) Ahasuerus 01:35, 29 July 2008 (UTC)

(unindent) Currently, series of short fiction don't show as series, AFAIK. I particularly find this a problem for "chapterbooks" (which have to become shortfiction as soon as you try to make them a series & then don't show as a series), but there are short stories from anthologies which are in series (different anthologies). If the series has at least 1 novel, the shortfiction in the same series show up under Fiction Series, but if there's only short fiction ...
Does the proposed new approach cover this? --j_clark 07:27, 31 July 2008 (UTC)

That is really a separate (although not unrelated) issue, IMO. Note that if you click on the series title, or otherwise go to the series page, all series contents, including shortfiction, is displayed. Whether to display on the author biblio page series that include only shortfiction is a choice with pros and cons, i think. I don't feel strongly about it, myself. -DES Talk 15:19, 31 July 2008 (UTC)

Invalid ISBNs again

Please see User talk:Mhhutchins#Tactics of Conquest for background. In this case, the published book includes an invalid (does not pass checksum test) ISBN, which was duly mentioned in the notes when the book was entered and verified. A relatively new editor has submitted a change to an ISBN that does pass checksum validation, and that correctly retrieves data about the correct title, and apparently the correct edition, from amazon.com and ISBNdb.com.

The question is: what should be done in such a case? A preliminary question, what additional verification would be needed to say with reasonable surety that the revised ISBN (which does not appear in the book itself) is the "correct" ISBN? The main question is, what do when do when and if such additional verification is made and passes.

It seems to me that we have a few choices:

  1. We could change the ISBN field to show the "correct" ISBN, with a note that it is not on the book itself. Such a note should give the number that *is* printed on the book, and how the "correct" ISBN was derived and verified.
    • This has the advantage of making the various ISBN-based links for the publication display work, and of documenting what is on the book for anyone who reads the notes.
    • It has the disadvantage that anyone who does an ISBN search from the book here won't find the correct pub record, and might create a needless duplicate record. It also has the disadvantage that anyone doing a quick check of the book against the title bibliography display may also think there is a mismatch.
  2. We could show the ISBN actually found on the book in the ISBN field, along with a note that it is invalid, and what we think the "correct" ISBN is and why.
    • This has the disadvantage that the links don't work.
    • This has the advantage that anyone working from the book will find a match in the db when using the ISBN as a matching criterion.
  3. We could record what the book has, and that it is an invalid ISBN, but not record the "corrected" ISBN at all.
    • This has the same advantages and disadvantages as choice 2.
    • This has the further disadvantage that the possibly useful "correct" ISBN is not available at all.
    • This has the further advantage that we don't need to get into figuring out what the "correct" ISBN is in the first place. Note that it is not safe to assume that the error is always in the check digit. A publisher could have made an error at any point in the ISBN string.

The last time we discussed this, there didn't seem to be a consensus for any alternative. My personal preference is for choice 1. Note that this only applies when we in fact have a pretty clearly correct ISBN, I am not in favor of blindly "correcting" the check digit and using the result. But whatever the rules is to be, i think we should have a single clear rule for this case -- it has already come up several times and no doubt will come up again. -DES Talk 22:04, 29 July 2008 (UTC)

I'll go with Choice #1, but only if we're able to independently validate the corrected ISBN, and, as Dave states above, we don't correct the check digit just to obtain a working ISBN. If we're unable to obtain the true ISBN, then we use Choice #2, with the pound sign to prevent an error from appearing on the pub page, and prevent someone from "correcting" it. Hopefully this time we'll reach a consensus on the issue. MHHutchins 23:48, 29 July 2008 (UTC)
I would agree with that. Would you agree that finding an ISBN that points at the correct title and publisher at multiple online vendors or biblio sites (like ISBNdb.com or fantasticfiction.com) or library catalogs and that is clearly similar to the stated ISBN, if any, (multiple being "at least 3" here), is good enough that we have the "TRUE" isbn? -DES Talk 00:08, 30 July 2008 (UTC)
Sounds good enough for me. There should also be other provisos: the ISBN must be one assigned to the publisher as indicated by other pubs from that publisher and fall into a pattern of assignment within the time period in question (the date printed in the pub). Also, I wouldn't consider fantasticfiction.com as a true source. It's a tertiary source at best (I believe it gets its ISBNs from Amazon or some other source but can't be certain.) I've also found that when a publisher assigns an incorrect ISBN, in most cases, it won't be used in OCLC or abebooks.com, because both require a valid ISBN for search purposes. MHHutchins 00:32, 30 July 2008 (UTC)
I'd go for choice 1 still too. The check that Mike proposes that the ISBN range for the derived ISBN actually belongs to the publisher is good too - it's one of the reasons I list common ISBN prefixes on the publisher pages as I find them, and I only create those after we've passed a reasonable threshold of verified pubs for that publisher. (Or does someone want to buy the official ISBN register for us?) BLongley 20:16, 30 July 2008 (UTC)
Another check I would propose is that if the derived ISBN is STILL too ambiguous, we don't use it. E.g. even if at least three good sources say it's this certain book for sure, a good source saying it's something DIFFERENT should stop us using that as the Catalog-ID, although of course we can still NOTE the fact that 75% of sources say one thing and 25% another and the book STILL has this annoying non-ISBN on it. What we do NOT want is all those nice links Al has given us taking us to DIFFERENT publications depending on which link we click. BLongley 20:16, 30 July 2008 (UTC)
Oh, and I think we agreed in the last discussion that the error message "(Bad Checksum)" is misleading and might encourage people to only look for errors in that digit - it should simply read "Invalid ISBN" or suchlike instead, and not point the blame at any particular digit. BLongley 20:16, 30 July 2008 (UTC)
I agree with all of Bill's comments here -- if any reliable source gives a different ISBN we don't use a derived number. Of course if it just isn't listed that is a different matter. I also agree on changing the error msg, and i think that had consensus last time. -DES Talk 21:05, 30 July 2008 (UTC)
My vote would be for option #1 and I agree with the ideas about verification with other sources.Kraang 22:31, 30 July 2008 (UTC)
Option #1 with the agreement on verification. Dana Carson 01:43, 31 July 2008 (UTC)
I think this is a good example of how we are NOT going to misuse this. The checks suggested by DES ("at least 3") and Mike ("right ISBN range for publisher") both apply. (Now somebody find me an example of my check applying too! ;-) ) BLongley 22:13, 31 July 2008 (UTC)
I go back and forth on this one. For example, the older DAW books had a catalog # such as UQ1095 and no ISBN. However, when DAW did their DB upload to Amazon they did it using ISBNs meaning that ISBN 0-87997-095-2 is in Amazon and is frequently used to refer to the UQ1095 publication. Also, I regularly run across publications where the catalog # is prominently displayed on the cover/spine and the ISBN is tucked away on the copyright page. With UQ1095 I derive the ISBN and with both publications I add notes explaining what numbers are found where and for UQ1095 make a point of noting that 0-87997-095-2 is not stated. Each time I do this though I'm left thinking "I'd love the list of publications in the title display to show the well known catalog # rather than the ISBN." My concern is an editor coming along with UQ1095 for example and not realizing it's already in the DB as 0-87997-095-2. Thus, I do #1 but don't like it. Marc Kupper (talk) 09:46, 1 August 2008 (UTC)

[unindent] As so often happens when we get into standards discussions, we've started to stray from the original issue: what to do with invalid ISBNs. With Bill and Marc's latest posts, we've wandered over into derived ISBNs, Although tangentially related, talking about derived ISBNS will only dilute the current discussion. The options that David gave us are concerning ISBNs printed in a publication that turn out to be invalid numbers. With that in mind, Marc, does it make more concrete your choice of #1? Sorry to be so tactless, but I'll gladly jump into a debate about derived ISBNs when this issue has been settled. Thanks. MHHutchins 13:11, 1 August 2008 (UTC)

To be fair, my options were, at least in part, about when to substitute a "corrected" ISBN for an invalid ISBN actually found on a publication. A "corrected" ISBN could be a derived ISBN, or one obtained by fixing a transposition or other typo and checking the result, or one obtained from secondary sources, but I directly raised the question of when a corrected ISBN is sufficiently confirmed for us to use.
All that said, pretty much everyone who has stated a preference seems to be supporting option 1. Can we agree on that, and proceed to discuss just what will make us accept a number as a "correct" ISBN? (Actually i think we are near agreement on that, also). -DES Talk 15:07, 1 August 2008 (UTC)
I handle incorrect ISBNs the same as derived ISBNs in that I tend to do option #1 but don't like doing it. Note that I regularly derive ISBNs and the first thing I do is to see if it points to the book on Amazon and also search for it using Google. If the ISBN is found on Amazon or I find it on many web sites as being for the publication then I'll use it. The thinking is that if it seems to be a widespread practice to derive a particular ISBN then I'll use it. I use the same logic when correcting an ISBN in that I'll look around to see if others are doing the same correction. I'm temped to add a "Derived ISBN" field to the db so that the standard ISBN field would always have the stated value and the derived ISBN would only be visible when you are looking at a publication and would be used for the links to other sites. Marc Kupper (talk) 07:10, 2 August 2008 (UTC)

"Collaborative" essays

I was creating variants for "George Scithers" titles (for "George H. Scithers"), when I came upon this record. Pulling out the issue in which this appeared, it was clear that this is not a "collaborative" essay, but six different essays by the six authors. I've seen this more than a few times and have corrected most when I see them. This pub had all the NASA tributes as one essay, and this pub had all the "On Niven" essays as one essay. This pub still shows them as one essay (Bill). I saw this one a few weeks ago, but haven't done anything about it. And I'm hopeful that when someone creates this month's Locus that the Disch tributes won't be entered as one essay. Shouldn't there be a standard which states that when there's a clear distinction between each author's contribution that each should have their own record? I wonder how others feel about it. MHHutchins 23:19, 17 August 2008 (UTC)

I think there should, see ISFDB:Help desk‎#duelling editorials for another such case. I noticed the On Niven case just today, because thr are two records for it on file, one with only a single author (Benford). I was going to check my copy of N-Space. The problem is that we have no good way to indicate that such separate works are part of a larger, overall group, as these are. -DES Talk 23:39, 17 August 2008 (UTC)
I don't think it's important to indicate that the essays are part of a larger group. Just looking at the pub itself should make that evident. Here's another example that I entered awhile back. MHHutchins 23:45, 17 August 2008 (UTC)
Yes, in a case like that, it's obvious (a bit less obvious if you find one essay on an author page). In cases like this it might notm be quite as obvious, particularly if the group title had not been uused as a prefix. But if you don';t think it important, there are always tradeoffs. (Actually the time I really want a group/member indication is where several works of short fiction with separate titles are also part of a less-than-book-length group, with an over-arching title. That we don't really handle well, and it happens more often than grouped essays, and tends to be less obvious from the titles alone.)
Can you make a mini-series for them? Just a thought. kpulliam 23:53, 17 August 2008 (UTC)
I'm generally supportive, but lazy. ;-) I'm pretty sure I let someone else do Requiem and N-Space and cloned afterwards. I know I've even mentioned waiting for someone else to do the work on another Niven title, so I could clone it for the UK version. (Which turned out to be the same "world-wide" version in the end I think, so that didn't help.) I'd happily add this as a desired goal, and everyone can feel free to add such detail to anything I've verified (but please take-over verification if you do), but I'm rarely that interested in that much detail about the non-fiction elements of works here. I'm concerned about the Fiction, I like pointers to the bibliographies and might find such in obituaries (although I've never yet purchased a magazine or book for such, but reminding me I already own such in a book can be good), but so long as we get the fiction right I'm happy. (And yes, I know, I asked Mike about a preface to a pub earlier, but that's personal interest about inter-author-bantering, not something I care about bibliographically.) BLongley 23:55, 17 August 2008 (UTC)

Dateless Printings Redux

(Moved from the 2007 discussion to begin a new one.)

Has there been any movement on this issue? Has the 'print # in day field' been working? Are there any other work arounds currently in use? (I've seen some where the info was put into the publisher field as a parenthetical comment "Signet (5th Printing)" but with the work to cleanup publishers I assume that's a least preferred solution - though I will point out, that solution CAN be harvested with a script without interference of actual publication days in other records). Kevin 03:42, 22 August 2008 (UTC)

I don't use this method, and haven't approve any submissions that use it. Judging by the date (July, 2007) of the last comment on this subject, no consensus was ever arrived at. Neither the standards nor the help pages were changed to reflection this method (thank God!). As it stands now, we still enter "0000-00-00" for undated publications, which is probably the best we're going to get. My wish is that those pubs would come at the end of the pub listings. MHHutchins 04:22, 22 August 2008 (UTC)
Well, then let me propose this. How about for all 'unknown' dates, where we know it follows some other date (Date of copyright, date of original printing by that imprint, date of 36th printing) we simply use a single nonsense 'day' code and the known year (and month). Like 1971-06-99. This will make them appear in 'the best of our knowledge' order (Excepting that an undated 4th and 5th printing will be clumped and possibly reversed). The only implementation question is can the software handle days greater than 31? If a true software solution is ever created, it will also allow for (at least with a raw SQL Query) a list of pubs with this workaround, and presumably printing information in the notes, waiting to be converted by hand.Kevin 04:50, 22 August 2008 (UTC)
This presupposes that we know the year of publication. If we did, it would have already been entered as 1971-00-00, and thus be in a rough order with other editions/printings. When we enter 0000-00-00, it means we don't know the year as well. MHHutchins 05:07, 22 August 2008 (UTC)
I didn't explain myself very well. Imagine 1965 copyright. Known 1970 2nd printing. Unknown 3rd. known 1985 4th printing. Enter as 1970-00-99. We don't know the year, but we do know it was after (or during) the known year of 1970. This causes it to appear in order but with a flagged day of 99. Kevin 00:00, 23 August 2008 (UTC)
That sounds reasonable, but I wonder how it should be displayed within the list. I suppose it would still display "unknown" as the publication date, but at least it would be in rough order. If you think this could be readily programmed by Al, I can move this discussion to the bottom of the page. I don't think many editors would be aware of it here in the middle. MHHutchins 00:09, 23 August 2008 (UTC)
Hmm At this point I think it needs input from Al. Two outstanding questions. Can the date field handle values over 31? Can the (unknown) logic be amended to also look for YEAR-MO-99 and display it as (unknown) in addition to the 0000-00-00 entries?Kevin 00:32, 23 August 2008 (UTC)
Days of 99 currently don't get accepted, they cause the date to become '0000-00-00'. This is one of the things that killed using it for printing number. Putting the printing number in the publisher looked good but prevents publisher regularization. I can't think of anything we can currently do easily that helps us without causing a different problem, we're going to need Al to do something. BLongley 09:10, 23 August 2008 (UTC)
Also note that invalid dates (especially year) create problems in other software platforms.--swfritter 13:27, 23 August 2008 (UTC)
I've been using the invalid date format though with YYYY-MM of the first printing and DD is the printing # though have not needed to deal with more than a 20th or so printing meaning the date itself is still valid. I think things are settling down enough that I'm going to re-setup my ISFDB development machine and to add the printing # field. Plus I want to add a first-printing field that'll contain the pub_id of the first printing. If it's set then those publications will be displayed in order underneath the first printing. Marc Kupper (talk) 07:44, 27 August 2008 (UTC)
On the other hand, couldn't we implement an 'end of the list' forcing by using a nonsense year, perhaps 7777-00-00. Again this is searchable (And by searchable I mean that if it's ever determined to be a bad idea it probably can be completely removed in about 5 minutes of global editingKevin 04:50, 22 August 2008 (UTC)
This had been discussed and it seems like a quick fix to me. We currently use 8888-00-00 to indicate a work that was announced but never published. See this famous example. Also see how it's listed on the author's summary page. MHHutchins 05:07, 22 August 2008 (UTC)
We are using the date format supported by our database, MySQL, which doesn't recognize dates prior to 1000-01-01 (not that many SF works were published prior to the 11th century) and imposes other limitations as described above. We didn't discover this issue until we were too far along to choose another database... Ahasuerus 03:53, 24 August 2008 (UTC)
Really? The birth and death dates of Lucian of Samosata are well before 1000-01-01 but they seem to work correctly. -DES Talk 05:52, 24 August 2008 (UTC)
We seem to have problems with BC dates though. BLongley 12:00, 24 August 2008 (UTC)
Well, the "Y1K" problem came up over 2 years ago right after the ISFDB1->ISFDB2 conversion and we may be using a newer version of MySQL after moving to the new host this spring. Perhaps it's something that the MySQL crew fixed in 2006-2008? Ahasuerus 19:17, 24 August 2008 (UTC)

The Analytical Laboratory (AnLab)

Tpi entered a change to Analog, November 1974, putting into the Notes field the AnLab results (which were actually published in the February 1975 issue). (For anyone who hasn't read Astounding/Analog: AnLab is a tabulated readers' poll.) I put this on hold, as I don't think this is something we've done before. Looking around a bit, I see in an earlier discussion here in Rules_and_standards_discussions#Series_in_Fantasy_and_Science_Fiction some tangential comments, some of them positive for including this content. Concern was raised about copyright, but I think that would probably not really be an issue here (reviews were more in view, I think).
I guess I'd raise two questions:

  1. Do we want to try to include this content? I admit to mixed feelings, myself.
  2. If we do, is this a reasonable way to do it? For the benefit of non-moderators, Tpi's submission adds this at the end of the pub notes:
    Analytical laboratory results for this issue (from Feb 1975 Analog): The Indian Giver (Pt.1) 1.91 This best of All Possible Worlds 2.68 When No Man Pursueth 2.76 A House by Any Other Name 3.46 Unlimited Warfare 3.8
I don't think that format will be particularly readable, but may be about as good as free-form notes entry will support. Another possibility might be to put a title note with each story; that would be a lot more trouble to do, of course.

Anyway. Comments? Thanks. Dave (davecat) 17:59, 27 August 2008 (UTC)

Well, I think that those AnLab results are very interesting data, and it would be very worthwhile to have easy access to it. And the results of the yearly Anlab polls (the yearly format was started somewhere in the turn of 70s and 80s) are in database as Analog awards. (by the way, I don't really understand why they are called "Analog awards" here, as in Analog (and in Locus, also) these results are usually referred as AnLab or Analytical Laboratory results.) Tpi 19:02, 27 August 2008 (UTC)
The notes are a good place for them - either in the pub notes or the AnLab title notes - or perhaps both. As for formatting. I am opposed to using HTML to format the text - it discourages other editors from adding information to the notes and it makes the data more difficult to use in other applications. In some cases I have used the simple <br> to force a line feed. Even that may be going a little bit too far. It would be nice if the isfdb display logic interpreted line feeds as line feeds.--swfritter 16:56, 28 August 2008 (UTC)
OK, I'll unhold the submission. Thanks. (Tpi, I guess this means you should feel free to add all those things, if you want to do the work. Dave (davecat) 18:03, 28 August 2008 (UTC)
Ok, I will enter these. In fact, I already have those on spreadsheat as I am putting year 1975 to the storage, and wanted to have that information available at least for me - adding that data is just cut and paste job. Tpi 18:28, 28 August 2008 (UTC)

Short Fiction Published Outside of Magazines

Short fiction is starting to be produced, either as online publications or eBooks, or as Audiobooks/podcasts, without having a parent publication at all. Two cases in point:

  • Tor.com has started posting short stories by known authors directly onto their website, accessible to anyone who registers a free account on the site.
  • John Scalzi is editing a collection of short works called Metatropolis for Audible.com - to be produced in the first place as an audiobook, though presumably the stories were first written and John has in fact announced that he self-published a few copies of the novella he wrote for it and is giving them away as contest prizes.

