ISFDB:Community Portal/Archive/Archive02

From ISFDB

Jump to: navigation, search

Contents

Fall 2006 tasks

(moved from Al's Talk Page)

I've done a bit of bug fixing to get started, but I figure I can do that forever and not close them all out, so I started working on the verification support (we can start prioritizing the bugs, a generate a "must fix" list that should be closed prior to launch). Alvonruff

We definitely need to prioritize bugs and group similar bugs together. Sometimes a simple fix can take care of two-three bugs. However, I am a litle leery of leaving more serious bugs in the code since they may be masking other bugs. More than once I was hesistant to record a bug because I wasn't sure whether it was a new one or a slightly different aspect of an old one. Ahasuerus 19:00, 1 Oct 2006 (CDT)

Implementing verification is actually not going to be too difficult (about half done already), but there is one important question: when a user verifies a publication against the primary source or a valid secondary sources, Alvonruff

Did we ever decide whether verifying against a "valid secondary source" should be sufficient to raise the verification flag? Even the best secondary sources out there are known to have bugs and phantom titles (including bibliographies compiled by Reginald :) ), but on the other hand, primary verification is not perfect either. Editors may misread publication data (ISBNs, spelling, etc) just as easily as the authors of reputable bibliographies, although it does eliminate one layer of transcription-related errors. Ahasuerus 19:00, 1 Oct 2006 (CDT)

should that verification immediately update the database, or should that verification be submitted for later integration by a moderator? The feature can't be used for spam, but it would still be possible to perform low-grade vandalism (of a sort). Alvonruff 17:46, 1 Oct 2006 (CDT)

I think it depends on whether we decide that "reputable secondary sources" should be allowed for verification purposes. If we don't allow them, then the risk of vandalism is pretty low and it can be made a one step process. However, if we do allow them, then we may want to make verification a two step process. First, an editor raises the verification flag and fills out a free text "source of information" field. The latter is required and the help text should say something like "Enter the source of your information. If you own or have personally seen the work in question and recorded its data, then enter 'physical verification'. If you are using a reference book or another source, then enter the name/title of the source here." It would help moderators decide whether to approve or reject a submission. Ahasuerus 19:00, 1 Oct 2006 (CDT)
Currently, users will have a checklist of approved secondary sources, as well as the primary source itself to select from. Only moderators will be able to edit the list of approved secondary sources (which may be print-based or online resources). So a user would click "verify this pub", and get a choice of "I've got the book", "Tuck", "Reginald1", etc... and can click "YES" (they compared the work to the reference) or "N/A" (the reference predates the publication, or is incomplete). If there is some new reference, the user can request to have it added to the list of approved references. I think the vast majority of people will simply have the book in question on hand, while a much smaller number of people will have reference materials. I'm looking to make the interface as simple as possible. Alvonruff 06:04, 2 Oct 2006 (CDT)
Oh, I see which option you ended up going with! Sorry, my short term memory is not what it used to be, so I had forgotten the outcome of that discussion. If that's the case, then I think it's fairly safe to make verification a one step process.
Some of the other tasks that I think we would like to have completed before we go live are: (a) copying Publication contents to avoid the need to re-key all stories in a monster anthology just because a new edition has come out, (b) deleting individual entries from Publications, (c) perhaps adding a free text field for Delete Publication and Delete Title to make it easier for moderators to decide whether the request is legitimate or not. A pseudonym editor of some sort would be nice to :)
If we could make the submission mechanism ask the editor if he would like to auto-merge the new Title(s) with similarly titled existing one(s), that would be even better, but I am not sure how long it would take. Having entered a bunch of anthologies, I have found that hunting down pre-existing stories and merging them with the newly entered ones is not only time consuming, but also error-prone :( Ahasuerus 09:04, 2 Oct 2006 (CDT)
Oh, and I forgot to mention that a few days ago a kindly stranger donated a very detailed DAW bibliography that used to be distributed by DAW. Ahasuerus 17:29, 2 Oct 2006 (CDT)

Tamu.edu downtime

On 10/03/06, Tameu.edu was down for a number of hours, which was commented upon by different posters on r.a.sf.w. This has been happening more frequently lately and I wonder if there is a way to find out more about the root causes of these occurrences? Ahasuerus 07:40, 4 Oct 2006 (CDT)

My first edits

I tried my first editorial changes and documented the issues I encountered on my own talk page (so I wouldn't clutter this page). If anyone has a few free minutes, I appreciate any comments. PortForlorn 01:47, 19 Oct 2006 (CDT)