Are these in the purview of this project, and if so how do they fit into the database schema? Should they be called books, if they are ebooks or audiobooks? Is Metatropolis simply an anthology? I don't think it's appropriate to call Tor.com a magazine, but neither are they an ephemeral publisher. (I tried to find some reference to standards on web publications in general but if they're here I've failed to find them). Netmouse 21:21, 18 September 2008 (UTC)

We already have a number of such separate ebooks, and i think some such audio books, in the database. See ISFDB:Policy#Rules of Acquisition #s 1 and 14. Ebooks are in if downloadable and stable. They are treated as separate publications, of type novel, anthology, or collection where appropriate. Where there is a separate ebook containing a single work of fiction shorter than a novel, the publication is of type "chapterbook". See this publication and this one for examples. Audiobooks are treated similarly. -DES Talk 21:48, 18 September 2008 (UTC)
I once asked about escapepod and podcastle. I have been planning to add these, but haven't had really time, and I have been unsure how to add them. They seem to be fairly stable, escapepod has been there for well over three years. And even if the sites might disappear at some date, the produced stories probably will not, as they are Creative Commons, Attribution-NonCommercial-NoDerivatives licensed, and will probably float on the net for forever. They could even be downloaded to isdf wiki, at least from copyright point of view - it would only be a maybe 10 gigabytes or something... :-) Well, serously - I have been thinking about adding them as magazine. There is an editor, there is regularity and there same form for all episodes, so they could be interpretted as an audiomagazine. Tpi 14:36, 19 September 2008 (UTC)
Are these formats designed and implemtned so that users routinely download and store local copes of these stories, so that they are not dependant on a single site that is a single point of failure? Are they in fact so used on a routine basis? If not, i would class them with web-zines that are OUT. Otherwise, i see no reason off hand why they should not be IN. -DES Talk 16:36, 19 September 2008 (UTC)
There are rutinely downloaded and stored locally. You can listen them online, but most people download them and use their mp3-players or similar devices. I personally listen them through my cellphone with bluetooth headpiece while driving. If and when they should be in, would the magazine classification be the best? Tpi 17:09, 19 September 2008 (UTC)
For now I'd propose a title-tag, Digital Content, be applied to these stories so that if at some point we either come up with policy or code/db to handle them that we will be able to spot the stories. Digital Content would apply to any story that has not been published in print meaning it covers web, pdf, ebook, audio, video, radio or TV broadcast, etc. I know some of those are analog media but I can't think of a word that would show the story was distributed or sold to the public but has not been distributed on dead-tree based media. Marc Kupper (talk) 18:38, 19 September 2008 (UTC)
Well, those podcast I was talking have usually been published earlier in print (almost all Hugo-nominated short stories from past two years have been featured there among others) , so there aren't only digital content. Tpi 19:00, 19 September 2008 (UTC)
I'd say a podcast of a published story is in. As long as a story has appeared in print it's ok to add publications in other media. For example, there are editors adding publication records that link to stories on the the Project Gutenberg web site as those stories did appear in print. Marc Kupper (talk) 06:55, 20 September 2008 (UTC)
Would it be ok to use "magazine" as format for these submissions? Tpi 15:15, 20 September 2008 (UTC)
I would suggest treating them as individual units - I would imagine normally as chapterbooks. While the the mp3's may be around for a long time, the virtual magazine wrapper in which they currently appear on the internet could disappear at any time or the contents of an issue could change which is something that recently happened to Helix. --swfritter 16:03, 20 September 2008 (UTC)
Personally I don't feel comfortable using the "chapterbook". What are the chapterbooks anyway - the definitions on wikipedia and and freedictionary talk about children's books - here that term seems to be used for something else - I am not familiar with the term at all - and I suspect there are a lot of people who are also very unfamiliar with it.
Those podcast call themselves as magazine [3]. Also that "magazine wrapper" can't at least legally stripped away, as the CC license covers whole podcast, not only the story: " the entire world has permission to distribute the podcast for free, provided they give credit for it, don’t try to make money off of it, and don’t change it in any way. " They are also paying market with contracts, so it probably isn't even possible for the authors to pull back they stories even if they would be pissed at the editors. But if the consensus here is, that they should be handled as chapterbooks, I am ready to start adding them as such. The right procedure then would probably be "New Anthology" => pub type "chapterbook"? Tpi 06:44, 21 September 2008 (UTC)
I noticed that it is possible to make contribution as "New magazine" =>pub type "chapterbook" maybe that? Tpi 06:51, 21 September 2008 (UTC)
Please do not use the "Magazine" type here, unless there are regular issues, collecting multiple works into distinct and permenant sets, which are set up to be downloaded and accessed as a group, and unless there is an identified editor. Probavbly not unless the magazine as such has an ISSN or an ISBN.
Since there is no "new Chapterbook" link, it is probaly esaiest to use "new Collection" and then change the type to Chapterbook, as a collection (or anthology) is functionally closest to a "chapterbook". Creating a new Magazine record creates the special unbderlying EDITOR record, used by all magazines but by no other type, and it would be better not to do that for a record which is not going to stay as a magazine.
"Chapterbook" on the ISFDB appears to have been an early error for "Chapbook", now unfortunately embedded in the code. According to Help:Screen:EditPub#PubType:
CHAPTERBOOK. This is used for anything smaller or flimsier than a standard paperback. These are often, but not always, saddle-stapled; publications from conventions are frequently in this format. This format is also used for separate publications of a single work of short fiction, even if bound as a standard paperback or hardcover. It is also used for an ebook or audio edition of a single work of short fiction.
I hope that makes it clear that this is the type for such publications, unintuitive as it may seem. -DES Talk 14:26, 21 September 2008 (UTC)
I might also note that Helix is also a paying market - presumably with contracts. I really do wish there were an easier way to enter these. The magazine method is so tempting.--swfritter 15:01, 21 September 2008 (UTC)
It sounds like there may be a need for a new title type but I don't quite understand what it is. Would "Digital Media" suffice?
An idea that just came to mind is that novel length works (40,000 words or more) would be filed as type NOVEL and that a new title type NOVELLA would be added. The existing display logic that deals with CHAPBOOK would be modified to also look for NOVELLA and the section header would be one of "Chapbooks", "Standalone Novella or shorter length publications", or "Chapbooks and standalone novella or shorter length publications." You would see that digital editions are available when viewing the publications. Has someone removed all of the CHAPBOOK titles from ISFDB? I can't find any. Marc Kupper (talk) 02:24, 23 September 2008 (UTC)
I don't see what difference there would be, besides the name, between the current Chap(ter)book type and your proposed Novella type. I agree that an improved name and improved display logic would be good for the current "chapterbook" type. Note that sometimes a short work shorter than a novella gets published separately -- would we also need a "Novellette" type? -DES Talk 22:59, 23 September 2008 (UTC)
I would strongly oppose "digital media" as a publication type, just as I would "acid-free paper" as a publication type. An anthology is still an anthology when publsied as an ebook. A novel is still a novel. No new type is needed. Som digitally formatted magazines (Jime Baen's Universe, for exampel) are being treated as magazines because they fit the general parameters: Stable issues, that people can possess individually, independant of the publisher (once they aquire them) with an editor. Digital chapbooks (or "short pubs" or whatever we want to call them) should still have the same type as short works individually and separately published on paper. 22:59, 23 September 2008 (UTC)
As for chapbooks. I agree it's a confusing term as "chapter books" are aimed at the children's market. Chapbooks are rare these days. It's my understanding that corner news-stands and news-stores used to sell books by the chapter or individual short fiction works. You bought one per week or however often they came out. I don't own any specfict chapbooks but have one non-specfict work. It's bound like a paperback with a flexible cardboard cover and a flat spine though one that's only 1/4 inch thick and is 9" high by 6" wide. It's half title page has "Number One International Chapbooks" at the top/center and was published in 1938. It's 38 numbered pages plus a blank leaf page. This chapbook is a story/commentary in itself and not a chapter from a longer work. I believe the reason this particular binding format ended up as an ISFDB title is to designate that it's not a 40,000 word novel. I see there's an article at wikipedia:Chapbook. That picture of a "modern day chapbook" is terrible. That's a pamphlet, not a chapbook and also seems to be something included with a product and never sold separately. I'll pose my chapbook and get a picture uploaded at some point. Marc Kupper (talk) 02:00, 23 September 2008 (UTC)
In modern poetry circles 9and I think also "literary" circles, a single poem, or a few poems, bound and sold separately in what might in other circles be called a pamphlet is generally called a "chapbook". Indeed I think that any literalry or fictional work published in pamphlet form may be called a chapbook. -DES Talk 22:59, 23 September 2008 (UTC)
I have approved five of them. 1, 2, 3, 4, 5. How do we think they look? Can the web address be made linkable? Put the titles in a series?--swfritter 16:44, 21 September 2008 (UTC)
I think the existing Chapterbooks look good, and are nicely sorted. But with current Chapterbook support (which I think is practically "none" after initial entry, although some merges or edits for content should still be possible) I don't think you'll be able to put them in a series. Contents yes, the Chapterbooks themselves no. BLongley 20:16, 24 September 2008 (UTC)
You got here first. :-) I was also going ask for comments. So, comments, please, anyone? Series at least, I think. Tpi 17:45, 21 September 2008 (UTC)
By the way - is there a way how remove that extra "Chapterbook" line in the contents-listing? Tpi 07:58, 22 September 2008 (UTC)
I believe not - CHAPTERBOOK title types are a one-shot at the moment. Al should make them disappear automatically if he makes them a "Container" title type like "ANTHOLOGY" or COLLECTION". (We have some small books of poems called chapterbooks, and even a chapterbook can have interiorart or an introductory essay, they're not always for a single piece of fiction.) BLongley 20:16, 24 September 2008 (UTC)
Another thing: Is there a way to get in the chapbook "editor" credit instead of "author" crdit? Now, if the title of story would be mentioned in title of the episode (e.q Escapepod EP001 - Imperial") the publication listing credits the editor as an author, not the real author. These are the changes I would like to have in record, probably easier to tell from picture that from my broken engrlish. Is this format even possible? Tpi 15:47, 22 September 2008 (UTC)
A series seems reasoanble. I don't think there is any way to cause an "editor" line to appear for a chap(ter)book, although there probably should be. Even on collectiosn, when there is an editor other than the author, i think we have to credit the editor as a co-author and add a note to explain. I'll look into whether and if so how we can get rid of the chapterbook line in the contents -- I agree that it dosn't look well. When the chap(ter)book ahs the same title as the work it includes (as it often does) such a line does not appear, IIRC. -DES Talk 16:29, 22 September 2008 (UTC)
I suspect an EDITOR record could be added if you chose "new magazine" and changed title type to "CHAPTERBOOK" on entry, but I agree with swfritter that such should NOT be done. Editors of Magazines are very different from Editors of Collections (that don't even get a co-credit under current guidance - put them in notes) and Editors of Anthologies (that do get credited). BLongley 20:16, 24 September 2008 (UTC)

What is an RPG?

I know Role Playing Games are out but I don't know for sure what an RPG is. I've heard there multi-player games involving simulated battles for example that have books and usually large table size boards with pieces. There's Dungeons & Dragons which I know to be a true RPG.

The reason I ask is I have a number of books that have a page or two of text followed by a question and two or three answers or two or three statements such as

  • If you pick the green button turn to page 80.
  • If you pick the red button turn to page 48.

You navigate your way through a story and either end up a hero, dead, or maybe not much happens. The sections are usually interesting enough that you could read the books cover to cover but there will be no continuity to the story. Are these books classified as RPG? If not, then what are they called?

Some of these books are just straight read and chose. With one of my books you also need one dice and besides choosing a choice you also roll the dice and then look up stuff in one of a number of tables to see how your score got adjusted. It was still a single player thing but it is more like a game in that you keep a score. I don't think two people could play at the same time as the book also had you jumping to sections based on your choices and the outcome of dice rolls. Sorry, I can't remember the title but the setting was future warriors and you were a leader of a small commando team that would get dropped planet-side. The warriors had these "jumper" or "bounder" rocket backpacks that allowed them to move rapidly by jumping. The bad guys seemed to be creatures that lived underground. Marc Kupper (talk) 03:55, 20 September 2008 (UTC)

I call these (or perhaps they were labeled as) 'Choose Your Own Adventure' books. If Spec Fic, I would call them in. As to the book that requires dice to 'read', that's borderline, which means it would probably be in unless someone complained. As to your original question, an RPG is either a Role Player/Playing Game, or a Rocket Propelled Grenade in my vocabulary, both out I believe. Kevin 04:05, 20 September 2008 (UTC)
I agree with Kevin above. t least soem of these were indeed called "Choose Your Own Adventure" books, I remember reading/owning some. (In fact I specifically recall one based on McCaffery's "Pern" setting.) The one you describe above sounds as if it were based at least loosely on Starship Troopers. Ones where dice are used and score is kept seem on the "debatable" side, ones where it is just a matter of makign choices seem clearly IN to me. A note on any such publication record about the special format would be a good idea, IMO. -DES Talk 15:45, 21 September 2008 (UTC)
See also http://www.isfdb.org/cgi-bin/tag.cgi?3121 for items tagged "Your own story". -DES Talk 16:26, 21 September 2008 (UTC)
We have the Combat Command series which I believe are books like this, and fit with existing SF worlds. And You Can Be the Stainless Steel Rat is also "A Role-playing Adventure". So long as it's SF, and a single reader can get a STORY out of it, I'm happy if it's in. Adding dice doesn't make too much difference to me. However, there is a category of books I'd leave out - "Duelmaster" was one such series. These were PAIRS of books, one for each player. Although often in a Fantasy setting, the fact that they're useless to a single reader makes me lean towards EXCLUSION. Most of the RPG pubs that get deleted rapidly are rulebooks (no actual story, just background material) or adventure "scenarios" - loose settings where the players make the story. Of course, when a set of players gets a good story out of such, somebody might well write it up as a proper story, which is why we have true novels and collections based on D&D, Warhammer, Magic the Gathering, etc. BLongley 17:58, 21 September 2008 (UTC)
I seem to recall we once had several "Fighting Fantasy" books of this type, but they're gone now. If it wasn't for this pub (and one review I suspect is incorrectly attributed) we might have been able to lose one of the Steve Jacksons entirely. BLongley 18:34, 21 September 2008 (UTC)
I happen to have several hundred RPG books/modules/sets/campaigns and it is as addictive as any 'written' work can be. I find it somewhat amusing that Dragon Magazine is allowed, while the rulebook is not. I am also glad that because of the rulings I do not have to enter them. I do think they are as much 'fiction' as anything more formally known as a novel. They would eat up a huge amount of memory to cross-correlate and do deserve doing, but it would harm the work here. The actual adventure modules and the adventure books are the connectors between the novels and the RPG. Do to the movement into 'computer game devices' the RPG's have suffered. RPG in printed form can delve into the same essentials that novels do, though they require one to think it out.
I suggest that people look into an easier example such as 'Empire of the Petal Throne' and check out the volumes of written material it has created. Then consider, M.A.R. Barker, it's creator will soon be dead and pretty much the material will lapse into some type of 'ERB' model. Is it worthwhile to find a good(fair) novel such as 'Flamesong' and yet not attribute it fully to 'EPT'?

[4]

What I am saying is I personally find rpg's speculative fiction even in the rule books. Just because the orientation to do something, seems a poor rationale for exclusion from the db. The rpg rule book is designed to promote thought, just as the 'classical' novels did/do. Exclusion makes great sense for the db because it (the db) would be overrun by gamers. Thanks, Harry. --Dragoondelight 13:47, 1 October 2008 (UTC)

Ghost written stories.

What do people think is the best way to deal with a ghost written story. Lester del Rey was overloaded so he had Paul W. Fairman ghost at least one book for him. The Runaway Robot is known as a del Ray book but Fairman was the real author. Making is a pseudonymous work which is what I currently tried means it is listed under Fairman and not under del Ray at all. Given that it known as a del Ray book and has no other name listed this doesn't seem like the right thing. Maybe make Fairman a coauthor so it shows under both? Dana Carson 00:51, 28 September 2008 (UTC)

Co-author seems like a good idea in this case, with a titel-level note. After all, a number of books explicitly credited to two authors have 95%+ of the work done by one author, usually the less famous or popular author. But this will need to be done case-by-case, i suspect. -DES Talk 02:26, 28 September 2008 (UTC)
Done. And from the discussion on rasfw they think the outline was done by del Rey so that even fits. Dana Carson 08:46, 28 September 2008 (UTC)
I played with different permutations at one point and wasn't particularly happy with the results, but I finally settled on what appeared to be the least of the available evils. I would first enter ghost written books using the book's officially credited author, in this case "Lester del Rey". I would then make that Title into a Variant Title by both authors, in this case "Lester del Rey+Paul W. Fairman".
The advantages of this approach are that:
  • it preserves the "stated" author's name as given in the Publication record, which is one of our basic principles
  • it credits the ghost writer in an easy-t-o-find cross-referenced way
  • it also keeps the book on the "official" author's Summary page.
The last one is particularly important since failure to list the book on del Rey's Summary page -- which is what an outright "pseudonym" solution would result in -- would make it effectively invisible to most of our users. You could even argue that the officially credited author (e.g. Shatner vs. Goulart) is a collaborator of sorts, although it would probably stretch things too far if we started using this approach on posthumous ghost jobs a la V. C. Andrews :-) Ahasuerus 23:03, 29 September 2008 (UTC)
I like that. Will do it that way. Dana Carson 21:25, 30 September 2008 (UTC)
When the nominal author was alive and presumably had at least some input into the ghosted work, i like the co-author method better. For phosthumous ghosting, or cases where it is known that the nominal author had no input at all, i prefer the VT method. Note that your three principles above are followed in both cases. -DES Talk 21:30, 30 September 2008 (UTC)


Chapter Books yet again

I've read (some of) the earlier discussion on chapter books (as opposed to chapbooks) & been ruminating on the issue ...

I have a lot of kids books that fall under our "speculative fiction" guidelines, ranging from picture books through to young adult. Many of these are less than 100 pages. Those that are ~ 100 pages are generally less than 40,000 words due to largish typefaces, well-spaced layouts, and sometimes illustrations. (Of course, there are the later Harry Potters, Brisingr, etc. which are no problem with length => NOVEL.)

Although my "definition" of NOVEL is a sort of sliding scale, with less pages as the target age goes down, the books for early readers with a lot of illustrations, I like to put as CHAPTERBOOKS (except when I want to put them in a series or have to edit, then the choice is SHORTFFICTION or NOVEL). (And for a series of chapter books to show up on the Summary page as a Fiction Series, the only choice is NOVEL).

There also seems to be a niche market for short "novels" for reluctant readers and those learning English as a second language, where the target age is teen, say, but the length is, say, 50 pages. Some of these fit the "chapbook" description, it seems to me.

Anyway, the particular aspect I've been wrestling with is:

... when such a publication is under SHORTFICTION and it looks like a novel, it's not obvious to find. Completists like me can't pick straight away that we are missing a book by that author. The problem for ISFDB is compounded when the item has been previously published in a magazine, anthology or collection (or vice versa), so one or more of the publications of the title would be reasonably deemed to go under SHORTFICTION.

The particular title that prompted this article is Andre Norton's Outside from 1974. My Blackie hardcover looks just like a typical kids/young adult hardcover from the time, but I couldn't find it under Fiction Series or Novels on Andre's Summary page. The book is 126 pages! I would have been tempted to enter it again as a New Novel, except I wouldn't expect any of her novels to have been missed in ISFDB. (I assume it's under Shortfiction due to application of the "less than 40,000 words" criterion due to a large typeface, well-spaced layout and illustrations.)

To get around the problem, I wonder if there is a way to display titles on the Summary page twice for such situations - once under Novels or Chapterbooks or Fiction Series, and once under Shortfiction? (automagically or by an editor setting a flag??) (Well, on pages other than the Summary Page, too, but I don't look at alphabetical, chronological, etc., much.) ... clarkmci/--j_clark 20:57, 5 October 2008 (UTC)

Brother! :-) I have exactly the same problem - people keep trying to hide BOOKS from me! You can find Outside here, but as only the CHAPTERBOOK PUBLICATION type is allowed and the TITLE type isn't, it'll never appear as a book unless reentered. You can find a properly entered CHAPTERBOOK as a separate BOOK for a while, e.g. Julia S. Mandala or Stephen Eley but as soon as you edit it it disappears into SHORTFICTION. Marc has proposed some variations for a SHORTNOVEL type or such, but apart from the name, CHAPTERBOOK would work for those of us with short "Novels" that other people want to demote on "Too Short" grounds. It keeps them as books and the lengthists get the contents recorded their way. Yes, it would require programming changes but apparently CHAPTERBOOKS used to work, if not as well as they should: but making them a title type like COLLECTION or ANTHOLOGY should be comparatively easy. BLongley 21:38, 5 October 2008 (UTC)
Any clue who has entered these Stephen Eley entries? They are neither Chapterbooks, nor novels, and he didn't author any of them. CoachPaul 22:17, 5 October 2008 (UTC)
Tpi entered them, after guidance on how to enter single shortfiction titles published on their own. Stephen Eley is the Editor. Which we can't represent properly at the moment. See User_talk:Tpi#Escape_Pod_12. BLongley 22:55, 5 October 2008 (UTC)
I'm not sure why the "displayed only once" logic is done and agree it's awkward at times that if you put a shortfiction in a novel series for example that it vanishes from the shortfiction list.
For now, I personally would prefer to deal with niche markets using series or tags rather abusing CHAPTERBOOK. With either of these it will also be easy to track down and convert the titles once we add support for CHAPTERBOOK, CHAPBOOK, SHORTWORK, etc. Marc Kupper (talk) 06:31, 6 October 2008 (UTC)
It's CHAPTERBOOK's that seem to be providing a lot more abuse than they are being given.--swfritter 19:18, 6 October 2008 (UTC)
They are a major pain at the moment, and I suspect we could move all but the EXAMPLE ones to something more workable in the meantime. I suppose the big question, for anyone that participated in the ISFDB:Community_Portal#The_Trouble_with_Tycho_What_is_a_novel.3F discussion, is whether they could live with a "The Trouble With Tycho" ANTHOLOGY containing JUST the "The Trouble With Tycho" Novella? I could. It would remove the NOVEL record that some people object to on grounds of length. Single-fiction-entry Anthologies would be easy to spot when Chapterbooks/ShortNovels/whatever finally get resolved. BLongley 20:32, 6 October 2008 (UTC)
Sounds good to me.--swfritter 15:07, 7 October 2008 (UTC)
Bill, the "The Trouble With Tycho" discussion is 8950 words. I'm sure Mr. "the word count trumps all" will know exactly which rule/classification that makes it. :-) The trouble is, I have no idea what you mean by the anthology containing just the novella. I see an omnibus. I believe that particular example is also more troublesome than most as variant names are also involved. Using anthology as a general rule seems to introduce "great astonishment" as it'll display the authors as "Anthology editors" on the summary bibliography pages unless you are thinking of making the editor "unknown" or "chapbook". Most people seem to use omnibus unless the book looks/smells like an anthology as the omnibus category is already being used for dos-a-dos. The Binary Star series uses #1-4 as anthologies and #5 being a test case using omnibus. Marc Kupper (talk) 16:04, 7 October 2008 (UTC)
The Omnibus/Ace Double you see is fine, the "Novels" 1976 here and 1983 there are the ones causing dissent. Changing the NOVEL there to ANTHOLOGY (for both title and publication type) will leave a visible ANTHOLOGY of the same name containing one Novella, and the ANTHOLOGY will show only one NOVELLA entry as the ANTHOLOGY content record would be hidden like all good "container" types. We could also change the NOVEL to a COLLECTION as swfritter mentions below, or even to a single-entry OMNIBUS - any of the "container" types will do to some extent and each will cause some sort of display confusion. But I get to see real Books separately, the fiction content is always the right "Novella" length, ISFDB hides the container types from us, and we don't get CHAPTERBOOK edits causing confusion. I'd keep the title and author matched exactly to the content for now, for easy discovery when the software allows. BLongley 19:02, 7 October 2008 (UTC)
Wouldn't collection make more sense?--swfritter 16:56, 7 October 2008 (UTC)
"Collection" is generally reserved for a publication containing works by a single author but I don't have an objection to testing out a "Collection" containing one work as a way of packaging publication of a single shortfiction. Marc Kupper (talk) 18:09, 7 October 2008 (UTC)
Quite possibly. Which looks weirder: an Anthology where someone is Editor of his own Novella, a Collection with only one item in, or an Omnibus with only one item in? Of course, we should have a standard note as well, explaining why we're using this format in the short-term. But I think we all can see now why CHAPTERBOOKS are undesirable. BLongley 19:02, 7 October 2008 (UTC)
I suspect if we called The Trouble with Tycho (on its own) publications an Omnibus for now, then people might assume it's actually the two-entry Omnibus/Ace Double Marc sees, incorrectly entered. If we call it a Collection people might assume it's a stub-entry where we don't know the contents, but made an educated guess as to ONE of them. Similar for Anthology, but an Anthology wouldn't usually have the same Editor and title as one of its constituent Authors and titles. Any or all of these options might be usable depending on what you're trying to differentiate a standalone publication of shortfiction FROM. BLongley 19:02, 7 October 2008 (UTC)
Most of the Chapterbook entries are essentially single story collections. Chapterbooks should work exactly like collections with only a change in name to differentiate them. The idea is, of course, that there be a software implementation to resolve what is basically a display issue. As below, they will be easy to find.--swfritter 16:17, 8 October 2008 (UTC)
I might also note that multi-story pubs are not precluded from being defined as Chapterbooks. The original definition for Chapterbooks seems to apply to anything that does not fit in any defined category.--swfritter 17:21, 8 October 2008 (UTC)

Image file names

Should we have a standard for image file names that we use with publications? Most people are using the publication tag but occasionally I see people using other names. We can edit MediaWiki:Uploadtext to include instructions such as the one I added recently about the image license tags. Marc Kupper (talk) 15:20, 7 October 2008 (UTC)

We discussed this shortly after uplaod was enabled. The pub tag approch was favored by several people, but ther was no consensus. The problems with it are (IMO):
  1. If you do the uplaod first, as I sometime find handier, the pub tag may not be assigned yet. This could be fixed by always doing the record first, but that means a non-mod would have to wait for approval.
  2. It is not uncommon for the same cover art to be used on multiple printings, soemtimes even on diffreent editions with no visible cover difference. Which tag should be used in such a case.
  3. If one is searching for a particular work of cover art on the wiki, the pub tags are not helpful. If you know the title and edition, you can search through the db first and find the tag. But if you don't, the puubtags are not useful search targets. The categories for artist and publisher may help, when the uplaoder has used {{Cover Image Data}}.
The advantages of using the pubtag are (IMO):
  1. It is guarenteed to be unique. This is not a major beneffit, IMO, because if you chose an existing file name on uplaod the wiki software detects this, warns the user, and prompts for a different name or confirmation that an overwrite is desired (for an updated version of the same image)
  2. Given the pub record form the DB, the file is easy to find. But then, if the image has been used, then there is a direct link to it from the pub record anyway.
  3. It takes no thinking on the part of the uplaoder.
When uploading was first enabled, I started using names of the form "Short Title-Artist:. For example Image:Sharp_Edge-Elmore.jpg is the cover of David Drake's The Sharp Edge, art by Larry Elmore. I still use this format from time to time, particualrly when I expwct the same image to be used on more than one pub. I suggest it as an alternate to the pub tag. it is much more human readable, and more searchable, and can be assigned beforfe the pub tag is. A number can be addded if the anem is not unique, for example if the same artist did different art for different editions of the same book, or for two different books with similar titles. I really dislike the apprently meaningless image file names the pubtag system provides, But if everyone else prefers that, we can adopt it as a standard, and change the wiki message and/or Help:How to upload images to the ISFDB wiki (which the wiki message should perhaps link to). If we do this, we really ought to give some thought to the share-image issue. -DES Talk 19:49, 7 October 2008 (UTC)
Publication Tags are NOT always unique, although they SHOULD be. I know from painful experience, I had to fix many "Doctor Who and the..." titles that all got a "DCTRWHNDTH" publication tag prefix. The result was several links in ISFDB going to the first Publication with that Tag, even if it was a completely different title. Editing Publication Tags is still possible, but I'd prefer to tell people to AVOID such unless they really know what they're doing - I believe Magazine editors have used, and maybe still do use, such functionality to get Standard Tags, but they don't have the "multiple printings" problems that book editors have. So although nobody has suggested it yet, can I please suggest that trying to guess the Tag in advance is not good. BLongley 21:20, 7 October 2008 (UTC)
I don't see a problem with using such and having them REUSED by other publications (which I think we'll get from cloning for instance): in fact it's a good pointer as to the original publication, when somebody notices that although a cover looks almost identical, the Canadian price changed, or the blurb from a review changed. When the art is the same, it's a good pointer and I'd prefer to see notes like "coverart is like this, but the prices are different" or "coverart is like this, but the blurb isn't 'Fantastic! from the Times', it's 'Terrific! from the Manchester Evening Post'. BLongley 21:20, 7 October 2008 (UTC)
The issue, IMO is that if use of pubtags is the standard, people will asusme that to get to the image page for the art on pub ABCDEFG2008 you go to Image:ABCDEFG2008.jpg. For pubs using cloned art, this will fail, violating the "principle of least astonishment". If there is no such standard, no one will make such an assumption. -DES Talk 22:00, 7 October 2008 (UTC)
I doubt they will. People will click on the image to get the bigger picture, they won't try and construct a URL from a Publication Tag immediately. Most people don't even notice that the Publication Tag is VISIBLE - it's right down at the bottom in case people want to add to "Bibliographic Comments: Publication:ABCDEFG2008". If they feel like trying for such, then the Tag used on the bigger image will still lead to the right place. But I really don't think most people will want to go from Pub to Image page, and I only add links from Image page to pub so I can find where I FIRST used it. I think that mostly all people want is "can I use the same image"? and they'll decide that from wherever they see it. The "principle of least astonishment" doesn't really apply as I for one would be Astonished if people really expected to be able to manually construct a URL to get to the Wiki page! Why do you assume they'd look for a .jpg rather than .jpeg or a .gif or .png for instance? BLongley 22:49, 7 October 2008 (UTC)
.jpg is the default extension for JPEG files created under windows, and thus is by far the most common form. Look at our filenames: we currently have 2 gif, 8 png, and no jpeg, and only 3 of the pngs are art, and those aren't cover art. All the rest are .jpg files at present.
As for who, I would construct not a URL but a wiki page name (which leads to a URL, but is shorter and more likely to be typed in, IME) based on an existing convention, I do so now for publisher and Bio pages. maybe no one else would, but I suspect that any wiki-experienced person might be likely to try this. -DES Talk 23:19, 7 October 2008 (UTC)
"Doing the upload first" is an obvious problem though, and one many non-self-approvers will encounter. (Thanks for pointing that out DES!) I think a gentle pointer to "choose a name that doesn't make your cover look like the DEFINITIVE cover, there WILL be other editions" is fine and suggestions that including a disambiguating publisher, imprint, year, printing number, or such is fine. Adding SOME notes with the upload will help searches anyway, although a warning that TLAs don't search well in the Wiki might be appropriate too. E.g. "1st NEL Ed." will not add anything to search results while "First New English Library Edition" might, even if the "New" is not going to be used. BLongley 21:20, 7 October 2008 (UTC)
I can agree with that. What do you think of the curent version of Help:How to upload images to the ISFDB wiki? -DES Talk 22:00, 7 October 2008 (UTC)
Better, but you already know what I think of the "Licensing" stuff. (For anyone that doesn't - I support only uploading LEGAL images, I just don't think requiring a justification for every individual image is worthwhile and is frankly discouraging uploads.) BLongley
Fair use images are perfectly legal -- the right to make fair use of published content is one fo the important rights secured by the copyright law. I might add that, given a skeleton template, {{Cover Image Data2}} requires exactly the same work as {{P}} does, and if it is the template name's length that bothers you, it would be trivial to create a shorter alias, say {{CID2}}. Whether having a rationale on each page is legally helpful or not is somehting that a lawyer would be needed to say with any certianty, my assumption was that following the practice of Wikipedia, where specialist lawyers have been consulted, was not a bad thing. -DES Talk 23:19, 7 October 2008 (UTC)
I'm with DES on the licensing tags. One of the significant issues for Wikipedia is that new editors tend to be unaware of just what the licensing stuff means and that over time the more senior editors have developed more awareness. The existing upload help page needs some copyediting. I don't have time to deal with that now. Marc Kupper (talk) 18:56, 8 October 2008 (UTC)

Image scaling

Should we have a standard for how large uploaded cover scans should be? The copyright and fair use law do not specify absolute pixel or percentage sizes. Amazon.com uses no more than 500 pixels but I have no idea if this has been tested in a court as valid for an across the board standard. It'll be easy to add a notice about the size as that can be done on MediaWiki:Uploadtext.

  • Marc Kupper has been scaling to no more than 500 pixels using Easy Thumbnails.
  • Ahasuerus uses 520 pixels.
  • DESiegel60 uses a variable size from 200 to 590 pixels though uses 575 for two covers. Image:Whisp_21-22.jpg is a cool cover.
  • BLongley uses 600 pixels for covers, using Paint Shop Pro Version 4.15 SE.
  • Gilgamesh uses 800 pixels.

I'm personally comfortable with 600 pixels as that usually makes it easy to read the small print on paperback covers. 800 pixels makes me nervous. Marc Kupper (talk) 15:40, 7 October 2008 (UTC)

Wiki image pages only show 600 pixel images so thats another reason to use that as a upper size. Dana Carson 19:46, 7 October 2008 (UTC)
I have generally been either using the image size as scanned at 96 dpi for covers (this low resolution is IMO plenty to handle the fair use issues), unless that is 600 or over in the larges dimension, in which case I scale down to something between 550 and 590, exact number pretty much on a whim. I see no benefit to scaling up. For signatures I scan at 800 dpi, crop tightly, and scale down if 600 or over, which they usually aren't. Logos i scan at more than 96, but not always 800, depending on the size and intricacy of the logo. Again I crop and scale down if over 600 pixels on the long access.
BTW, thumbnails of images over 600 pixels can be available if Al installs the ImageMagic library. He estimated some months ago that this would take most of a day of his time. By the way, I generally scale with the IrfanView program, or the editor that came with my Epson scanner. -DES Talk 19:59, 7 October 2008 (UTC)
Oh, the main fair use issue is that the resolution be lowered enough that a publication-quality printed copy could not be made from the image. Real publication quality generally means at least 1200 dpi, and a moderately convincing fake at least 600 dpi. For a Mass-market pb 600 dpi means an image of over 4,000 pixels in the larger dimension. Anything under 1000, even under 2000, will not be of printable quality at full size. Furthermore, the reduced resolution issue goes largely to the "marketplace harm" aspect of the 4-way fair-use test. Since there is usually no significant market for cover art images separate from the books they advertise (unless someone is printing and selling pirated books), quite probably even full-quality images would still qualify for fair use. My guess (and it is only a guess) is that Amazon limits image sizes not for copyright reasons, but because they are good enough for web display at the sizes they use, the storage saved means something when you have as many images as they do on file, and a larger image slows the page-load time, and might cause users to click-away before the page loads, which might hurt sales, the ultimate sin for a retailer. This was probably more important when more people were on dial-up, and in those areas of the world where slower connections are common. Besides, the difference of appearance at typical web display sizes is so small that it just isn't worth it to them to store larger images. Most of these reasons (making things fast for users, and reducing storage space) apply just as much to the ISFDB, so setting an image size of 500-600 pixels as a general standard seems like a good idea to me. -DES Talk 20:26, 7 October 2008 (UTC)
There is also the point that fair use standards only matter if the copyright owner sues, and a publisher is unlikely to want to sue Amazon for listing its books for sale. Amazon might decide not to list them any more. -DES Talk 20:28, 7 October 2008 (UTC)
I scan covers at 300dpi, crop, resize to 600 pixels high. Simple. I even add a P template link to the notes to remind me where I'm using it. (DES adds all the other stuff eventually.) BLongley 20:42, 7 October 2008 (UTC)
I recently added a help-link right on the upload page so I could remember all those handy DES templates and usually use both the "P" and the standard template. I suspect we could add the pub-tag to Template:Cover Image Data. Marc Kupper (talk) 00:57, 9 October 2008 (UTC)
I'm copying this down to the section on templates for better visibility, and responding there. -DES Talk 18:56, 9 October 2008 (UTC)
For signatures I scan at 1200 dpi, crop, zoom, go find the sig, crop again, rotate as required, zoom, crop, repeat and eventually resize to something that looks like a sig rather than a colour-blindness test and upload. Those I happily add the appropriate template to (which I can never remember and have to go find) as I want to build up the Artist Signature Library. Publisher Logos I do the same for, and add the template (again, can never recall it and have to go find it) but I suspect they'll be of less use. But feel free to go prove me wrong. BLongley 20:42, 7 October 2008 (UTC)
A reasonable approach. My scanner allows my to do a first crop on the preview screen (scanned very quickly at low res) so that for high-res scans like sigs, it only bothers to scan the area within the marked limits. This speeds things up a lot, it cuts the time to scan a sig from around 5 minutes to under 1. -DES Talk 21:51, 7 October 2008 (UTC)
I get the same sort of preview, but if it CAN then JUST do the bit I select, I don't now how to choose that option. Still, I had to wrangle with software installation that assumed I have a C: drive. (Lexmark programmers seem to have been a bit presumptive, even Windows doesn't HAVE to be installed on C: to work.) I have a sentimental attachment to this particular scanner (it's my 10th anniversary long-service award, and far more practical than the engraved glass triangle thingummy I got for being part of the "Team of the Year") but I'm open to acquiring a better one as well. Suggestions/Recommendations? BLongley 23:27, 8 October 2008 (UTC)
Part of the problem is much of the software that comes with hardware was written by hardware manufacturers. They may be great at engineering a peripheral but universally suck at writing software. I have a Microtek ScanMaker 3800 USB flat bed scanner that someone gave me. It works fine and replaced an ancient HP SCSI scanner that someone else gave me. I have no idea if the MicroText software can be installed on something other than C:\
  • For the most part the Microtext works fine.
  • I don't use the preview mode thing as it resets ALL of the scanner settings meaning I leave it permanently set to scan a bit more than paperback size at 300 DPI and I crop the results in Paint Shop Pro.
  • My main complaint with the Microtek is that it has a 1 1/4 inch border or frame around the edges of the glass. If I ever see a scanner with at least one borderless edge then I'll get it as that'll allow me to scan the inside of a book without cracking the spine. The lid floats and can remain flat up to about 1/2 inch and can easily be removed to scan thicker books.
  • 60 seconds to scan a pb at 1200 DPI plus an additional 15 seconds uploading over USB.
  • 20 seconds to scan a pb at 300 DPI including USB transfer time.
  • 15 seconds to scan a pb at 96 DPI including USB transfer time.
  • 10 seconds to take a preview of a pb though the default cropping it guesses at is frequently way off. Adjusting the crop is easy meaning someone can preview a book and then scan it to a file in less than a minute. One downside is that the crop is a blinking red/white box. It's easy to see but as you can't zoom in on the image it's also difficult to nail the crop right at the edge of something. The plus side is the software remembers the crop and other settings meaning if you mainly do pb books you can have a file ready to upload to ISFDB in 15 seconds.
  • I did a test of a cover for The Best of Leigh Brackett and it was about 3 minutes from getting the book off the pile to having it show in the ISFDB pub record but now that I can see the results I see I did not crop the bottom well.
  • I'll probably stick with my normal system which is I scan at 300 dpi, rotate if needed plus do the final crop in Paint Shop, and make a 500 pixel "thumbnail" for uploading to Amazon or ISFDB. As Paint Shop allows me to zoom in/out which the mouse wheel I and to move the image around with click-hold-drag I also use the cropping time to look at the cover for evidence of artist signatures. While I'm using three software packages (scanner, paint shop, and easy thumbnails) the first is a keyboard shortcut, and the latter two are available when I right click on a file with the thumbnail being a one-shot-no-questions-asked thing that creates the no-more-than-500-pixel image named tn_filename that can be uploaded.
  • If I see a signature then I can switch back to the scanner app, drag the crop box around where the signature is and scan that area at 1200 DPI. When it's done I drag the crop to be a bit larger than a pb. Marc Kupper (talk) 01:05, 9 October 2008 (UTC)

(unindent) A little off topic, but I noticed that some you are having problems with scanner software. I would like to recommend a third party software driver called VueScan from Hamrick software, http://hamrick.com/. I've been using it for several years now, ever since the manufacturer of one of my older scanners quit supporting it. The driver runs under both Mac and Windows operating systems, and supports over 1200 scanners, both flatbed and film.--Rkihara 02:41, 9 October 2008 (UTC)

Template Use

split off from above thread. -DES Talk 21:27, 8 October 2008 (UTC)

...And for those templates you find handy, you can (if you like) copy the skeleton call onto your user page, to which there is a link from every wiki page (when you are signed in). This will make looking for them quicker. -DES Talk 21:51, 7 October 2008 (UTC)

You haven't checked how long my user page actually IS, have you? :-( BLongley 22:58, 7 October 2008 (UTC)
No I haven't. But at least one user has created a sub-page of his user page jsut to hold such skeletons in an easily accessible place. -DES Talk 23:03, 7 October 2008 (UTC)
I consider such an indicator of a failure in our HELP. I can never find the templates easily. I use a few that I've learnt, would use a few more if they were easily found, and don't think "oh, once you've found them you can put them somewhere you'll easily find them again" (when I actually WON'T, due to having to keep loads of other stuff there if I want it here at all) a particularly helpful response. BLongley 23:49, 7 October 2008 (UTC)
Fair enough, this may be due to my being to used to working on a wiki, where the numbers of useful, indeed essential, templates was in the hundreds. How and where do you thing the help should be improved to make it easier to find and use the templates we have? The upload page (and the Help:How to upload images to the ISFDB wiki page both link to Category:Image License Tags, which includes links to every template to be used on an image page, and each of the templates displays as part of its documentation a skeleton call, designed to be copied and pasted. How can the be improved on? What would make such things easier for you to find/remember and use? What would make other templates easier? -DES Talk 00:02, 8 October 2008 (UTC)
Putting the useful templates (not the Image License Tags, things like Title and Publication) on Help:Editing would make them quicker to access from where I actually use them. Don't list them alphabetically, list them in order of usefulness. Keep the names short, and the list short: I can cope with TWO clicks to get to the little-used ones. BLongley 19:34, 8 October 2008 (UTC)
See Help:Editing#Templates and tools. Is that something like what you had in mind? Help:List of ISFDB Templates is intended to be a more comprehensive list -- I'm still working on it. -DES Talk 21:27, 8 October 2008 (UTC)
Yes, you're moving in the right way. P, T, A are good. (You've got A mixed into T at the moment but I expect that'll be fixed before I wake up.) The specific CID stuff is of no use on general Wiki pages though, so leave that on the comprehensive list. BLongley 23:05, 8 October 2008 (UTC)

Tags in Templates

Marc's post copied from the "Image Scaling" thread. -DES Talk 19:03, 9 October 2008 (UTC) I recently added a help-link right on the upload page so I could remember all those handy DES templates and usually use both the "P" and the standard template. I suspect we could add the pub-tag to Template:Cover Image Data. Marc Kupper (talk) 00:57, 9 October 2008 (UTC)

Yes, we could. What we can do without changing the template is to include a P tag in one of the parameters. i have been using the edition parameter for this, look at Image:Space Infantry-Velez.jpg as an example. I could change the template to add a Tag parameter, but without having the IF parserfunction available, it would be a little hard to make it look well when the tag was not supplied, and making the tag required would mean entering it for the 300 odd cover images now on file, and requiring future users to supply it in every case. I can do that, but I don't want to make the template any more daunting than it already is. Suggestions, please? -DES Talk 19:03, 9 October 2008 (UTC)
Finding where an image is used is already a double-step check in the wiki, and finding its use in the database is impossible online - which is why I do use the P template. I could live with it being a very strong recommendation at least, and if the recommendation is accomplished via "it'll look awful if you don't supply the publication Tag or ID" that's OK by me. I'm NOT volunteering to back-populate our current covers though. BLongley 18:43, 10 October 2008 (UTC)
I'll try a draft set of changes on {{Cover Image Data}}. Is there any chance that you would look at {{CID2}} for use on your future uploads? It has a fairly short name, and uses the same arguments as {{P}} does. I urge you to consider it.-DES Talk 19:47, 10 October 2008 (UTC)
I'll give it a go. But why isn't it just a C template yet? Is C already used for something useful? BLongley 20:50, 10 October 2008 (UTC)
I don't really like single letter names for things, i find it easier to remember ones that are a little longer and more specific. But I will create {{C}} as an alternate alias, OK? -DES Talk 20:59, 10 October 2008 (UTC)
I seem to have remembered to use it successfully for a couple of dozen uploads today, so thanks for that. BLongley 23:13, 11 October 2008 (UTC)
I see (or C) you did. More than 4 dozen, actually, if I am counting correctly. Category:Cover images now has 400 items in it. Thank you very much, both for the covers and for using {{C}}. It will save me some work, later, and I do think it is a good thing. -DES Talk 23:30, 11 October 2008 (UTC)
Of course, adding a tag parameter will not let you find multiple uses in the db, when there are such, directly. But most cover images are used on only a single pub record, anyway. -DES Talk 19:47, 10 October 2008 (UTC)
I think you underestimate the amount of cloning going on. My Amazon covers appear on a lot more publications than I suspected. But maybe we've got most of the reprints by now. BLongley 20:50, 10 October 2008 (UTC)
4% of the publication records share images and if you look at only publications with images then 9.4% of the publications share an image with at least one other publication. At present there's some 2425 images that are shared by 5110 publication records. These images are shared by 10 and 7 pubs respectively: [5][6][7][8][9][10][11][12][13][14] and [15][16][17][18][19][20][21]. Marc Kupper (talk) 01:43, 12 October 2008 (UTC)

First example changes

Please look at three example pages where a new (draft) version of {{Cover Image Data}} is in use. If this meets with approval, I will copy it over the existing template, and adjust the documentation as needed. The three examples all use the data from a cover I uploaded last night.

  • User:DESiegel60/CID-test1 A page where a publication record number is specified
  • User:DESiegel60/CID-test2 A page where a publication tag is specified instead
  • User:DESiegel60/CID-test3 A page where neither record number nor tag is specified. This is what existing images will look like until/unless edited to insert the rn or tag, and what new images will look like if the pub rn or tag is omitted.

Please tell me what you think of these examples. -DES Talk 20:59, 10 October 2008 (UTC)

Since we are getting more scans uploaded to the Wiki, linking them back to the Publication records which use them sounds like a good idea, but there is nothing that would prevent the two from getting out of sync, is there? Ahasuerus 02:32, 11 October 2008 (UTC)
There is no way that I know of to make such a link automatic. If an image is used by multiple pubs, the feature I have drafted would record only one, presumably the pub from which the image was actually scanned or otherwise taken. I suppose it would be possible for an SQL script to scan a db backup for all image URLs with the ISFDB.org domain, and construct a table matching pub record numbers to image file names, which could be put on a wiki page, but that would be a fair amount of work, I suspect. It seems to me that while it is not uncommon for an image to be used on multiple pub records (vis cloning or other copying) it is rare for the inital pub record on which the image was used to be deleted or hav its tag or record number changed. Am I correct in that? Because only such a removal of pub records (or a manual change to the image's wiki page) would make them really out-of-sync, although cloning cvould easily make the link incomplete in that it will point to only one of multiple uses. If Feature:90159 (Link locally hosted images to their wiki description pages) is implemented (which i would think not overly hard) it will help with links the other way. Do you want a list of all pub records that use a given image? Such a list could be created manually as images are added to pub records, but there would be nothing to indicate if an editor failed to update it when s/he should have. -DES Talk 02:49, 11 October 2008 (UTC)
I do delete duplicate Publication records reasonably often, which will affect Wiki side images sooner or later. I guess as long as we are using two manually linked applications (the database and the Wiki), they are liable to accumulate discrepancies over time as Publication, Title, Author, Publisher, etc records are deleted, merged and otherwise mutilated. Not much we can do about it, but it often makes me hesitant to add anything permanent to the Wiki rather than to the database. Oh well! Ahasuerus 04:09, 11 October 2008 (UTC)
Well, when I delete a duplicate I make sure that any useful info on the duplicate is carried over to the remaining publication, so I doubt we'd lose all links: the Wiki to original database publication link could be broken but the database to Wiki ones wouldn't be. Of course, if I was deleting Manga or RPG then there'd be no obvious indication that there's a stray image left on the Wiki unless you check the Image properties. But I can live with extra unlinked Wiki entries, particularly if they're ones only added to provide a coverart entry for a publication. I can't really see anyone trying to get BACK to the wiki entry from a publication coverart entry (OK, DES says he would (or is it just "could"?), but I can't see any reason to do so unless it's for Wiki-image-credit policing stuff.) BLongley 23:13, 11 October 2008 (UTC)
I think the main issue being discussed above is possible loss of wiki->db links, which is exactly what you say could occur. When deleting pubs, looking at the image URL just enough to see that it is (or is not) on the isfdb.org domain tells you if there is a possible orphan image. Unfortunately, because of possible cloning, there is no easy way to be sure if an image is an orphan (not used on any DB record) unless a query looks for the image file name through the entire db. -DES Talk 02:24, 12 October 2008 (UTC)
I think the wiki->db links loss is not an issue. I can't see many people stumbling across a wiki image page and wondering what this "ISFDB" thingummy is. Losing the db->wiki link breaks image displays in the online ISFDB, but we already know it's going to break when using the ISFDB offline unless we start including the images in the backups. I don't want to link the database and wiki too closely: it would make it non-transportable. Images are nice (that's why I add so many) but the loss of the wiki side doesn't affect the bibliographic data in the database: just (at worst) removes justifications or evidence for some of the entries. BLongley 20:46, 12 October 2008 (UTC)
If you really think it isn't an issue, why did you think it was important to insert such links in the first place. These are precisely the links that the current draft changes are designed to create. As to how they would be used, no I don't see someone stumbling across the covers on a random google search or whatever, and not knowing what the ISFDB even is. oh it's possible, but that isn't the case i care about, nor do i think it will be the common case. What I do think likely is that an ISFDB ediutor who uses the wiki finds an image page, parhaps via the artist or publisher categories, and wants to know what publicatuion the image came from. But if the original pub has been deleted or merged, there is no way to know where the image is being used, or what pub record describes the pub it was scanned from (or copied from elsewhere for) unless the editor who deltes or merges the pub record also updates the wiki-page, and that may not be a reasonable expectation. -DES Talk 23:54, 12 October 2008 (UTC)
Did you look at the examples lined above? Do you think they are improvements on the current template, or not? Or improvements, but not sufficient? -DES Talk 02:49, 11 October 2008 (UTC)
They look reasonable enough. I haven't done a whole lot with cover image pages (aside from creating them) so far, so I am not the best judge of their usefulness, though. Ahasuerus 04:09, 11 October 2008 (UTC)
They all look pretty awful to me, but no worse than before. I think if we HAVE to have all those words about justification I'd like to lose the colours for all but the link back to the publication. That's the only bit I find useful, and the pinks and blues and purples don't help me. (We only chose that horrible green for ISFDB links to make people complain and get it improved, we seem to have failed in that. :-( ) BLongley 23:13, 11 October 2008 (UTC)
I have come to rather like the green, actually, but if you have a better suggestion, by all means share it. The other colors are cloned from the corresponding templates on wikipedia, and serve to distinguish the information that identifies the work from the justification for using it. Still they are not essential, and if others besides Bill dislike them, they can go. -DES Talk 23:30, 11 October 2008 (UTC)
The part I particularly wanted comments on was the new Publication Link field, and how it looks both on linked (examples 1 & 2) and on unlinked (example 3) uses. Everything else about the template is unchanged. -DES Talk 23:30, 11 October 2008 (UTC)
The green on "Resolution reduced?" really should go as it seems to be the same green as for links to ISFDB. The colours on Publication Link version 3 do seem to be even more horrible than our green, but is lost when we have it surrounded with other horrible colours. I think what I'd like is the little "arrow" graphic for links to ISFDB to vary depending on the type of link (Title, Publication, Author ((Do we have a Series link? If so, that's missing from the help so far)) ) but I'm a lousy artist. BLongley 23:56, 11 October 2008 (UTC)
Now I look closer (I've only tried "C" today) I see they actually all miss showing the TITLE as a clickable link. I'd like that to be maintained. BLongley 23:56, 11 October 2008 (UTC)
I would like the title to be a clickable link, but I can't make that work unless eaither a) Al installs the PaerserFunctions extension so I can use an IF in the template, or b) we backfit every instance of the template, and are confident that it will never be used without a record number or tag being specified. Without an IF function, if I set it up so that the title is clickable, no title will be displayed if the rn or tag is ommitted.
On second thought, ther may be a way to make that work, I'll try tonight or tomorrow if I have time. -DES Talk 00:31, 12 October 2008 (UTC)
On the colors: I can't control the "arrow" separately from the link text, as the wiki software generates the arrow automatically. I can make the green for the "resolution reduced" different, and i will. I could vary the colors for Title, publication, author, and sereis links to the ISFDB, at the moment we have one color that means "link to the DB", but that could change if we wish. We do have {{Series}}, which is included in Help:List of ISFDB Templates. I didn't add it to Help:Editing in the interest of keeping that list short, I only included the most often used of the linking and image templates. -DES Talk 00:31, 12 October 2008 (UTC)
I wonder if the full Fair Use Rational needs to be in the template or if we can just mention the image is used under fair use with that linking to a page with the details. Also, if we ask that all images be under 600 pixels can we default the "Resolution reduced" question to Yes. If someone uploads an image that's larger than 600 pixels it won't scale and we can then decide to change the resolution to "No" or to re-scale the image so that "yes" qualifies. Marc Kupper (talk) 05:13, 12 October 2008 (UTC)
There are different rationales for different types of images, and for some types (author image, for example), different rationales for each image. Also, anyone viewing these images on our wiki ought to see the rationale, because it helps indicate why the image may not be reusable. I think the rationale should stay on the image pages. -DES Talk 11:27, 12 October 2008 (UTC)
Going back to the "Cover Image Data" issue alone (even I agree that other uses might need more work), I'd still prefer that there is less wording on the resulting page when a CID template is applied, and I like Marc's suggestion that we could link to the rationale and reduce such. There should probably be a sort of cascade too - anything we currently use under "Fair Use" will eventually need to be reclassified under "Public Domain". (Which is one of the reasons why I don't like stamping every image with ONE justification.) But the people finding such pages are IMO unlikely to be ISFDB editors and can go find their own justifications, it's not our problem. We've got nothing in the ISFDB software that encourages work on the wiki images. BLongley 21:03, 12 October 2008 (UTC)
While the precise wording might be debated and changed, i really don't think we could or should cut the current wording down by too much. Linking is possible, but at the least we would need multiple pages to link to, and what is the advantage. It is not as if the template wording is stored multiple times. The point of having it there is so that users, and particularly potential reusers, might possibly read it, and slso so that copyright holders, if any vist, see the argument immediately adjacent to the image. -DES Talk 23:30, 12 October 2008 (UTC)
You say "people finding such pages are IMO unlikely to be ISFDB editors and can go find their own justifications, it's not our problem. But I think to some extant it is, morally if not legally. ISFDB:General disclaimer#Copyright says "All original content of the ISFDB, and of the ISFDB wiki, is released under a Creative Commons License". This invites reuse. IMO this means we ought to notify users what pages are not origianl content subjet to free reuse, and indicate what the conditions of use are. This is IMO an additional reason for having the fair use information, including attributions and rationale, right on the image description page. We do already link to a page on fair use. -DES Talk 23:30, 12 October 2008 (UTC)
As to "...anything we currently use under 'Fair Use' will eventually need to be reclassified under 'Public Domain'" yes this is true, eventualy, but not soon. None of the works we now use under Fair use can possibly by PD before 2018, and that is for works published in 1923. We don't have many book covers from 1923. Works from, say, 1945, will not be PD until 2040. By that time we can probably develop tools to handle the conversions. (In general any work published in the US from 1923 through 1977 will be in copyright for 95 years from the date of publication, provided there was a copyright notice and a timly copyright renewal.)
You say above: "I don't like stamping every image with ONE justification." Would you prefer that wew do as wikipedia does, and require an indivicually created justification for every use of every image? Part of why they do that is that they use images for a wider variety of purposes than we do, and evben so the "individual" justifications are largely created via complex parameterized templates, but some individual manual writing is still required for every image there. -DES Talk 23:30, 12 October 2008 (UTC)
I re-read the rationale for the sample templates and agree that pretty much all of it should be there. Is this your own wording or copied from another source such as Wikipedia? The reason I ask is if "...confirm the accuracy of such bibliographic data..." could also have some wording about the painting or artwork itself added - "...confirm artwork and the accuracy of such bibliographic data..." Marc Kupper (talk) 19:14, 12 October 2008 (UTC)
I started with the text from Wikipedia, and altered it to fit the ISFDB's situation. There is no reason that I can see not to make the change you suggest. -DES Talk 19:18, 12 October 2008 (UTC)

Second example changes

In response to the comments above I have devised a new test version of {{Cover Image Data}}. In this version, ther is no separate Publiction Link field. Instead the "Edition" display, which comes right after the title, is made a clickable link. (I used edition rathe than Title because the link is to a publication record, not a title/bibliography record.) The three exampels use the same data as the ones above, and again the third example shows what happens if no publication tag or record number is provided. I also changed the color of the "resolution reduced" field so that it is distinct from the link color (it now uses a significantly deeper green).

Please look at three example pages where a 2nd new (draft) version of {{Cover Image Data}} is in use. If this meets with approval, I will copy it over the existing template, and adjust the documentation as needed. The three examples all use the data from a cover I uploaded recently.

  • User:DESiegel60/CID-test1a A page where a publication record number is specified
  • User:DESiegel60/CID-test2a A page where a publication tag is specified instead
  • User:DESiegel60/CID-test3a A page where neither record number nor tag is specified. This is what existing images will look like until/unless edited to insert the rn or tag, and what new images will look and act like if the pub rn or tag is omitted. Note that the link apears, but goes to an invalid record.

Please give me your opnions of the merits (or demerits) of these changes.

One comment is when you completely rearrange a talk thread to please do the rearrangement, save it, and then add new text as a separate edit. At present it'll take a fair amount of work to understand what's "new" that needs attention and commenting vs. the needed reorganizing the thread to keep the bloat down. Marc Kupper (talk) 19:03, 12 October 2008 (UTC)
I inserted some secondary section headings and reduced indents accordingly, and added some text at the end. No text was moved or inserted in the middle of the thread, the only new text is at the end, specifically the "Second example changes" sub-section. no substantive changes were made before this subsection. In fact, i didn't even do spelling corrections in earlier text. -DES Talk 19:13, 12 October 2008 (UTC)
The adding of section headers, unindenting, spelling correction of "iamges" to "iamges" in the earlier text, and addition of the Second example changes confused http://www.isfdb.org/wiki/index.php?title=Rules_and_standards_discussions&diff=131222&oldid=131213 well enough. Anyway, I see that this section is the only new part and the rest was organization. Marc Kupper (talk) 19:26, 12 October 2008 (UTC)

(unindent as I think I can be on topic now...) test3a confuses me. Is it possible to not have a link? A link to an invalid page, particularly one highlighted in green, seems like a bad thing. Marc Kupper (talk) 20:02, 12 October 2008 (UTC)

No, not unless the ParserFunctions extension is installed, which gives me the IF functuon. That was the reason for test3. If I add a link, that link will always be there, whether we have the info to create a valid link target or not. Even test3 has an invalid link. I can default the link text to something nasty and in red, as in test3, but not if the link is to normally be on the edition display, and we still want to show the edition even if the link is invalid. I could put the link on the edition and have a nasty red default text simialr to the one in test3, but then it would replace the edition in such cases. (The consequences are perhaps even more severe if I put the link on the title field instead.) Is that desireable? -DES Talk 23:39, 12 October 2008 (UTC)

Making choices

There have been no comments on the various examples for several days. The basic options, as I see them are:

  1. Retain {{Cover Image Data}} as it currently is, making no changes. Links to the pub record can be inserted manually, but are not automatic. The documentation can be changed to encourage such links and describe how to make them.
  2. Use the "first draft" version, the one displayed in the test1, test2, and test3 examples above. A new "Publication Link" field would be added, which would contain a valid link to a pub record if the proper publication record number or tag is supplied, and an invalid link with a warning message otherwise. Other X Image Data templates could be changed to match, where appropriate.
  3. Use the "second draft" method, (the one in the test1a, test2a, and test3a links above) where a link is provided on the "edition" field (or, if preferred, on the "title" field). This link would go to a valid record if the proper record number or tag is provided, and to invalid record number 0 if no rn or tag is provided. There would be no obvious visual indicator of an invalid link. This would make it more desirable to back-populate the existing hundred or more uses of the template with the proper record number or tag.
  4. Wait for the IF functionality to be installed on our wiki. Marc left a message on my talk page suggesting that this might happen in the foreseeable future. This would allow a link similar to that in the "second draft" method, but with the link disabled and perhaps a visual indicator if the parameter is not supplied. But there is no telling (or at least i can't tell) when or if this would be available.
  5. Do something different than any of the above.

Note that if we go with either #2 or #3 above, converting to number 4 when and if the IF extension is installed would not be hard, nor would any data be lost.

Which of these 5 options would people prefer? -DES Talk 21:58, 15 October 2008 (UTC)

Resolution Reduced part of the template

I need to look at the other thread about the puke green background for the "Resolution reduced?" thing... Marc Kupper (talk) 19:34, 12 October 2008 (UTC)

I re-read the thread and believe there's no need to color code the "Resolution reduced" text at all. As I noted earlier, it could default to "Yes" and now see that we could eliminate this as "Resolution reduced" is subjective. I'm looking at a random Amazon image and see that a college student could crop it and post the image on their own web site/blog as an example of their "beer drinking mindset" or that it could be cropped and sold as a refrigerator magnet. The thing, we as moderators can do, is to ask, and enforce if needed, that images be under 600 pixels which seems to be the consensus. Marc Kupper (talk) 19:59, 12 October 2008 (UTC)
I can live with removing ALL colours that aren't on links (and have suggested such). But I'm a bit confused that moderators are now suggested as policing the wiki as well as ISFDB edits? BLongley 21:08, 12 October 2008 (UTC)
Moderators are also wiki-sysops, and as such do whatever policing the wiki needs. I have been policing vanity bios, and indeed bios in general, for example. -DES Talk 23:44, 12 October 2008 (UTC)
As to colors on the "Resolution Reduced" section: I can default it to yes if we wish, and make the Yes text colorless. (I'd like to retain the red alarm color on the NO version, which should be rare). My intent ws to require each uploader to make a positive statement about reduced resolution, but if we are limiting the image size to 600px or less as a policy, I agree this is somewhat redundant. When I devised this part of the mechanism, i didn't yet know that the automatic rescaling that Wikipedia and some other wikis have wasn't installed here. -DES Talk 23:44, 12 October 2008 (UTC)
Since I have shell access to the virtual server that runs the Wiki, I can easily search for image files over, say, 200Kb in size and then post a list here for follow-up. I can even make it a weekly job and have the results e-mailed to whoever wants to volunteer to police this area. Ahasuerus 23:42, 12 October 2008 (UTC)
Here are the 200K+ files as of a few minutes ago:
sizedatelocationcomments
289019Jul 10 21:39BKTG05370.jpgOk - It's a complex picture that resulted in a large file.
232325May 13 18:26Asfdec2006.jpgOk - It's a complex picture that resulted in a large file.
292518Jul 6 14:36Galaxy No. 3.jpg774px × 1146px image that should be scaled to 600px and re-uploaded.
309874Jun 3 07:03Analog june2003big.jpg639px × 1000px image that should be scaled to 600px and re-uploaded.
258894Jul 20 01:31NVDGSSW1967.jpgOk - Both halves of an Ace double.
1863748Jul 21 19:01280-8016 IMG.JPGMarc Kupper uploaded this as part of a talk thread somewhere. As the copyright has expired it's not a copyvio but I'm not sure if we need the image either.
252767Jul 6 14:05Galaxy No. 1.jpg770px × 1128px image that should be scaled to 600px and re-uploaded.
289544Jul 21 03:43SHSNDSTRSP1978.jpg629px × 1754px image that should be cropped, scaled to 600px and re-uploaded.
230725Aug 30 14:40WDNSTR1968.jpg638px × 1048px image that should be scaled to 600px and re-uploaded.
322796Jul 21 03:49XLWTNGP1975.jpg694px × 1754px image that should be cropped, scaled to 600px and re-uploaded.
248539Jul 19 16:22The Man Who Folded Himself.jpg892px × 1348px image that should be scaled to 600px and re-uploaded.
279449Sep 6 10:47Mysteria copyright page.jpgMarc Kupper uploaded this as a source-record for a rather complicated copyright page. I'm undecided on if scanning this page in its entirety and at a resolution that someone can easily read the text is a copyvio. I'll probably delete this at some point.
1860761Jul 21 19:04280-8017 IMG.JPGMarc Kupper uploaded this as part of a talk thread somewhere. As the copyright has expired it's not a copyvio but I'm not sure if we need the image either.
207747May 13 18:32Asfjan2007.jpgI think this image hit the radar as it's got progressive encoding to speed up web display. We can reduce it from 203K to 137K bytes by removing the progressive encoding and can knock it down to 41K with 20% JPEG compression as the image is fairly clean.
227601May 13 18:510hucyw89lr62kfgzgn8iz21bkye8vm3.jpg472px × 751px deleted image that should be scaled to 600px and re-uploaded.
235688May 13 21:01Asfdec2007.jpgOk - It's a complex picture that resulted in a large file.
Ahasuerus 21:33, 15 October 2008 (UTC)
I converted Ahasuerus' table into a wiki table with links, and added comments about each image. Marc Kupper (talk) 17:36, 22 October 2008 (UTC)
Four (SHSNDSTRSP1978.jpg, WDNSTR1968.jpg, XLWTNGP1975.jpg, The_Man_Who_Folded_Himself.jpg) have been rescaled and uploaded - the images you're pointing at are in "temp" or "archive" directories that I've no idea how to clean up, if we even can. BLongley 19:48, 22 October 2008 (UTC)
I have shell access to the server, but I can't delete files in the "images" directory tree, so we'll need to ask Al what he wants to do about this. Ahasuerus 20:14, 22 October 2008 (UTC)
Galaxy No 1 and No 3 are obviously mine, but I can't see them actually used in ISFDB. I suspect I reloaded under a Publication tag name but I can't even find the magazines they're applicable to. BLongley 20:37, 22 October 2008 (UTC)

(unindent) This issue applies identically to the current template, and to the first and second example draft templates--it is identical in all three versions. i have therefore moved this section so it is not a sub-section of the "2nd example" discussion. -DES Talk 21:43, 15 October 2008 (UTC)

  • I believe we have a consensus for 600 pixels to adopt as policy. The lack of the image scaling library is convenient in this respect.
  • There was not as much interest in the color of the text. I brought it up as the design guides advise limiting the colors and that it seemed unnecessary to highlight LowRes=yes. Presumably if we ever have a LowRes=no then it should be a free-use image (copyright expired, etc.) and in that sense it's not clear if the lowres notice needs to be highlighted. I suspect it's going to be a very rare publication cover, or any image on ISFDB, that "needs" to be high resolution other than artist signatures. Marc Kupper (talk) 17:50, 22 October 2008 (UTC)
600 pixels works for me. Ahasuerus 20:16, 22 October 2008 (UTC)
600 fine by me too. BLongley 20:37, 22 October 2008 (UTC)
MediaWiki:Uploadtext and Help:How to upload images to the ISFDB wiki updated with 600 pixel limit. -DES Talk 20:44, 22 October 2008 (UTC)

Third example changes

I have just created a version of the template using conditional logic. This venison puts a link on the edition label if and only if the Rn or Tag parameter is supplied. Please look at three example pages where a new (draft) version of {{Cover Image Data}} is in use. (they use the same data as the examples above.) If this meets with approval, I will copy it over the existing template, and adjust the documentation as needed. The three examples all use the data from a cover I uploaded last night.

  • User:DESiegel60/CID-test1b A page where a publication record number is specified
  • User:DESiegel60/CID-test2b A page where a publication tag is specified instead
  • User:DESiegel60/CID-test3b A page where neither record number nor tag is specified. This is what existing images will look like until/unless edited to insert the rn or tag, and what new images will look like if the pub rn or tag is omitted.

Please tell me what you think of these examples. -DES Talk 07:19, 22 October 2008 (UTC)

This looks good with some observations being.
  • I don't think the Rn parameter needs to exist as someone can use "Tag=1234" and it'll work. It may be better to call the parameter “Publication.” I suspect I’m not the only one that routinely uses {{p}} with both tags and record numbers depending on whichever value I have handy.
  • When LowRes is "yes" I don't think the green background is needed but agree with the red background or better, a yellow background because that red does not stand out well from the pink of the Fair Use Image Rationale. I'm sure Bill will cheer css3's reintroduction of the marquee tag.
  • Incidentally, I noticed in the examples, you used "|LowRes=yes". I was not aware that "yes" with a lower case Y was supported as the doc page states "Yes" or "No" otherwise a malfunction will occur. Is using "yes" going to be a permanently supported feature or a malfunction that happened to do what you were thinking the page should do?
  • Down in the bottom where you have "specific edition of ..." can you italicize the title? Marc Kupper (talk) 16:57, 22 October 2008 (UTC)
The functionality appears to be fine and I will leave the aesthetics to our color-enabled editors :) Ahasuerus 17:37, 22 October 2008 (UTC)
  • I can combine the "Tag" and "Rn" parameters into a single parameter, say "Pub". I thought it would be easier to document and understand if they were separated, but the record number and pub tag are, in effect, alises's for each other. This would also simplify the tempalte code a bit. Ok, I'll do that.
  • I'll change the red background to yellow, and make yes the default as discussed above. I'll change the documentation to suggest that users ommit the LowRes parameter unless the value is No, while remaining compatible with existing calls.
  • LowRes is implemented by constructing the anme of another tempalte to call. I had thought that meant the parameter value would be case sensitive. and i had not attempted to change that -- if I used "yes" insted of "Yes" that was an oversight on my part, and if it worked that was unexpected. I'll look into the matter further.
  • Yes I will italicize the other mention of the title.
Thanks for your responses. -DES Talk 18:38, 22 October 2008 (UTC)

Publisher edits: Location prefixes

I would like to propose that it be considered acceptable to edit (in the DB, en mass) publisher (or imprint) names of the form "City:Publisher" (for example "London: Blond & Briggs" or "New York City: Tower Books" or "New York & London: Longmans Green") to remove the city prefix, leaving just "Publisher", provided that the city designation is added to the publisher's notes field, and optionally to its wiki page (if it has one). If there is an existing record for "Publisher", the two may be merged. If needed for disambiguation, a country (or other location) suffix may be added so "London:MegaBooks" might become "MegaBooks (UK)" while "New York:MegaBooks" became "MegaBooks (USA)". Such changes may affect verified publications without checking with verifiers. Does anyone agree with this idea? -DES Talk 22:28, 7 October 2008 (UTC)

Oh, and along with this would go a comment in the relevant help to avoid such city prefixes. -DES Talk 22:29, 7 October 2008 (UTC)

I'm definitely against such "en masse" edits without MUCH thought, and I'm REALLY against affecting verified publications without checking with active verifiers and editors. If people are trying to rush such changes through, I would prefer removal of the powerful tools Mods have for such until they can be re-introduced with "can't be done against verified pubs" checks (which Mods can change, but only one-by-one, so should be a disincentive), or even DE-Moderatorisation of Mods that might use such. BLongley 23:39, 7 October 2008 (UTC)
This was a proposal. I would not act on it until and unless there was a clear consensus that such action was acceptable. Was that not clear? If (and only if) such a consensus develops, than such edits would be proper. I would be happy to debate the merits of the proposal, and that debate will, i trust, include enough thought. -DES Talk 23:55, 7 October 2008 (UTC)
The "Such changes may affect verified publications without checking with verifiers." last line is a BIG issue for me, and I don't like to see it tucked away as almost an afterthought. It should be a separate, and frankly, more important issue than the rest of your suggestions. BLongley 00:28, 8 October 2008 (UTC)
To my way of thinking, any proposal on an acceptable way to use the Edit publisher function will include that line. Publisher edits that we can't agree can be done without pub-by-pub approval are, IMO, publisher edits we can't agree should be done. it will be a part of any and every proposal I make for what can be done with the publisher edit function. -DES Talk 00:44, 8 October 2008 (UTC)
I suspect a lot of the dual countries like "New York & London: Longmans Green" are actually referring to TWO publications and so should be split into such. "Blond & Briggs" and "Blond and Briggs" could probably be merged. Not sure which way, but so long as the Wiki pages are kept in step (and if you CAN start a Wiki page or add publisher notes, DO SO!) I'm OK. But deal with "&" versus "and" as ONE separate discussion or we'll get nowhere. (Again!) BLongley 23:39, 7 October 2008 (UTC)
I rather suspect not. I am pretty sure that a number of publishes, having developed an international presence, routinely placed "New York & London" on all their books for a time. I have even seen lists like "New York, Toronto, London, Paris, Warsaw, & Sydney" and I doubt very much that this indicates 6 separate publications. In some cases a publisher that had both US and UK offices published separate editions of books in each country, but not usually of all their books, and I don't think that the existence of such dual pubs can be at all reliably inferred from "New York & London" on the copyright page. -DES Talk 23:55, 7 October 2008 (UTC)
I generally agree that the city/country does not need to be included in the publisher field but I'm not going to remove it when approving someone's edit.
If we should find two publications that are identical except one says "New York, London" and another says "New York, London, Sydney" then it's two separate publications as far as I can see. Thus, I don't mind that someone may include what seems like "unnecessary information" in a publication record as that information may well be a distinguishing point some day. All I hope and ask for is that when someone enters something into a publication record that it accurately reflects what's in the source.
To simplify some of the database work we have normalization rules for all of the publication fields except Publisher. If someone wants to start a thread that proposes normalization rules then I'm all for it. However, I'm against going back and modifying existing publication records other than by physically inspecting them and ensuring that the publisher field follows the normalization rules. Thus I'd like any rules related to the publisher field would be from the time of adoption and moving forwards one publication record at a time. Marc Kupper (talk) 19:47, 22 October 2008 (UTC)
The ONE thing I'm instantly in agreement with is that City prefixes should be avoided, and help should be updated. I know this will upset a lot of people entering data from Library catalogues, but they should learn that Libraries aren't doing it right in the first place. BLongley 23:39, 7 October 2008 (UTC)
I enter a lot of pubs from OCLC records, which often use this form, and it won't upset me. -DES Talk 23:55, 7 October 2008 (UTC)
I think Ahasuerus is another big user of OCLC, and probably responsible for more entries than you've managed so far. I'm not sure when he'll have time or health to respond, but there's some trivial issues ("Spaces around slashes" for instance) we can deal with in a week or two (most editors can respond in that time), and some we'll take months over (non-reversible edits). Your proposal is one that SHOULD take months, if needed. BLongley 00:28, 8 October 2008 (UTC)
If we need to take months discussing whether to approve such edits, fine. I set no deadline for discussing this matter. I'm not arguing for starting such edits tomorrow (although i have done such edits without consultation in the past, and would have continued had it not become plain that clearer consensus on the acceptable use of the publisher edit feature Wes needed). Not until it seems clear that all active and untested editors have had plenty of time to respond, and a consensus has developed among those who have responded, would i expect to put this into active use. -DES Talk 00:44, 8 October 2008 (UTC)

(unindent) I see all kinds of interesting points raised over the last couple of days -- which I will comment on as soon as I defeat these microscopic invaders -- but it occurs to me that it may attract more contributors if we could somehow change the tone to be a little more relaxed. Tea and cookies, anyone? :) Ahasuerus 02:19, 8 October 2008 (UTC)

I am sorry if my tone seemed hostile or tense. These are significant issues, and I want to discuss them seriously, with the understanding that we are all people of good will, and all want the ISFDB to go forward in the best possible way. I don't want to jump into doing things unwisely. I want to start discussions on these issues so that we can get to a resolution after appropriate thought, discussion, and perhaps changes of mind. I don't assume that everyone will simply and automatically agree with me, nor that I will never change my views. I merely intended to put forth a limited proposal, which i hope people will agree with, either as proposed or after discussion and change. Failing that, perhaps the discussion will lead to some other useful agreement on what we should and should not do with the edit publisher record feature. I have definite views on how I think it should be used, but they may be wrong (or at least unwise) or not welcome to others. I intended the proposals to be positive, forward-going, and polite, and to invite discussion and differing views from any and all editors. -DES Talk 17:01, 8 October 2008 (UTC)

(unindent)As far as "city" prefixes go, keep in mind that library databases typically have at least two separate fields for publisher information: "place of publication" (field 260a) and "name of publisher" (field 260b). It's only when the library software has to display this information for human consumption that it concatenates the two fields using " : " as a separator. Since we originally had only one free text "Publisher" field, we were forced to stuff both data elements into it, which has increased the number of Publisher records significantly and fragmented publisher-specific information. Now that Publishers get their own records with multiple fields, I think it makes sense to move our "place of publication" data to the Publisher level, although we have to be careful to indicate when the publisher was based in City A and when it was based in City B.

One major issue to consider is DES' Megabooks example. A number of late 19-early 20 century publishers had offices in London as well as New York and it's important to record whether a particular edition was published by "Megabooks (London)" or by "Megabooks (New York). I am not sure if this distinction has been made clearly in the last few decades, but the recent tsunami of mega-merges and acquisitions makes me wonder if there have been books published in "New York London Berlin" and then, separately, in "London Tokyo Warsaw" or something equally bizarre. Perhaps we could ask somebody currently in the publishing business, e.g. Andrew Wheeler, formerly with the Science Fiction Book Cub?

Re: "and" vs. "&", I agree that it makes sense to standardize one way or the other since it's one of the things that causes dispersal of publisher information in our database and in the OCLC catalog. The only exception would be publishers like "A&B Printers" or some such. Another major offender is ", Inc.", which roughly half the cataloging librarians enter and the other half omit. Do we want to create a list of these "problem" conventions -- ", LLC" being another example -- and whack them all at the same time?

Finally, the question of verified publications is a tricky one. I am sure that in 95% of all cases the verifier won't mind if we adjust Publisher-specific data, but we can't vouch for the other 5% since the universe is "not just stranger than we imagine, it is also stranger than we can imagine". How about this as a possible solution: write a script which finds all Publisher records for each proposed change based on some rules -- e.g. change "&" to "and" in "Bill Longley & Friends" or move "London" in "London: Panther" to the Publisher level -- then finds all Verifiers for all affected Publishers. Once the list has been compiled, we can post a canned message on each Verifier's Talk page about the proposed adjustment to "their" records? Ahasuerus 05:20, 11 October 2008 (UTC)

To add to that - I'd like for any mass edits/changes to also document in the notes what was changed and why. For example, If we agree that "A and B Press" should be "A & B Press" then set up a wiki project page for this at Project:A & B Press and in the publication notes we append
  • Oct 22, 2008 Publisher field updated from "A and B Press" to "A & B Press" per Project:A & B Press.
The Project:A & B Press page should have a list of the publication records that were modified as a result of that project and if it's more than a straight substitution of one string for another, collapsing "Incorporated" "Inc", etc. all into "comma space Inc." for example, then the project page should also note the original text for each publication. (Marc Kupper (talk) 19:58, 22 October 2008 (UTC)
To my way of thinking, requiring that level of detail makes the mass edit function useless: it would actually be less work to make the edits pub-by-pub than to capture a list of all affected pubs. I could see recording that on a particular date "A and B" was changed to "A & B" or whatever, but if the proiject must list individuial pubs, I would be inclined to make any changes that seem appropriate one pub at a time, which currently records no history at all. -DES Talk 20:10, 22 October 2008 (UTC)
I hope Marc meant "publisher notes" rather than "publication notes". And that he doesn't really want every publication that was updated to be noted. (We can't even easily find such as the searches get broken down by year, however few pubs there are.) Maybe just the publications that we're over-riding a verifier on, if the verifier hasn't responded in the months we've given them to respond? And I don't think we need a "Project" page for such when there'll already be a "Publisher" page far more easily visible. For instance, I can see the day we decide on "NEL" or "New English Library". Whichever loses can have details of why we changed it. (E.g. "'NEL' is too short a name to search for in Wiki pages".) After a load of Star Trek books arriving today, maybe "Titan" versus "Titan Books" would be a good example of a mass edit approved of after sufficient consultation.BLongley 22:13, 22 October 2008 (UTC)
There's some publishers that are almost extinct after pub-by-pub regularisations already, maybe I can find a simple example of a publisher that can be regularised out of existence after a few pertinent questions. Unfortunately I have to resume work tomorrow (I don't know if people have noticed that I've been restricting myself to no-brainer edits for the past few days I've been ill) and so such an example might have to wait till the weekend. But I don't think we can remain stymied on Publisher edits, even though I DO really want to hold back on stomping on Verifiers data. I've done loads of uncontroversial edits with the new tools, I know some have used them for controversial edits, there IS a balance to be found here. BLongley 22:13, 22 October 2008 (UTC)
This thread, and other adjacent threads, were an attempt to get agreement on what sorts of edits are noncontroversial enough to do without pub-by-pub checks with verifiers. Some people seem to feel that the answer is "none". To me, that means we remain stymied. I'm not advocating that anything anyone wants to do is just fine. I'm trying to get agreement on the sorts of edits, or procedures for making edits, that we can see as generally acceptable. But I think that any procedure which includes individual pub-by-pub checking with verifiers is a non-starter. -DES Talk 22:25, 22 October 2008 (UTC)

Publisher edits: "and" vs "&"

Another proposal (following up on a comment of Bill's in a related discussion): In future it will be acceptable to merge publisher names of the form "Name1 & Name2" with otherwise identical records of the form "Name1 and Name2".

Option 1: We will pick (after discussion) a preferred form to be used in all cases, and any such publisher names may be promptly changed to the preferred form.

Option 2: Any mod may do this, based on such research on the actual form used by the firm in question, as satisfies the mod in question.

Option 3: Any such changes must first be discussed, to decide which form is prefers in the particular case.

In any case, such changes may affect verified pubs without pub-by-pub discussion.

The help will be updated to reflect the new standard. -DES Talk 00:12, 8 October 2008 (UTC)

  1. The "picking" should be done based on what the publisher uses. Ideally the same set of double-verified source publications should cover the entire span that the publisher was active to detect if perhaps they changed their usage.
  2. I'm not sure why you have this. Any editor can work on any publisher.
  3. The discussion should be minimal if #1 is done well.
  4. Add one more is that the publisher's wiki page, under the canonical name identified by #1, should have a section documenting the efforts of #1, a talk page with #3. As I mentioned earlier, this will work better once we have a code update that'll notify someone that they are using a name that we feel should be converted to another name. If an editor says their publication has the "wrong" name then that'll need to be documented.
  5. Wiki pages for the non-canonical names should be added to note that they are the non-canonical version and to ask if someone has evidence that that non-canonical usage occures in in the publication(s) to bring it to the attention of community portal or something similar. Marc Kupper (talk) 02:17, 9 October 2008 (UTC)
What I meant by option 1 was that we would agree on one standard for the whole ISFDB, either always use "&" or always use "and" to be applied to every publisher regardless of that publisher's practice. The selected style would be applied in bulk to all publisher records, changing any relevant verified pubs in the process. What I meant by option 2 was that any mod could select, and impose without discussion, a preferred form for any given publisher, and make the change in bulk, possibly changing verified pubs in the process, without discussion. What I meant by option 3 was that anyone could propose a form for a given publisher, but consensus or at least non-objection must be obtained, probably on a spacial project page, before bulk changes should be made. Once such consensus is reached, bulk changes may be made, possibly affecting verified pubs without pub-by-pub discussion. These are "options" in the sense that they contradict each other, they are three different methods of handling such situations, and i am proposing that we agree on one of the three. I did not intend them as three steps in a single process. -DES Talk 18:45, 9 October 2008 (UTC)
Your items number 4 and 5 are IMO good ideas after any of the three options, and please consider them added to the proposal. -DES Talk 18:45, 9 October 2008 (UTC)
  1. I don't want to standardize on either "&" or "and" for the whole of ISFDB and instead would use whichever form the primary source of the name uses. If they use both then we would use both. At some point we may add support for canonical publisher names and that would be an appropriate place to point the "&" and "and" forms of a name to the same name. Right now we can't merge names without also loosing what someone deemed to be relevant information at the time they added or verified a publication.
  2. Who appointed the mods as gods? The moderators exist to provide for some consistency but not absolute consistency. I'd hold that any editor (moderators are editors too) is invited to research and document publishers. The person makes their case on the publisher's wiki page, provides notice to people that have verified relevant publications plus anyone else that's interested, and will not disregard anyone's input. It's my belied that the effort should be based on a foundation of what's stated in publications with much less weight given to secondary sources, recollection, belief, opinions, faith, etc.
  3. I believe there needs to be be pub-by-pub discussion if someone brings it up as part of a merge proposal but have been thinking that if an editor is not active, or chooses not to participate, that their publications are "fair game" for getting modified by a merge project. It probably means a period, say two months, from the gelling of consensus to executing a merge. Why that long? Simply because it may take a while for some people to track down what they believe are relevant publications. Even then, I would not put an absolute date on it. If someone needs more time the window is left open.
In any case, I'm thinking the merge logic should be modified to at least log the old/new names for each publication so that someone can look back to see what used to be there and if needed, that we can undo a merge, either entirely or for individual publications. This logging can be done in the publication notes field but ideally we add a new table (Al was working on this before life got in his way).
As I've said before, I've regretted every single data merge I've done on my own databases and here we are talking about disregarding what individual editors have chosen to use as the identified publisher of a publication. Marc Kupper (talk) 23:04, 9 October 2008 (UTC)

Publisher edits: verified pubs

It is my belief that publisher regularization cannot usefully proceed until and unless we agree on what sorts of publisher edits can, and what sorts cannot, be acceptably done en masse without consulting verifiers on a pub-by-pub basis. It is also my belief that if the answer is "none", such regularization will be hampered if not stopped dead, and that as discussions proceed, more things will be brought into the category of "regularizations approved for change en masse". Some of these may require general discussion before a particular edit, some I think will not. I have made two suggestions for such categories of edits above (removal of city prefixes, and merging "&" with "and") and I may make others. I do not intend these to be "stealth' proposals, which i why I am (partly in response to Bill's comments) making this post to emphasize this common feature, which may well be shared by other such future proposals. -DES Talk 00:59, 8 October 2008 (UTC)

I believe we want to identify and merge out names a publisher is *not* using but unfortunately, absence of evidence is not evidence of absence meaning that a strong case would need to be made for a name in the ISFDB database that appears to be “fiction.”
I’m undecided on if suffixes such as “Inc” should be normalized out of the database. While in the verified publisher names project we are not including the suffix on the wiki article title we are documenting exactly which suffixes are in use in the publishing name articles so in that sense the suffix is not lost. I’m personally leaning towards that the canonical database name for a company such as DAW would be “DAW Books, Inc.” as that’s exactly what they have used on every single publication I’ve ever seen from 1972 to present. Marc Kupper (talk) 03:55, 9 October 2008 (UTC)
My argument is that there should be exactly one canonical name for any given publisher, even if they have used multiple forms of the name on different books. The only time I would favor permitting multiple names is when there has been a change of ownership or structure, such that the publisher afterwards was arguably not the same entity as the one before, particularly if the editorial team and policies were significantly changed. What form of the name the publisher used on publications is to me, while not irrelevant, far from definitive. If a publisher at the same time used on some books one form an on other books a different form, I want those merged unless there is positive evidence that the difference had significant meaning worth preserving. -DES Talk 18:51, 9 October 2008 (UTC)
If that's the case then nearly every time we have seen a different name then it's a reflection and, perhaps evidence, of some sort of ownership/policy change, etc. I'd submit that few of the names in the database are purely the invention or construction of or by ISFDB editors. Even when we have constructed names, such as Imprint / Publisher it's been an effort to document what's stated or what we know if the record is from a secondary source. We also revisit publications and at times decide to change how we documented its name. For example, we may have learned that a particular bit of additional data is important, or that a detail that we thought was relevant is not so we or remove things from the publisher string. Marc Kupper (talk) 22:05, 9 October 2008 (UTC)
I'm not totally against regularisation of publishers on pubs verified as having a slightly different name, but I think we should make every effort to contact such verifiers LONG in advance of such changes to warn them we're proposing such, and give them a chance to defend their version of the name, or move such information to notes if they think it's needed. I don't think it's quite on a pub-by-pub basis - you can usually spot the one or two hold-outs very quickly and could ask such about their view of the publisher in general. BLongley 22:10, 9 October 2008 (UTC)
I also think we can start with VERY small differences, such as "spaces around a slash or not" as ONE topic, or "If a publisher is CURRENTLY entered as Publisher / Imprint, we want to change that to Imprint / Publisher" as ANOTHER. We are very good at going off topic and leaving threads in such a mess that nobody can actually spot whether there's a consensus on anything mentioned. For instance, I think Marc has already gone off-topic. This is about when/whether it's OK to adjust verified pubs, not about WHY we want to adjust them, or what we want to adjust them TO. BLongley 22:10, 9 October 2008 (UTC)
I disagree that publisher regularisation has to stop if we can't agree on when it's OK to adjust verified pubs - Kraang and I have been doing some regularisation for months without having to address this issue. And for some publishers we don't need a general discussion, there's only one verifier you would have to talk to. There's still much that can be done before we even have to talk to somebody, but if we're talking then let's try and keep it to separate, simple cases, and engage the people that would be affected. It's slow, yes: but I think it has to be. Not so slow that we CAN'T use the powerful tools when we have agreement though. And the more agreements we get, the sooner we move forward. BLongley 22:10, 9 October 2008 (UTC)

Fictional authors

Apparently "Mars", a work which seems to lie in that vague area sometimes called "fictional essays" or "in-universe POV essays" has been recorded (in a VT) as being written by "Survey Commission Office, Kahora, Mars" which is listed as a pesud of Margaret Howes. I must say, I dislike in general the idea of crediting fictional characters or entities (such as "Survey Commission Office, Kahora, Mars") authors in the DB. Yes, authors frequently assign "credit" for stories, parts of stories, or in-universe essays to fictional "authors" as a form of realism (or sometimes as a joke). But that is all part of the story. We wouldn't list the Sherlock Holmes stories with a co-author credit for "John D. Watson", nor do we list the Callahan's Bar stories with a credit for "Jake Stonebender". I realize this is an odd case, but I would prefer that this be discussed a bit more before this precedent is followed. Previous discussion on this issue can bee seen at User talk:BLongley#The Best of Leigh Bracket. What do others think about this kind of thing? -DES Talk 16:54, 8 October 2008 (UTC)

I'd tend to agree with you - if a name is clearly fictional and is part of the story we'd add a note documenting what was said, probably both in the publication and title records for the story. For this case I've started the note at THBSCKTTAD1977 but will leave the existing fictional author in so people can see it. Marc Kupper (talk) 18:32, 8 October 2008 (UTC)
Well, we have Kilgore Trout here. But I don't think I've ever seen Jake Stonebender or John D. Watson credited on a title page or even a Table of Contents entry, so there is SOME separation between "exactly as recorded" and "what the fiction says the author is". In this case, I'd have interpreted the "authors" or "artists" differently if pushed, but rather than pre-colour people's conceptions I'll let people go look and decide for themselves. It's not a great example of a problem though (IMO), but it's a start. BLongley 21:50, 8 October 2008 (UTC)
There's also Walter Bupp, who was a pseudonym for John Berryman or maybe Randall Garrett, the credited author of some stories in Astounding/Analog, but is also the narrator/protagonist of those stories. I have to say that this is a case for using the fictitious name as author, IMO. I don't feel the same about a section of a longer work, credited to a character within that work. -- Dave (davecat) 15:37, 9 October 2008 (UTC)
As I started this 'problem', my visualization of what happened may help. I saw 'Map' listed as a title of an essay. That seemed odd. It was not listed on the 'Contents' page. It would have been in the 'Afterword' by Leigh Brackett. Note LB is credited specifically in the contents for it. The 'Afterword' is split into three sections and the maps follow those. Afterword starts on pg 414 but at 418 their is a Leigh Brackett, place, date. That can give a person the idea it ended there, but now I think it should be seen as ending only that portion/section. Under that portion ending is Addendum. Addendum is at the far left and text is underneath. First line is 'About the Maps of Mars.' Then below it starts the paragraph explaining and crediting Margaret Howes for the maps. The third paragraph is this. "Far from minding , I was delighted. And in due course, the maps arrived, a pure labor of love, together with keys and 'several closely-reasoned pages' concerning the reasons for placement, etc,. The several closely-reasoned pages bothered me. The fourth paragraph tells how good the maps were and my editors decided "to include them in the collection." That credits the map specifically, but nothing else. LB lauds the map work. Fifth paragraph. "The cities and canals, of course, are fantasy. But the maps themselves ...not so. Listen to the Cartographer." Sixth paragraph onward quotes M Howes. The final paragraph is back to Brackett and thanking Howes and fans. Signed. L.B. and dated. Date different than Afterword first section. Then comes the section Marc showed you.
The problem is should I have parsed the 'Afterword' into sections. As a single essay of three parts I believe it did not need to be separated. In my mind it is like reading a story, with a story being told in it. I would not do that and probably should not have separated this. That the third essay is a story is probable, but why can not an essay contain a story. That the 'maps' need to be credited I have no doubt as we separate such according to the 'taste' of the editor. In this case it is doubly important, because Marc did a super check and found a literary reference to this work in another source. He established beyond doubt that Margaret Howes deserves credit. The rest I leave to you all. I think everyone has done a good job at a very complex task.
The importance of the topic is the proper crediting of material in books by others who the author brings up usually in non-specific ways. Sorry for the confusions. Thanks, Harry. --Dragoondelight 13:49, 9 October 2008 (UTC)
I absolutely agree that proper credit should always be given when we known who to credit. It seems absolutely clear from your description above that Howes should be credited for the map. (And I am generally for listing maps.) I am not quite as clear on her credit for the section of the Afterword, but if you are satisfied I have no argument. My only question was whether "Survey Commission Office, Kahora, Mars" should be listed as a pesud for Howes (or for that matter, for LB, if we had concluded that she wrote the section in question), which does not IMO really depend on whether to list the sections separately, or whether to credit Howes or Brackett. Indeed I'm not really discussing this particular case at all as much as the general policy issue for future cases. By the way, note that the subtitles of some of the Sherlock Holmes works were statements like 'An excerpt from the memoirs of John Watson, M.D.", so he was credited on the title page. Now those are not SF, but some SF from the same period used similar devices, I think possibly including works by H. Rider Haggard. I would dislike to see an author record for Allan Quatermain (hero and narrator of several of Haggard's books). -DES Talk 18:07, 9 October 2008 (UTC)
I found a Doyle example that does belong here: The Lost World. I checked, and it seems nobody has yet created "Professor E. Challenger", "Lord John Roxson", "Professor Summerlee" or "Ed Malone", so I don't think people are crediting authors based on subtitles just yet. I agree they SHOULDN'T in future either. BLongley 19:06, 9 October 2008 (UTC)
Back on topic: now you've all had a look, I'll now say I wouldn't have made "Survey Commission Office, Kahora, Mars" an author at all - it looked more like part of the title or sub-title to me. I MIGHT have made "Margaret Howes" or "Margaret M. Howes" an author for the bit before the map - but as I had no evidence as to whether the middle initial was real or part of the fictionalisation I bailed and have left the decision to others. Has anyone confirmed the reality of the "M."? BLongley 19:06, 9 October 2008 (UTC)
Oh, I've just thought of an example that I think better matches your concerns DES: Dray Prescot. Would that be a more relevant example? BLongley 19:13, 9 October 2008 (UTC)
That is a very relevant case, although it wasn't quite the sort of thing I was thinking of. I was more concerned with such credits on maps, essays, and other such included works. (For example, IIRC, The Door into Shadow -- which i must verify -- has a map with a printed credit to a fictional character, a dragon no less, apparently to indicate that it is an "aerial survey") Both Dray Prescot and Kilgore Trout were, as i understand it, given as the sole author credit on book covers and title pages of at least some editions of the relevant books, which makes them both pseudonyms and names of fictional characters, i guess. The whole issue is tricky, and I'm not sure exactly where to draw the line. -DES Talk 19:52, 9 October 2008 (UTC)
If they get cover credit as the author they should get listed in the ISFDB since people might be looking under that name. Otherwise I'd say not listed if it is at all obvious that it is a fictional name. I'd think story in a magazine where they get TOC credit falls under the listed side. Dana Carson 20:12, 9 October 2008 (UTC)
Going the other way then, Fantastic Beasts and Where to Find Them and Quidditch Through the Ages need "Newt Scamander" and "Kennilworthy Whisp" author credits. (Actually, under current rules they should anyway on TITLE page basis, not just Cover credit.) And "Albus Dumbledore" gets credit for an introduction. How "obvious" are these as fictional names? (I'll leave aside the facts that these are CHAPTERBOOKS, have fictional publishers, and fictional prices. I don't INTEND to buy my canned-worms by the case, it just happens that way.) BLongley 21:19, 9 October 2008 (UTC)
I'm not sure how to word the "rule" that cover all of Bill's worms while also constructing a "do not cross" line in the duroplast. For example, while Dray Prescot is clearly fictional I support that he is listed as the author as his name is on the spine, and the title page will say something like "Title by Dray Prescot as told to Alan Burt Akers" with the copyright being by Dray Prescot. I see that Alan Burt Akers, who is also fictional, has credits for the Prescot books and essays. Thus when it comes to wording the "rule" it's going to be something along the lines of the following for Template:TitleFields:Author which I believe is only used for the Author field of publication contents. (it's not the same as Template:PublicationFields:Author which is the author field of a publication itself nor Help:Screen:AuthorData which is for the author record itself.)
Fictional names - A common story telling device is to include short stories, essays, songs, poems, and illustrations, etc. credited to or "by" a fictional name. Often times these will be presented in an "in-universe POV" for the publication/story and will be be written, performed, or drawn "by" one of the story characters.

Rather than crediting the fictional author or artist in ISFDB and then doing variant titles and pseudonyms back to the real author or artist you should credit the author or artist of the main story body directly and would document the fictional author/artist in the publication, and ideally title, notes.

For example, a story could include an illustration of an Aerial Survey "by" a dragon. Rather than adding an INTERIORART with "Dragon" as the artist you would look at the copyright page, etc. to see who is credited as the artist and credit him or her. If the artist is not credited enter the artist as "uncredited" and in the publication/title notes you would document that the work was credited to "Dragon" in the publication. (should we have a special sub-case to credit as "Dragon" rather than "uncredited"?)

Note that fictional names are not the same as pseudonyms. If an entire story (either novel or shortfiction) is "by" someone you would credit the name as stated and would set up the variant titles and pseudonyms as needed.

Also note that if an essay, song, or poem is included as part of a longer work that you do not need to add title records for each of these. The intent of this section on fictional names is to note how to handle fictional authors/artists standalone works that are also part of a larger work where you would normally be adding a new title record had the work been credited to a non-fictional name.

At present that's a bit too long - if someone has ideas on how to tighten it up then fire away. Marc Kupper (talk) 21:47, 9 October 2008 (UTC)

(unindent)I agree with Dana's underlying reasoning. As long as a "naive" reader -- I typically use a 12 year old boy to help with visualization -- who has never heard of Kilgore Trout, Dray Prescot or Kennilworthy Whisp may think that the credited "author" is a real person, I think it's useful to set up a pseudonym record. If, on the other hand, the attribution in the book strongly suggests that Dr. Watson and Lord John Roxson are characters and not real people -- or the attribution is clearly fictional, e.g. "dragon" -- then I don't think it would be useful to capture what is basically an in-universe convention. Granted, some naive readers are more naive than others, and a 5 year old reader may think that dragons are real, so the line is a little blurry, but it will likely get the job done in 90%+ of all cases. Ahasuerus 02:12, 10 October 2008 (UTC)

I can agree with Ahasuerus on this, i think. The suggested help above is not a bad start, but probably needs polishing. There will always be odd cases not well covered by any rules we devise, unless the rules are so long and complex that they are useless for all ordinary purposes anyway. -DES Talk 05:02, 10 October 2008 (UTC)
I'm still hoping for some feedback on either how to make the fictional names section smaller or perhaps that we need not bother with adding a section about them as I suspect the issue does not come up often. I'd personally rather that the help pages be much smaller and perhaps to link to articles that give the full details. Marc Kupper (talk) 21:38, 10 October 2008 (UTC)
I'd say it's not worth bothering about. "Survey Commission Office, Kahora, Mars" was a mistake and should be rectified. But Kilgore Trout and Dray Prescot are fine by me, and applying the "12 year old boy" test to "Kennilworthy Whisp" makes me think that only someone that's never even looked at a physical copy (such as the person that created the Amazon data for the "Omnibus" (just a pairing of the two books, it seems)) would be confused. BLongley 21:34, 12 October 2008 (UTC)
I would like help pages to be shorter and link elsewhere to the reasoning behind them though, so that's still a good goal to aim for. BLongley 21:34, 12 October 2008 (UTC)

Fanzine Reviews post-lexical match

Now that Reviews have been changed to link to Titles directly, do we want to revisit the issue of Fanzine reviews? I assume that at the very least we will want to delete the sentence "fanzines are not indexed in the ISFDB" from Help:Screen:NewPub. Ahasuerus 01:55, 10 October 2008 (UTC)

Any chance of reconsidering the decision to index fanzines? -DES Talk 05:09, 10 October 2008 (UTC)
Well, the decision was made a while back and we now have hundreds of fanzine issues entered and linked from the Fanzines page, so it seems highly unlikely that we will ever go back.
As an aside, I was neutral when the decision was made since I wasn't sure that we wouldn't be duplicating what other, more qualified, folks had been doing. However, after reviewing some of the other Web pages that cover the same territory, it's my impression that we can do a reasonably good job of it, especially given that our software and our standards are generally superior. We also provide an easily downloadable backup file, which makes our information more durable than a random collection of Web pages. Ahasuerus 06:41, 10 October 2008 (UTC)
There is far too much of value in fanzines to even think about not indexing them. Besides that Al and Bill are probably as happy as I am to be in the database. I don't think there is anyplace else that has the ability to accommodate the amount of searchable cross-indexed integrated data, book reviews and interviews being of particular value, that the isfdb is potentially capable of. I will mostly be done with my magazine data entry by the end of this year and I intend to enter my small selection of fanzines. As far as fanzine reviews go - the prozine reviews of fanzines usually don't list a specific issue so it would mostly be fanzine review in other fanzines that would be entered unless the prozine reviews linked to a single generic fanzine pub.--swfritter 17:10, 10 October 2008 (UTC)
I think they are a valuable addition, and not just because it's the only way I'll appear here. One of the reasons I HAVEN'T been doing much IS because there are good sites out there - e.g. the BSFA index was inspired by the ISFDB anyway, is pretty good already, and might well be able to provide data as a mass import rather than having us re-typing it. Actually, I should probably ask for the data anyway: Michael Cross doesn't seem to be updating the site any more, and still links to some of our older web-site addresses. BLongley 19:28, 10 October 2008 (UTC)
Fine, I withdraw the suggestion. You may be right, and in any case if i'm not doing the work of entering fanziens I suppose I have no reason to object. Please feel free to drop the side-topic and return to your scheduled discussion of how to move ahead. -DES Talk 19:51, 10 October 2008 (UTC)
The issue holding up the entry of fanzines right now is Letters. Issue A is whether or not they will have a specific entry type (or my preference a sub-type). Issue B is the criteria for entering them - all or just by well-known names (which means I would be out). My own philosophy is developing into "while I'm there I'd better do it because I may be the last person who ever touches this pub". I am more and more tending towards inclusiveness although I don't expect anyone to enter data any more comprehensively than they have the available time, interest, energy, or motivation.--swfritter 16:43, 11 October 2008 (UTC)
I don't think Letters are an issue really. Or if they are, they shouldn't be. I agree help on such is a mess, but I have no problem with Letters or Reviews by notable people (or not so notable, like me) being entered individually rather than as part of a bigger section. I suspect the bigger problem is that allowing fanzines in is that reviews of such fanzines might become perceived as the norm for inclusion in other publications, not just fanzine reviews of fanzines. I quit magazine editing as I felt pressure to generally add more data than I care about. So long as there's no pressure though, I have no problem with anyone adding more data to ANY publication. I'm redoing all my early verifications to add cover-art and other prices and printing number details, for instance. That's something I choose to do, and if people want to follow that, fine. If they don't - I'm not going to badger them to do so. What we have is pretty useful already: setting an example of what could be done is fine too. I lean towards inclusion as well. BLongley 23:21, 12 October 2008 (UTC)
OK, "fanzines are not indexed in the ISFDB" has been deleted. Now we need to figure out what to do about fanzine reviews. It occurs to me that since Review records are linked to Title records and since EDITOR records for fanzines and magazines are usually merged for the same year/editor, it means that even if we decide to link to individual issues, we may not be able to. Hm... Ahasuerus 16:42, 15 October 2008 (UTC)
I guess one option is to enter them as essays and series them. I think that points up the need for a generic non-linking, non-extra-data-generating REVIEW type or sub-type. As far as level of detail Bill - perhaps use the transient verification so we have some idea who entered the data or the last person to touch it. When a pub is Verified editors have a tendency to agonize over making changes to somebody else's work.--swfritter 17:05, 15 October 2008 (UTC)

Literary nonsense - in or out?

Charles Lutwidge Dodgson, aka Lewis Carroll, writes in a genre known as Literary nonsense. Is this part of speculative fiction, perhaps as a subset of fantasy literature? Marc Kupper (talk) 22:34, 12 October 2008 (UTC)

Have you got an example of anything we have here that might not be appropriate, or should have here and don't? My first thought at seeing the title was "Edward Lear" and I'm thinking "Hilaire Belloc" as I post... but I'm in an "inclusionist" mood at the moment. BLongley 23:31, 12 October 2008 (UTC)
I don't know that we should be concerned with "Literary nonsense" as a category or genre, and that is not how I would classify Carroll's better known works. I am convinced that Alice in Wonderland and Through the Looking-Glass are fantasies, and classic ones. I would at least argue for The Hunting of the Snark. (I am less sure about Sylvie and Bruno or others of his works.) But that doesn't mean that everything labeled "Literary nonsense" is an example of fantasy. After all, "Alice" and many other classic fantasies are also labeled "Children's literature", but that surely doesn't mean that all books so labeled are IN. -DES Talk 00:04, 13 October 2008 (UTC)
I suppose you could argue that SF writers like Daniel M. Pinkwater write "literary nonsense" much of the time, but it also makes a twisted kind of sense, especially to pre-teens :) But the underlying problem seems to be that our Definitions section, which covers fabulations, magic realism, and slipstream, doesn't mention surrealism or related areas.
My first thought is that as long as the work can be classified as "fiction" and is more or less speculative rather than just surrealistic, we will want to include it. Thus Yuz Aleshkovsky's Kangaroo, "[a] nightmarish journey through a surreal vision of Soviet history beginning with the trial of the rape & murder of a kangaroo", wouldn't be included since it's surrealistic rather than speculative, but I am open to suggestions. Ahasuerus 01:03, 13 October 2008 (UTC)
The question was triggered by The Hunting of the Snark: and other Nonsense Verse which happened to be next on a shelf I was verifying. Much of the material fantasy, some is fiction, and Jabberwocky had me puzzled as far as classification went until I saw that Wikipedia uses nonsense verse. I like Ahasuerus' use of surrealistic as that's a good description of much of the material in this book. Marc Kupper (talk) 01:10, 13 October 2008 (UTC)
From Definitions section which Ahasuerus mentions. Links are to Wikipedia.
Oh, and we should probably clarify that "non-genre SF" means SF published outside of the traditional genre venues like magazine and specialty publishers and, more likely than not, unaware of the established genre conventions. I think I recall at least one of our editors being confused by the term. Ahasuerus 03:24, 13 October 2008 (UTC)
I'm still confused. What are the established genre conventions that someone should be aware of? Marc Kupper (talk) 05:56, 13 October 2008 (UTC)
Well, the idea here is that we want to list books that are speculative in nature even when they are published as "literature" rather than "science fiction/fantasy/supernatural horror". Thus, we want to list Updike's Toward the End of Time and the speculative stuff written by the likes of Gabriel García Márquez, Margaret Atwood, Salman Rushdie, Doris Lessing and other folks who wouldn't know a tachyon from a dhampire :-) Ahasuerus 14:33, 13 October 2008 (UTC)
It seems like a slippery slope into people adding the Torah, Bible, Quran, Scientology, etc. :-) However, I was not planning on drifting that far. (I see that the rules are silent on historical fiction for example and presumably all romance books are "fantasy"). Literature is a rather broad brush - I understand and agree with what you wrote but am thinking perhaps one more rule on the definitions section that says "If a particular author or work does not seem to be included or excluded by these definitions then please ask at Rules and standards discussions. The intent is that we want to include works that are speculative in nature even when they are published as "literature" rather than as a work that most would recognize as science fiction, fantasy, supernatural horror, etc. A key element for speculative fiction is that the story setting is unlike the past or present real universe in various important ways." Marc Kupper (talk) 05:23, 15 October 2008 (UTC)

ISBNs and the SFBC

Question in play here is whether to use the ISBN# or the SFBC catalogue # in the pub record. For about the last 10+ years the SFBC editions have included the trade editions' ISBN on the copyright pages and the back covers. The designation "Book Club Edition" has been removed from the front flap. All that's left to distinguish the different edition is the catalogue ID# (besides the physical size difference on most, but not all) and the lack of certain data on the copyright page (a number line, usually, but not always). These differences are easy to see with book-in-hand. But one of this site's primary uses is for searches, and an ISBN will never bring up an SFBC specifically (other than those newer omnibuses unique to the SFBC). Inputing an SFBC catalogue # won't bring up an SFBC title either, though it could at some point, but not if those ID#s exist only in the notes. If the SFBC had never done anything but reprints, it could almost be a footnote, but with the unique omnibuses and the few first hardcovers it produced, in some cases the only hardcover edition (PKD's "The Preserving Machine" comes to mind) or the true first edition (Herbert's "Hellstrom's Hive") that option doesn't come into play. The SFBC issues are many, this being just one. For any to be resolved, or at least handled in a consistent manner, data needs to be as accessible as possible.--Bluesman 16:04, 15 October 2008 (UTC)

If an SFBC edition had its own ISBN, then it should clearly be recorded in the iSBN field. Do the "unique omnibuses"now have their own ISBNs? But when the ISBN is merely copied from the trade edition, I'm less sure. Granted that in other (non-SFBC) cases the same ISBN is used for multiple printings, so that an ISBN does not uniquely identify a particular printing, and not always a particular edition. But ther is a case to be made for leaving such "borowed" ISBNs in the notes field, and recording the catalog number, on SFBC editions. And the case is IMO stronger when the borrowed ISBN is for a trade-paper edition, as I gather is the case for at least one recent publication. -DES Talk 16:32, 15 October 2008 (UTC)
The unique omnibuses that preceded the takeover by Guild/America did not have an ISBN. All the ones since seem to, though the earlier ones only put it on the copyright page and not until the later 90s put it on the back cover (this is also true of a series of novellas they put in mini-hardcovers similar to the old PermaBook ones of the early 50s that were done in 1997). Since these are unique to the edition, they should lead to that pub in a search, and then the SFBC catalogue ID# rightly belongs in the notes. When the only identifier between a trade edition and an SFBC edition is the catalogue ID#, shouldn't it be more prominent than a footnote? --Bluesman 23:46, 15 October 2008 (UTC)
I am for leaving ISBN's in the ISBN/Catalog # field when it is known, and putting the SFBC# into the Notes field. In a perfect db, we would have fields for both, but I wouldn't prioritize that request very high. CoachPaul 16:42, 15 October 2008 (UTC)
The problem is, in some cases these ISBNs appear to be not really the ISBNs for the SFBC edition. See User talk:Bluesman#Dust for the recent example that I think spawned this thread.

If there is an ISBN in the book, and there is no dust jacket, how is someone to tell if it is a SFBC book or not? An ISBN in a book belongs in the ISBN field! CoachPaul 00:30, 16 October 2008 (UTC)

the data on the copyright page is different in a trade edition, as it includes the "Library of Congress Cataloging-in-Publication Data", which is always missing from the SFBC reprints.--Bluesman 14:58, 16 October 2008 (UTC)
100% of the ISBNs on the reprints are not for the SFBC edition, the primary reason to keep them in the notes.--Bluesman 23:50, 15 October 2008 (UTC)
By the way, the plural of "ISBN" is "ISBNs", not "ISBN's", in my opinion. Similarly a decade is referenced as "1990s", not as "1990's". -DES Talk 17:22, 15 October 2008 (UTC)
That's a sticky one - ideally the "ISBN" field allowed for multiple values and we input both the SFBC number and the stated ISBN. Someone doing a search with either would find the record. We can put both into the field using "#12345 / 1-123-12345-X" for example but the publication search form seems to use exact-match meaning a search for 12345 or the ISBN alone fails. Even if we fixed the exact match we'd still be dealing with someone searching for 112312345X and that would fail to find the SFBC publication. At present I can't think of a rules-only solution that introduces no or minimal astonishment for everyone. If a publication had both an SFBC and ISBN I would tend to put the SFBC # in the Catalog #/ISBN field and to document the ISBN on the notes using both the 112312345X and 1-123-12345-X forms so that someone using Google can spot the record. Marc Kupper (talk) 02:25, 16 October 2008 (UTC)
I just checked four books awaiting the 'SFBC' issues. None had ISBNs, three had numbers, all of which you had to assume had significance, all four had 'book club edition'. I do think SFBC as a producer/provider of reading material needs to be highlighted for the db users to find. In all four cases in the ISFDB, the book publisher is put before SFBC. I would prefer to see SFBC / publisher because the publisher actually is supposed to be producing it for the club not for direct consumption. I also have the tendency to think SFBC book number trumps ISBN in this case. SFBC books often do not meet the highest publishing standards and when they can be confused with a 'first-class' product through the ISBN search the db user suffers. In this case. [22] . The book was sold through Amazon (not Amazon vendor), but not as a SFBC. My point for this is the seller (Amazon) may not differentiate between them, and so when using the ISBN a vendor need not differentiate it either. ISFDB owes the db users the best parsing of the book differences as it can provide. Thanks, Harry. --Dragoondelight 15:24, 16 October 2008 (UTC)
I am reasonably certain that the SFBC produced all of their early reprints and omnibuses, as they were all exactly the same size and "quality", and did not have them printed/bound for them by the various publishers. All the omnibuses I have seen, pre Guild/America, are Nelson Doubleday regardless of who had the original copyright. Don't know if that would affect the order in the Publisher field or not. I do agree with the db giving as accurate a finding to the users as possible.--Bluesman 22:08, 16 October 2008 (UTC)

"Not safe for work" images

As some of you probably know, Glen Cook's first published novel, The Swap Academy, is very rare and even Cook doesn't have a copy. It's a porn novel, and not a very well written one at that, but it's something of a rarity, so I thought that I should scan and upload the cover art. However, it occurs to me that the cover is not entirely "safe for work" and we obviously wouldn't want our users to get in trouble when accessing the site from work. I wonder if it might be possible to mark images as "not safe for work" so that they would have to click on a warning message, e.g. "This image may not be safe for work, please confirm that you want to view it", before displaying it. At first I thought that it would require a software change, but since we store images in the Wiki, I wonder if there is already an add-on that handles these issues? Ahasuerus 16:27, 16 October 2008 (UTC)

We might also want to include most of the pulp covers under that designation. It's pretty obvious that all astronauts should be women because their (mostly) bare bodies are impervious to the vacuum of space.--swfritter 17:30, 16 October 2008 (UTC)
I don't want to get into the business of trying to judge what someone else will consider "safe for work". Most employers will probably not favor anyone using the ISFDB while at work much anyway, no matter what kinds of images are displayed. To the best of my knowledge, there is no way to implement such a process from within the wiki, and if there were, i for one wouldn't so mark any images, no matter what their content -- Indeed I would make a point of finding any raunchy cover images I could, and uploading them. Notice that project Gutenberg's overall image for their SF shelf is at least arguably "not safe for work" as have been some of our banner images -- Images of classic pulp "scantily dressed" women, while clearly not porn, probably fit the anti-harassment "hostile work environment" criteria in place at many firms these days. Wikipedia has also declined, on principle, to have any such warning labels on any of its images, even ones with explicit sexual content (such as those illustrating articles on various sexual practices). While not every Wikipedia policy fits the ISFDB, IMO that one should. -DES Talk 20:40, 16 October 2008 (UTC)
Unfortunately -- and I speak as someone who watched Pasolini's Salo, or the 120 Days of Sodom twice back in the day -- there is a great deal of difference between using a non-work-related Web site and a sexually explicit one, at least as far as most American employers are concerned. The former may be frowned upon or, depending on the frequency, eventually get you in trouble, while the latter can easily get you fired the next day. Granted, we only display small images which are rarely more than borderline erotic in nature, so browsing Amazon.com is probably more dangerous than using our site, but I wanted to make sure that we are least aware of the issue. Ahasuerus 22:07, 16 October 2008 (UTC)
True enough, although few if any of our covers would be in the "fire now" range, I think, even for the raunchier pulps. Perhaps we should add to the general disclaimer. But I still object on principle to self-censorship of relevant images. -DES Talk 04:40, 17 October 2008 (UTC)
I personally don't think it's an issue as people don't see the image until they drill down to the publication level. An option would be to reduce it to 300 pixels or fewer before uploading. Marc Kupper (talk) 06:22, 17 October 2008 (UTC)
Note that it is already reduced significantly when displayed on a publication page. A user sees the full 5-600 pixel image only after clicking on the thumbnail. I haven't measured, but I suspect that the thumbnail is no more than 200 pixels -- compare for yourself. It is surely less than half the size of most uploaded images. -DES Talk 13:32, 17 October 2008 (UTC)
Well, here is the cover in all its glory -- http://www.isfdb.org/wiki/index.php/Image:SWPAC1970.jpg -- and again, it's not safe for work. Make of it what you will :) Ahasuerus 03:21, 18 October 2008 (UTC)
I will grant, that is more explicit than almost any SF cover I have seen. I presume it is the cover art for this pub, correct?. That does raise a related question, at lest in my mind. The work is NONGENRE, included for completeness because the author is "over the threshold". Should we, in general, show cover art for nongenre pubs, work-safe or not? I'll restate this in a new section, to avoid topic drift within this one. —The preceding unsigned comment was added by DESiegel60 (talkcontribs) .
How about this compromise then -- for sexually explicit images, do not enter the URL In the URL field, but make it a clickable link in the Notes field with a "may not be safe for work" warning? Ahasuerus 17:24, 22 October 2008 (UTC)
Let me be clear. I will not, as a matter of principle, agree to treat "not safe for work" images in any way differently from any other images, unless they are legally obscene. (obviously, unless they are covers of books with pub records, there is no particular reason to upload or link to any such images.) I will not refrain from scanning and uploading them. I will not fail to link them, nor link them differently from other images. i will not label them in any way. If there is a consensus that it is policy to do so, i will withdraw from the project. To me this is not compromisable. If others choose not to link or upload such images, that is their choice, and no policy currently requires uploading or linking to any cover image. -DES Talk 18:53, 22 October 2008 (UTC)
I don't think there's actually any requirement to agree to any ISFDB Wiki policy here. After all, I've not agreed to any of your Cover Image Data policies. (I upload stuff and point at the ISFDB entry, you come along and add loads of other stuff I find ugly, unnecessary and confusing.) I haven't quit because you have a policy to add such. If you started DELETING such then I'd ask for you to be removed from such a position that allowed you to do so, or I would stop using ISFDB for images. If someone starts adjusting your Image uploads to add "NSFW" metadata would that make you quit "the project"? (I'm not sure if you mean ISFDB or "ISFDB + Wiki" or just Wiki.) Or would you just do as I have and use the Wiki to support the database, realising that the Wiki side is not the important side and everything we add to the Wiki MAY be useful but is not essential? BLongley 22:50, 22 October 2008 (UTC)
I take it that if something has enough consensus to constitute "policy" then editors are expected to comply, whether they agree or not. That is what a policy is, IMO. Whether it is Policy that "all images must include a license tag" might be debated. On Wikipedia that is a firm policy, images that violate it are routinely deleted by automated scripts, and users who routinely upload such can be blocked. That isn't the case here, and I don't see the need for it to be at this time. We have a policy that legal name is recorded "lastname, firstname", for example. Complying with that (once an editor has been notified) is not optional. We have a policy of not including non-genre works by non-genre authors. If such are discovered, they are routinely deleted. Choosing to add such books is not an option for an editor.
If there is a general agreement here, amounting to a policy that editors are expected to comply with, to treat NSFW images 'on pub records differently, either by labeling them, by not actively linking in the same way as other cover images are linked, or by not linking to them at all, I would leave -- do no further work on the ISFDB. If some people started removing, labeling, or otherwise changing such links, I would ask them to stop, and ask for a consensus decision on the matter. If someone were simply adding such labels to the wiki pages, I would ask for discussion also, but less urgently. If there was a policy requiring such labels to be on all wiki pages with such images, I'm not sure what I would do, but I would probably leave the ISFDB.
I might add that with publisher notes on the wiki side, we are beginning to have at least some "essential" data on the wiki. I'm afraid I don't agree that, in all cases, data is less important simply because it happens to be stored on the wiki. Much of the data so stored is less important, but not all, and there is no guarantee that this will always be the line of separation. -DES Talk 23:31, 22 October 2008 (UTC)
Oh by the way, US law is clear. Unless images are legally Obscene, by the Miller test, no one can be prosecuted for displaying or hosting them. Basically this means hardcore porn, with explicit images of actual sex, well beyond the Playboy level, or even the Hustler level.

(unindent)I think Bill was talking about "ISFDB Wiki polic[ies]" while DES was explaining what he would do if we decided to "treat NSFW images on pub records differently". Wiki side policies are generally more nebulous, although we enforce a number of basic ones, e.g. "Bio" and "Biblio" pages shouldn't be used for advertising and self-promotion.

Anyway, I think it may be advisable to take a step back and analyze how much of an issue this is at the moment and whether it is likely to become a problem down the road. After all, the scaled back images displayed by the "Publication display" logic are quite small and I am not entirely sure that they would be truly "not safe for work" even with dealing with explicit images. Perhaps we could do a poll of our user base by posting a question on rec.arts.sf.written? Ahasuerus 17:11, 23 October 2008 (UTC)

I would not like to see marginal s-f entered into the database just because it has such artwork. Much of s-f is pre-teen lit although when you look at the subject matter of some YA material it becomes clear that times have changed.--swfritter 18:39, 23 October 2008 (UTC)

Cover art for non-genre pubs?

We include "non-genre" works to help complete the bibliographies of authors who have written a "significant amount" (currently undefined) of SF, not because it is otherwise within our scope. But such works, once entered, are pubs like any others. Should we encourage, discourage, or prohibit the uploading of or linking to cover art for such works? This question was inspired by the discussion above, in which it became apparent that this pub had this cover (not safe for work, explicitly sexual). But the question is more general, should pubs like this mystery novel have cover images linked if possible, or should those be considered out-of-scope? -DES Talk 13:32, 18 October 2008 (UTC)

I think cover images serve a couple of different purposes when used within a project like this one. First, they are valuable in and of themselves since many people enjoy and/or collect cover art. Second, they can be used for verification and disambiguation purposes, as we have seen over the last few weeks. The downside, of course, is that they take up a lot of disk space and make our "full backup" file much bigger than they would be otherwise. At the moment, cover scans add up to about 50Mb of disk space and we have barely scratched the surface.
Applying these considerations to NONGENRE works, their cover art is generally of less interest to SF readers, but I can see how a Silverberg completist (to use a random example) may be interested in the covers of Silverberg's non-fiction, YA books and soft core porn from the 1960s. As far as verification/disambiguation goes, covers can be quite useful regardless of the genre. Finally, disk space is not a major concern in this case since NONGENRE works are but a small percentage of what we have on file.
Overall, limiting the use of cover art for NONGENRE works doesn't seem to be indicated, except, perhaps, in cases of sexually explicit or otherwise "not safe for work" (graphic images of gore and corpses?) material. Ahasuerus 02:16, 19 October 2008 (UTC)
I've added some coverart for NONFICTION today. For some possibly NONGENRE Doris Lessing as well. And for Movie Tie-Ins that probably go beyond the "copyright is believed to belong to the publisher or the cover artist" as well. ("Paramount Studios", for instance?) Another of the reasons I hate the boiler-plate entries suggested/recommended/demanded for images. Some of the Chris Foss verified signature entries could come from his work on Asimov's Non-Fiction, etc. And Boris Vallejo seems to rarely have created anything that would be acceptable to our fundamentalist Muslim editors... BLongley 23:11, 22 October 2008 (UTC)

Hyperlinks for Navigation Through Magazine Issues

I've been working mostly in magazines and it would be convenient for both editors and users to be able to move backwards and forwards through the issues without having to return to the magazine page. Ideally the "previous" and "next" issue buttons would be written into the application, but I don't think that's going to happen. I'm proposing instead that hyperlinks to adjacent issues be entered into the Notes, as in this example. The links would be placed at the bottom (top?) of the "Notes" window, and marked off by html comments for identification and stripping if necessary to do so for any reason. Comments?--Rkihara 16:49, 17 October 2008 (UTC)

If someone really wants to do the work I don't mind. This is actually an uncommon way for the average user to access the system. The most common way to access the system is through author and title searches.--swfritter 17:33, 17 October 2008 (UTC)
I think ít is fine idea. I browse magazines very often, and it is not at all uncommon way access info for me. I am not too familiar with html - how do you add such a link? Just [http ://whatever.the.link.is next months issue] (without the extra space, of cource)? By the way, top on "notes" rather than in end. The lenght of notes vary, and when the link would on the top, it would always be in approximately same place. Tpi 18:33, 17 October 2008 (UTC)
See the notes in Edit Mode. Notice the comment section header which does not show up on the screen. This would be an indicator that this method would not work anyplace but in the isfdb environment. If this were put at the top of the notes it would require putting such a section before and after the links.--swfritter 18:42, 17 October 2008 (UTC)
Links on top would make browsing faster, since the cursor wouldn't have to moved. I'll change the example. The links for most of a magazine run could be generated through a macro, so only cut and paste would be required. Magazines with irregular publishing dates would require a bit of tweaking.--Rkihara 19:21, 17 October 2008 (UTC)
I believe that lack of magazine navigation tools has been commented upon on rec.arts.sf.written a few times. A programmatic solution would be much better, of course, but if someone has the time and the energy to apply a band-aid for now, more power to him :) Ahasuerus 22:23, 17 October 2008 (UTC)
No one seems to have a problem with it, so I'm proceeding ahead and putting in the navigation links as I edit each pub.--Rkihara 17:29, 18 October 2008 (UTC)
Here is a spreadsheet with the pub ids and tags. You may want to think about using pub ids because they cannot be changed by a user.--swfritter 18:56, 18 October 2008 (UTC)
Thanks! I've been cutting and pasting the code listings from the magazine page. This is more convenient. Pub ids are probably easier to deal with than tags, so I'll switch to those.--Rkihara 01:41, 19 October 2008 (UTC)
Hope no one minds, but because there are too many pubs verified by other people to easily notify, I will be installing the navigation hyperlinks without notification.--Rkihara 06:49, 19 October 2008 (UTC)
You might find this useful. I have done 1960.--swfritter 15:04, 19 October 2008 (UTC)
Thanks again! That will speed things up even more.--Rkihara 16:24, 19 October 2008 (UTC)
With a few more adjustments I should be able to kick such a list for most pub series in about five minutes.--swfritter 16:34, 19 October 2008 (UTC)
The hyperlinks for April 1944-Dec 1944 on your spreadsheet are a little scrambled.--Rkihara 17:13, 19 October 2008 (UTC)
No need to correct, I'm working around it.--Rkihara 18:11, 19 October 2008 (UTC)
Yes. A sorting error in the sql - fixed. Also, I cannot process pubs until the date has the month part filled in - 1953-00-00 for instance does not work correctly. Quarterly and sporadic titles will also be some trouble.--swfritter 22:23, 19 October 2008 (UTC)

Replacing the verifier for a publication

User:Bluesman added some notes to Venus, Inc. and also noted that the verifier has not used talk page in almost a year. That is true in that User:Animebill has not contributed on the wiki since 25 November 2007. I'd like to propose that if someone is inactive for more than a year that a current editor would be free to copy/paste a "stale" verification into the notes and to mark the publication as verified with their own name. I'd assume we'd post a notice on the talk page and also try to contact the editor via e-mail to see if they plan to return Marc Kupper (talk) 04:51, 18 October 2008 (UTC)

In the absence of a better plan, I'm ok with this, and would be ok with a shorter time period too. CoachPaul 06:25, 18 October 2008 (UTC)
Obviously the best solution would be software support for multiple verifications, which I think has long been requested. Failing that, the above sounds fine. I'd be good with a term as shorts as, say, 8 months. -DES Talk 13:35, 18 October 2008 (UTC)
With Primary (Transient) verifications, isn't that the same as "multiple"?--Bluesman 04:52, 20 October 2008 (UTC)
BNot really. True multiple verifiactions would support a variable number of varifiers, rahter than just two, and would not overload the "transient" verificarion type -- each verifier could mark his or her verification as permenant or trasnsient, perhaps. -DES Talk 04:55, 20 October 2008 (UTC)
I would suggest placing such information in bibliographic notes for a pub when applicable. The data is of no interest to the casual user. I don't know that is necessary to e-mail them. If they still have an interest they will eventually access their talk page. Eight months sounds good to me.--swfritter 16:41, 18 October 2008 (UTC)
Sounds reasonable as long as we leave a note on the original verifier's Talk page. As we have seen over and over again, it's always possible that two verifiers have a slightly different edition/printing/cover... Ahasuerus 17:00, 18 October 2008 (UTC)
It seems to me that the primary verification leaves a 'Verified' tag at the end of entry. Personally when looking for information I go to verified books. Could Verified (t) for transient be added to show against the book and the same guide line for no activity? Are you talking about a note on the user page, added by committee, that this user has been inactive for x period or longer, therefore please leave note of changes and (supersede the previous verification, if you wish to assume that responsibility). That makes it easier to leave a note and retains the 'respectful' attitude of the group. In a sense any user seeing this is notified to watch over their 'clutch' of verifications. --Dragoondelight 12:18, 20 October 2008 (UTC)
I don't see a need for a note added by committee. Any editor can use the Special:Contributions page (there is a link on every user and user talk page, labeled "User contributions") to see the date of the most recent wiki-edit by any given verifier. If that date is more than 8 months past, any editor may leave a note of the verifier's talk page that the editor is talking over the verification of a given publication. Then the existing verification data can be copied onto the bibliographic notes page of the wiki, and the editor can un-verify the pub, and reverify it, which will put the editor's name and timestamp on the verification. Normally this would not be done unless there is new or changed information to record. if the verifier has email enabled, the editor could optionally send an email to the verifier before taking this step. If no one objects, I will add this to Help:How to verify data. -DES Talk 19:26, 20 October 2008 (UTC)
Also if I am parsing this correctly, by 'copy/paste name data' into the notes, creates a separator for notes, which is especially useful for data that can not be verified from the book, and that would need to be added as an instruction to do this.
As for time, computer failure could make one lapse for three months or so, but rarely would leave the user in the 'wind' for six months. I suggest this 'policy change' also to include making sure members are aware that notes to the effect that they are on vacation to the 'Antartic' for a period would be greatly appreciated. I have noticed several moderators do this now. Thanks, Harry. --Dragoondelight 12:18, 20 October 2008 (UTC)
Such a visible indication might be useful. I advise you to add the suggestion to the ISFDB Feature List. -DES Talk 19:26, 20 October 2008 (UTC)

Reviews of Ace Doubles

A basically unresolved issue - the double or the the novels within. Now that we can link to reviews we can have the best of both worlds by linking to the ace double and the two halves which in most cases are reviewed as though they are two separate entities. It would be the editor's option to do either the Double or the novels/collections within or to do it both ways. Any subsequent editor who comes upon a pub where it is done one way should have the option of doing it the other way provided they don't remove or modify the existing review entry. See the reviews for the above in this issue of Galaxy. In addition, it would make sense that a review entry can be added for a significant discussion of a shorter piece within a collection/anthology; or the shorter pieces reviewed in a magazine like Locus many of which show up on award nomination lists.--swfritter 18:29, 20 October 2008 (UTC)

Astounding/Analog Essays - Who Wrote 'em?

Very Ugly!!! Most of these are signed "The Editor". Editors have been doing them three different ways. Some have ignored the fact that they are signed "The Editor" and entered John W. Campbell, Jr. and place them in the series (most of these were probably entered from secondary sources). Others have been crediting only as to "The Editor" and then adding to the series. Other have been assigning John W. Campbell, Jr. as a pseudonym of "The Editor" and then adding the Campbell entry to the series. Help strongly suggest use of "The Editor" unless there is incontrovertible proof as to authorship. This is probably a special case. There is little reason to believe that anybody but Campbell wrote them since his magazines were not committee edited as were a number of other magazines. There are those of us currently entering data who would like a consensus so that we can enter the data in a way that will sometime in the future correspond to the way the data will eventually be corrected by whoever volunteers for the task of removing numerous titles from the series and placing the correct titles in the series along with (depending on the consensus) unlinking cases where pseudonyms were used. My inclination, in this case only, is to make John W. Campbell, Jr. a pseudonym of The Editor. I might add that, luckily, not all of the essays have been added to the series - and I might suggest not adding anything to the series until there is a resolution.--swfritter 19:45, 20 October 2008 (UTC)

I agree that during JWC's editorship, he is pretty well known to have written all the editorials, and that when something was signed "The editor" it was intended that everyone know this mean JWC. I think that our records should reflect that, in some way. Whether this is best done by the pseud method, or by simply crediting these to JWC, or by some other method, I'm not sure. But simply crediting to "The Editor" with no indication that this is actually JWC seems mistaken to me. -DES Talk 20:03, 20 October 2008 (UTC)
I agree, with a quibble I'll get to in a bit. As far as the editorials go, if it just says "The Editor" I would (& have) enter it that way & then make that a pseudonym for Campbell; in that case it's the Campbell one that goes in the series. Similarly for other things signed "The Editor" (or just "Editor" or "Ed") which have real opinion-type content; it's pretty clear that Campbell cared enough about these things to insist on expressing things the way he wanted them expressed.
Other things such as In Times to Come & (even more so) AnLabs seem iffier, though what's there (beyond the mere numbers in AnLab) definitely sounds like Campbell. But for these I was convinced early on (by mhhutchins & swfritter) that it was iffy enough to enter it as it appeared, & I really think that makes sense. Campbell wrote, more than once IIRC, that Kay Tarrant had done more for the magazine than anyone could imagine, & I'd be surprised if she didn't do an In Times to Come from time to time. I hate to vote against consistency, but I think the editorials & such like are different in this case. (Mind, I think we ought to enter them all as The Editor; it's the application of pseudonyms after that I'm talking about.)
My quibble is with the original statement Most of these are signed "The Editor". I have relatively few Astoundings, but during most (at least) of the Campbell years of Analog they were almost always signed at the end "The Editor", but headed with a title & "An Editorial by John W. Campbell". For those I'd have no question: enter as Campbell, pseudonym of Campbell, Jr. And, indeed, looking at that (Editorial (Analog)) series I see that the only ones saying "The Editor" during the Campbell years are 1960 & 1961. (Probably not surprising, at that. I personally went through almost everything from then until Campbell died (& beyond), & if that's what the magazine said that's what our entries said after I did them.) -- Dave (davecat) 20:39, 20 October 2008 (UTC)
I agree with your distinction between editorals an other opnion essays, and largely factual essays such as AnLab or "In Times to Come" -- I should have mentioned the issue myself above. -DES Talk 20:56, 20 October 2008 (UTC)
As an afterthought: for anything as long & full of opinions as one of Campbell's editorials, if it's signed "The Editor" I'd enter it that way but make it a pseudo for whoever the editor is. Not just Campbell. For minor filler-type stuff, errata notices, etc. I'd assume some lowly member of the editorial staff might be doing it, but not a several-page editorial when there's one very much active editor at the helm. (Committee-edited things are obviously a different matter.) -- Dave (davecat) 21:03, 20 October 2008 (UTC)
It looks like the John W. Campbell credit started with the December, 1961 issue; before that it seems to have been "The Editor" only. Of course, that leaves us with the issue that John W. Campbell, Jr. is the canonical name; all have been correctly processed. The Jr. came off the editor credit when the mag switched from Astounding to Analog. The other columns are not as significant in that they do not give as much insight into Campbell's philosophy. Making Campbell a pseudonym of "The Editor" is self-documenting; otherwise we would have to leave a note in every issue explaining the signed name. And it is also one of the core standards.--swfritter 15:15, 21 October 2008 (UTC)
I vaguely think I may have found one or even two, during the years when they were headlined as being by Campbell, where that got left off; if so, it/they was/were still signed "The Editor" at the end. If I'm right, I think I credited The Editor, made it a pseudonym, & put in a note. Given Analog's copyediting standards, I'm confident that such an omission was not significant but just a slip. (And I may merely be remembering the first one or two when, working backward, I reached the point where there wasn't a Campbell credit at the head.) -- Dave (davecat) 16:25, 21 October 2008 (UTC)
The top of the page credit to John W. Campbell seems to have started with the December 1961 issue.--swfritter 18:18, 27 October 2008 (UTC)
I agree that we will want to enter these editorials as by "Editor" and then set up variant titles. That way our user will know that the essay was not signed by Campbell and that it was attributed to him based on secondary sources, so there is a small chance that it was actually written by somebody else. Ahasuerus 17:29, 22 October 2008 (UTC)

"Tor Double" as a Publisher

It looks like "Tor Double" is currently set up as a full fledged Publisher, presumably on purpose. I assume that it was done that way to make it easier to find all Tor Doubles? Ahasuerus 03:34, 24 October 2008 (UTC)

I'm sure there's more than a few other "publishers" that were created in order to list publisher series (as opposed to author series), e.g. "Ace Specials". Now that we have publisher Wiki pages, I believe these series should be moved to the Wiki pages. See how Bill did the Corgi SF Collector's Library for a good example. MHHutchins 04:27, 24 October 2008 (UTC)
I agree. I would be inclined to create a list of these on the Publisher:Tor wiki page, or a page linked to from that, and change the publisher to Tor. There is a downside: it would not be quite as easy or obvious to a user to find the list from a publication record -- it might be a good idea to also include a note with a link to the wiki page in each such pub. But given how strongly people have felt about publisher edits, i won't make such changes without wide support. -DES Talk 12:02, 24 October 2008 (UTC)
There is already a list on the wiki page Publisher:Tom Doherty Associates#The Tor Doubles, with pub links to the db. It seems to have been created by Marc Kupper -DES Talk 12:21, 24 October 2008 (UTC)
Well, it's quite easy to create the Wiki pages from the database if the Publisher has been set to "Tor Double" or such a distinguishing imprint - the Corgi one was rather harder as there was no such convention in the database, and I've had to track them down from secondary sources (and it's still not complete, months later). The Tor Doubles list looks rather better than mine but the Corgi SF Collector series weren't numbered, and unlike the Ace and Tor doubles, many exist in both plain Corgi and Corgi SF Collector versions. I suspect the SF Collector pubs would look quite good in Publisher Display if they were given a Publisher-name of their own, but a lot of them aren't my publications and I'm loath to go ask every verifier if I can change them when I'm still not entirely sure which ones would need changing. BLongley 18:38, 24 October 2008 (UTC)
Also, I don't think splitting publishers to create publisher/publication series is always going to be a good idea if it leads to a lot of fragmentation: if I knew that "Corgi" didn't include "Corgi SF Collector" publications I could go find them with a second search, and I can cope with our current "Tor" and "Tor Double", or "Ballantine" and "Ballantine Adult Fantasy", and even "Ace", "Ace Double", "Ace SF Special [1]" and "Ace SF Special [2]" (although note that the Wiki pages for the last two don't work). We could get into some major fragmentation with some publishers though: would we want "Tandem" split into "Tandem Science Fantasy", "Tandem Fantasy", "Tandem Sci-Fi", etc? BLongley 18:38, 24 October 2008 (UTC)
I support the creation of Wiki pages to cover short-falls in the database or the display software: for instance you may have seen me updating Series:Star Trek Pocket Books recently as we can't show the books in both the Pocket and Titan Series. But I want those to feed back into improvements in the database. What I can't support is the movement of useful data OUT of the database into the Wiki. That's data loss to me - just because we have a good Wiki page for Tor Doubles doesn't mean we can remove the distinguishing features IN the database. I'm open to suggestions as to better ways to KEEP the "Tor"/"Tor Double" distinction in the database somehow (it doesn't have to be via a different publisher name) but we shouldn't be copying useful data to the Wiki and then destroying it in the database. Someone found Tor Doubles and Ace Doubles etc useful enough to distinguish, saying "it's on the Wiki, we don't need it in the database" is just plain wrong to me. BLongley 18:38, 24 October 2008 (UTC)
There is nothing sacred about the relational databaseDES Talk 19:47, 24 October 2008 (UTC)
Grrr.. just lost an hour of responses due to Wiki edit conflicts. Might find them tomorrow. But you're starting from the wrong POV: The database IS sacred, the Wiki is Support and that's it. Go reread Help:Contents/Purpose. BLongley 21:58, 24 October 2008 (UTC)
If you hit an edit conflict, there is no reason at all to lose any text. See Help:Edit conflict. On this point Help:Contents/Purpose is outdated and should be changed. When it was written the wiki carted only clearly supplemental data and could be seen as secondary. It is now being used to carry significant amounts of essential data, and even the "supplemental" data has become sufficiently organized, and present in sufficient quantity, that the wiki can no longer be automatically considered "Secondary". It is true that "The ISFDB Wiki supports the ISFDB", but it is equally true to say "The ISFDB relational database supports the ISFDB". The ISFDB per se is the collection of data housed in both, and which might be housed in some different form in future. Both the wiki and the relational db are merely tools to serve the ultimate end, gathering, preserving, and displaying bibliographic information about speculative fiction in a useful manner. -DES Talk 22:26, 24 October 2008 (UTC)
I am sure Bill will post his reasons when he wakes up tomorrow, but in the meantime let me explain how I see the interplay between these two beasts.
And I'll be glad to listen to them, and honestly consider them, and agree or give what seem to me valid resons to disagree. I'll try not to get overly excited. -DES Talk 01:16, 25 October 2008 (UTC)
It is certainly true that there is nothing special about the fact that we are using a relational database, MySQL, to drive the application. After all, we used a home-grown database written in C (if memory serves) for years and we may use something else, e.g. BigTable, in the future. The key thing, however, is that the ISFDB application is currently tightly integrated with the database and performs all kinds of housekeeping for us so that we don't have to worry about deleting Authors with no Titles or accidentally deleting Titles with associated Publications. It's not perfect (empty Series records, anyone?), but it gets the job done.
The Wiki, on the other hand, is not tightly integrated with the application/database, therefore any significant changes on the database side, e.g. Publisher merges, Publication deletions or Series renaming will not be automatically applied to the Wiki side. Until and unless we are able to integrate the two -- something that we requested a long time ago and Al promised to look into at some point -- we will always run the risk of certain Wiki pages becoming effectively invisible when the associated database information changes. Based on experience, this may not be a huge risk, but I keep it in mind when deciding whether to enter, say, Publisher specific information in the Publisher record or add it to the associated Wiki page.
On the other hand, the Wiki works very well for keeping track of projects, organizing free text pages like Help, communications between editors, etc etc and in that sense it's an indispensable part of the ISFDB effort.
Finally, let's try to keep these issues in perspective and not get too excited about them. The world will not end whether we put certain bibliographic information in one set of tables or the other :-) Ahasuerus 23:22, 24 October 2008 (UTC)
User:Ahasuerus is correct that the db is currently more tightly coupled to the application, and more customized for the ISFDB project's needs, than the wiki is, and there is no knowing when or how or if that will change. That has many advantages, and some disadvantages. There are steps that can be taken on the wiki side -- using categories and lists, for example -- to make it easier to find any given page of interest, and less likely that pages will simply disappear or be overlooked. But it is true that such strutures must be built, to some extent on a one-by-one basis, and don't come automatically as the db/web application does (of course that didn't drop from the sky, it represents lots of hard work. But that may be less apparent to those of us who use it but never helped build it.) -DES Talk 01:16, 25 October 2008 (UTC)
Both the wiki and the db are tools to help us achieve our common purpose. It would be possible, but ill-advised, to keep all the publication level data in the wiki -- that is not what the wiki does best. It would be possible, but ill advised, to keep much of what is now stored on the wiki in the db. Thinks like the publisher pages are in the middle, IMO. Putting them in the db would enforce a certian amount of organization and consistancy on them, but while we are still determining what information we want to store, and how we want to format it, such consistancy might actually be a bad thing. And when data is inherently inconsistant, full of exceptions, a wiki may actually be a better way to store the data in any case. We may need beter orgination of the wiki pages, to keep significant pages from becoming orphans -- we ma need top generate wiki-lists from db dumps on a regular basis, to help keep the two in sync. -DES Talk 01:16, 25 October 2008 (UTC)
I am more than willing to discuss, for any particular case, whether a particular kind of data is best recorded in the db records, on the wiki, or both (or neither if it seems of no value). All that pushes my buttons (perhaps overly) is what seemed to me like "if it's important, it must be in the db, not the wiki. Case closed." That I am unwilling to accept. Probably that wasn't what bill really maent, and if I had seen his longer text that was unfortunately lost (and I know that can be very frustrating) I would have realized that. -DES Talk 01:16, 25 October 2008 (UTC)
I never said "must be in the db, not the wiki". I am perfectly happy to have all the database data also available in the Wiki, or someone else's Wiki, or someone else's database, online or offline. What pushes my buttons is the suggestion that we can move data out of the database into the Wiki and it will make no difference. It makes the data unavailable to me offline, or would require massive amounts of programming to use - I can work with the database backups, I can't work with the wiki data even if it's available in the backups. Have you noticed how many people have a working copy of the ISFDB (data + user interface software) locally? Less than a handful. How many have got ISFDB data, user interface, and the whole ISFDB Wiki working locally? I suspect "None". How many people have used the data? Enough that I keep coming across our data on other websites. Some wouldn't even exist if it wasn't for the fact that we provide the data for free. BLongley 15:43, 25 October 2008 (UTC)
I want a lot of the Wiki data incorporated in the database - the Wiki is a useful tool for demonstrating what sort of things we should be aiming to collect, and how we think they might be structured. The Publisher data is a good recent example - it's gone from being completely anarchic free text, to having each publisher recorded separately with links to the Wiki, to having some fields recordable in the database (Wikipedia Entry, Web Page, and Note). Hopefully we can look at what people are doing on the Wiki and learn from that - e.g. if people are often recording addresses for publishers, we can add an address field, or an address history table. If they're recording ISBN ranges, we can add that as a separate field rather than overload notes even more. Some things I doubt we'll ever do - incorporating coverart images into the main database is probably never going to happen as it will massively increase the size of the backups and put people off using the data. But we might learn enough from such Wiki-work to reconsider whether actual images should be on the publication or the coverart record. (I suspect we will need both - maybe PUBART for exact covers and COVERART for "same artwork, even if they stomped all over it with different text".) BLongley 15:43, 25 October 2008 (UTC)
The big issue (for me, at least, and I suspect I'm not alone) is that I finally found a place where I could add data, and get it back in useful format. Moving useful data from the database to the Wiki is about as welcome as replacing all your dollars with the Zimbabwean equivalent, or storing it in an Icelandic bank. It might look like a painless conversion at the time but the consequences are rather severe. I use and contribute to ISFDB because the data we put in is as useful as the data we get back. The Wiki is a useful area to record stuff that doesn't, and maybe never will, get into the database. ISFDB + Wiki will be superior because of that - but ISFDB + Wiki is an online-only resource, ISFDB (database alone) is something I do not want to lose data from. BLongley 15:43, 25 October 2008 (UTC)
I was not, in this instance, suggesting moving data out of the relational db. I was and am suggesting moving it from one field of the db (publisher) where it fragments publisher listings, to another field, (notes) where it doesn't. I am suiggesting tha each publication that now has a publisher listing of "Tor Double" will insted have a notes entry of "Tor Double #nn" plus a link to the existing wiki-page, which has links back to all the other Tor Double pubs. The online searcher could use these links to find all Tor Doubles, the offline searcher could search on the notes field entry. No data is lost, for either kind of user. -DES Talk 17:16, 25 October 2008 (UTC)
I know you didn't suggest "moved to the Wiki", Mike did. You agreed and then pointed out that something in notes might be appropriate - thanks, that's a bit better than a wholesale move. That's still a move from structured data to unstructured data though, and makes SQL searches a bit more complicated. I'm not sure where to draw the line of publisher regularisation: we have some so standardised that they're useless (e.g. "Scholastic" I think should have it's Point Horror/Point SF/Point Fantasy split out). "Tor Double" is quite useful at the moment and not especially hurting searches online - if someone searches for "Tor" alone they'll find both as the ISFDB interface adds the wildcards. Offline, I've learnt to end every publisher search with a "%" anyway - and usually start with one too in case people have been regularising the other way round. BLongley 19:31, 26 October 2008 (UTC)
There are some bits of data that I would think should move out of the relational db altogether: publisher street addresses for example. Those vary over time, and are IMO better kept on a wiki page than in any sort of relational db field. But the Identificatioin as part of the set of Tor Doubles is not such a bit, IMO, nor are any of the bits for which we have fields set up. -DES Talk 17:21, 25 October 2008 (UTC)
Addresses are something I'm not sure anyone really wants anywhere. I suspect that if we had enough of them we might spot a connection between imprints or publishers not currently noted, but so far it doesn't look worth the effort and I've stopped recording such. Although it's nice to see when a publisher is so into SF it will call its HQ "Astronaut House". ;-) BLongley 19:31, 26 October 2008 (UTC)
, nor is it inherently inferior to the wiki. Both the relational db and the wiki are tools to support the overall ISFDB, which is composed of both of them. I cannot agree that moving data from the relational database to the wiki database can sensibly be called "data loss". In some cases such moves are advantageous, in other cases they would be unwise, in yet other cases such moves have both benefits and costs. In the present case I am not advocating moving this information out of the relational db: what I am advocating is moving it out of the "publisher" field (currently being overloaded with this data) and into the notes field, along with a link to the relevant wiki page. As the wiki already has links to the publication records, this will provide a direct chain of links form any one of the Tor Doubles to all of them.
"Tor Double" is not a publisher, nor an imprint, nor even a line. It is a publisher marketing series. If we had support for publisher series, or for placing books in multiple series, that would be a reasonable way of dealing with the matter. We currently don't. Placing this in the publisher field is a hack -- an overloading of that field -- and it contributes to the highly undesirably and currently severe problem of publisher fragmentation, which makes publisher searches far less useful than they should be. We don't put "DAW Collectors Book #123456" in the publisher field, nor in the ISBN field if there is also an ISBN, although this would make searching on the DAW number easier. We put it in the notes. "Tor Double #37" should be treated similarly. There is a slight cost to this -- it won't be as easy to search for the Tor Doubles from the web interface, although it will be easy enough using SQL on a local db backup. Fewer than 40 pub records are involved: this change would be easy enough.
You (Bill) say that "someone" found this useful enough to do. I don't know if great thought was given to setting it up this way, or if that "someone" did it quickly on little thought, and others copied the format more or less automatically. If it is in fact useful, let those who think so make that case. The fact that someone did this at some time is not a good reason not to change it to a better way. I think that moving this data out of the publisher field is a better way. If you disagree, what are your arguments? -DES Talk 19:47, 24 October 2008 (UTC)
DES wrote "There is nothing sacred about the relational database, nor is it inherently inferior to the wiki." At present the db is inferior in that changes to the structure take much more planning and can only be done by a couple of people while the Wiki can be changed by anyone. The wiki's downside is little support for cross referencing things. Another item is that the wiki is not included in the "official" ISFDB database distribution meaning the already considerable efforts to organize magazines and publisher related things are not included. Marc Kupper (talk) 21:24, 24 October 2008 (UTC)
Backups of the wiki are, unless i am mistaken, already available to anyone who wants them, but perhaps they should be more closely integrated with the backups of the relational database. i suspect that the time is rapidly coming when the relational database will not make sense without the wiki -- no that is saying too much, but that significant data will not be available without the wiki, and significant features will not work properly. I don't see a problem with this. -DES Talk 22:26, 24 October 2008 (UTC)
Wiki backups are not publicly available at this time since they contain passwords, e-mails and other private information. I used to download Wiki backups, expunge all private data, delete old versions of Wiki pages and then post the "neutered" result. However, this functionality hasn't been available since the move to the current host 6 months ago. Ahasuerus 23:22, 24 October 2008 (UTC)
As ISFDB does not support including a title in more than one series people long ago started using the publisher field as a way of creating publisher series. I believe this occurred before we had the ability to view all of the publications by a publisher though once that happened it accelerated the process.
It was quite easy for me to create the Tor Double wiki-list as someone had set this up as a "Publisher" with the main effort being that the list was missing some titles. Fortunately, Tor numbered the books meaning it was not too much work to track down the missing titles on AbeBooks, etc.
I don't have a problem with someone editing the publications to replace "Tor Double" with "TOR, A Tom Doherty Associates Book, New York" <wink> and also adding a publication note
<li>This publication is part of the <a href="http://www.isfdb.org/wiki/index.php/Publisher:Tor_Double"><b>Tor Double</b></a> publisher series.</li>
There's also Ace Double, Belmont Double, and Signet Double. Marc Kupper (talk) 21:24, 24 October 2008 (UTC)
Ace doubles may be a special case, given their very long and significant history. It could be argued that they acted as an imprint. However, when and if we have support for publisher series or multiple series, i would like to see the ace doubles have simply "Ace" in the publisher field. The other "doubles" you cite should probably be converted now, but lets take one at a time and then we will have a prototype to work from. Ace is as god as any. So far I see exactly one person objecting to moving the "Tor Double" indication from the publisher field to the notes field. I have not seen any substantive arguments for this position beyond "The db IS sacred", which I find less than persuasive, or even responsive. I admit to one downside to the move: it would be harder (but not impossible) to directly search for the books involved as a group. I think that the gains outweigh this, but admit that a rational person could differ on that point. Does anyone but Bill actually object to such a move? Does anyone have substantive arguments against it?. -DES Talk 22:26, 24 October 2008 (UTC)
I object to the change!Kraang 02:25, 25 October 2008 (UTC)
Please explain what the gains are, DES. And how we could "search for the books involved" if you made the change right now. Don't get me wrong - "when and if we have support for publisher series or multiple series" I'll happily support some more standardisation of publishers. BLongley 16:01, 25 October 2008 (UTC)
The gains are: list the actual publisher/imprint more accurately; Let a publisher search on Tor return all the Tor books, not just the Tor books that are not doubles; reduce publiaher fragmentation; reduce field overloading; link to the existing wiki page that provides organized information not appropriate for any single publication; provide info in the publicatiuon db record on the number in the Tor doubles sequence. Online searchewrs could easily go from any Tor Double to the wiki-page, and from there to all pubs involved, or could search the wiki for the phrase "Tor Double". We could also set up a wiki page on Publication series that might be easier to find. Offline searchers could search the notes field, rather than the publisher field, for the text "Tor Double". I don't claim that there are no tradeoffs here, but I do think there would be a net gain. What do you see as the net losses? -DES Talk 17:17, 25 October 2008 (UTC)
But we wouldn't "list the actual publisher/imprint more accurately" - we've already got people complaining that Tor US and Tor UK and Tor Australia are being confused. BLongley 23:06, 26 October 2008 (UTC)
An ISFDB publisher search for Tor does return all the Tor publishers - along with publishers with "sTORy" or "sTORies" in the name, or "ediTOR" or any publisher based in "TORonto". Lots of room for improvement there, I agree 174 results is not good. There might be about sixteen relevant results, but I'm in no hurry to remove one of the useful ones just yet. Change the search, not the data. BLongley 23:06, 26 October 2008 (UTC)
"reduce publiaher fragmentation" is a good idea until it means loss of imprint data, or whatever people have found needed to add for disambiguation. That's why it's been so slow so far. BLongley 23:06, 26 October 2008 (UTC)
Searching for "Tor" in the wiki is useless as three-letter words aren't indexed. "Tor Double" works better, yes, but the Wiki software is fundamentally broken for such short searches - we'll encounter similar problems for NEL, NAL etc. BLongley 23:06, 26 October 2008 (UTC)
"Net losses" - well, just having to add another table into SQL queries is a pain, and affects performance. And such a move does overload the Notes field so "reduce field overloading" is another benefit you cite that I disagree with. BLongley

Omnibus author/editor

According to Help:Screen:NewPub:

Author - The name of the author of the publication. [...]
  • Editors, authors, translators, etc. If the book is the work of an author, use that name. If it is an anthology, use the name of the editor. If it is an omnibus leave the field blank.

The last sentence seems to be wrong since you can't even submit a New Pub form without entering something in the Author field. Do we want to change it to something like "If it is an omnibus, then enter the name of the author(s) whose work is reprinted in the omnibus"? Ahasuerus 00:58, 27 October 2008 (UTC)

That seems to be what we in fact do and it seems sensible to me. Any reason not to make this correction in the help. -DES Talk 03:22, 27 October 2008 (UTC)
I agree - correct it but wondered what the original intent was. Marc Kupper (talk) 07:06, 27 October 2008 (UTC)
I vaguely recall that the way Authors were entered for Omnibuses changed some time in mid-2006, so perhaps it was an attempt to capture the functionality as it existed then. Ahasuerus 13:09, 27 October 2008 (UTC)
I'd rather shorten and to delete the part about "If it is an omnibus leave the field blank." If there is a desire to include "omnibus" then change "If the book is a single-author collection ..." to "... collection or omnibus ..." Marc Kupper (talk) 07:06, 27 October 2008 (UTC)
"... collection or omnibus ..." sounds reasonable. Ahasuerus 13:09, 27 October 2008 (UTC)
Change made, see if you like the result. -DES Talk 16:19, 27 October 2008 (UTC)
Personal tools