Done and the results added to ISFDB Editing Guide when appropriate. Ahasuerus 18:30, 20 Oct 2006 (CDT)
Thanks much for the extensive comments and advice.
We are here to serve Man! (Well, other species too as long as they go well with ketchup). Ahasuerus 15:20, 22 Oct 2006 (CDT)
I was glad to see that you could reuse them for community editing notes also. As always, answers lead to more questions. Right now, I have three.
1) Now that I've finished correcting the titles for Mercedes Lackey, what more is needed to mark her entry in the Top 300 list as internally consistent?
Al is currently working on what we call the correctness flag, but it's not quite ready for prime time yet. For now, whenever I finish the consistency pass for an Author, I click on the Author's Wiki link up top (in this case Author:Mercedes Lackey) and enter {{subst:consist}} there. I then update Bibliographic_Projects in_Progress#Short term_projects with the Author's name and a brief description of what has been done. Come to think of it, I am yet to enter our October edits there and we really need to do it within 30 days while the edits can be readily accessed via Special:Recent Changes. Ahasuerus 15:20, 22 Oct 2006 (CDT)
2) I noticed that some of the publications' titles were entered as "Series Name" + "Book in Series Number" + "Title" (e.g., http://www.isfdb.org/cgi-bin/title.cgi?5500). I have the '92 pub in my hand and it seems that all this info was on the cover and repeated on the title page. I've read through the editing guidelines and this page and haven't seen a clear recommendation for what to do when the title page gives redundant info. I'm tempted to revise the pub titles to eliminate the series and sequence number info but I'd rather find out if there is a preferred approach such as "exactly reproducing" what is on the title page.
This is one of the big and non-trivial differences between Publication and Title records. Publications are supposed to reflect whatever is specified on the title page, which may have "The Prestige: A Novel" printed in the hardcover edition or perhaps simply "The Prestige" in a paperback reprint. Some editions will include series information on the title page and some won't. There are special rules in the MARC-21 standard and in the OCLC guidelines for recording Publication information, including "subtitles" and many other data elements, but that's for trained professionals to use and abuse. We can't expect our editors to follow these rules without getting a BA in library science first, so the rule of thumb is to enter whatever you find on the title page.
Titles, on the other hand, are a higher level of abstraction and would normally drop any obvious subtitles (e.g. ": A Novel") and series data. If the titles of two or more editions are clearly and meaningfully different, e.g. "Fugue for a Darkening Island" has also appeared as "Darkening Island", then we have to create two or more title records and make all but the first one "variant titles" of the title used on the first edition.
Sometimes it can be a judgment call: Is the proper Title of Tolkien's famous book "The Hobbit" or "The Hobbit, or There and Back Again"? And at what point does the difference between the titles used on two different editions become big enough to make one of the title an "alternate title"? Ahasuerus 15:20, 22 Oct 2006 (CDT)
3) I've also noticed that some of the Amazon URLs for cover images are now broken links. I assume that the titles have expired at Amazon and the image has been deleted. Has there been any discussion of capturing the thumbnails and saving then somewhere permanently or what should be done about broken image URLs? PortForlorn 00:28, 22 Oct 2006 (CDT)
That's a good question. I don't know if we have given it much thought considering all the copyright and disk space implications/ramifications. Al? Ahasuerus 15:20, 22 Oct 2006 (CDT)
Amazon often has an image link which displays on their page as "No Image Available". The urls are NOT images of the text "No Image Available", but are nontheless unique, so there must be information imbedded in the url which indicates whether or not an image is available - I haven't cracked that as yet (well I haven't tried...). Most of these come from dissembler which includes the url in its submission. If I figure out how to detect a blank image, then I'll change dissembler not to incude them. I don't think any images have timed out. Alvonruff 16:19, 22 Oct 2006 (CDT)
I've come up with a couple more issues:
4) I found yet another example of the strange ISBN numbers in the DB (see Open Display Bugs). This time I tracked it back and discovered that someone has entered an Amason ASIN number (B000BKRV38) in the ISBN field. This if found in this pub listing: http://www.isfdb.org/cgi-bin/title.cgi?37834 But there is no way to edit the pub and correct the ISBN without encountering the bug when you click on the pub title. Is there any other mechanism for reaching the pub edit page? PortForlorn
Al will correct me if I am wrong, but if memory serves, this is something that Dissembler used to do before Al added extra checks to the algorihm. Once the editing software is fixed, we can probably run a wholesale conversion of all ISBNs that are actually ASINs in disguise. Ahasuerus 01:23, 24 Oct 2006 (CDT)
5) The title "Tommorow, the Stars" is in a strange state. It is listed here (http://www.isfdb.org/cgi-bin/pl.cgi?TMRRSTRS291951) as authored by Heinlein, but the Contents shows it as an anthology by Pohl and Merril writing as Heinlein! The pub listing (http://www.isfdb.org/cgi-bin/title.cgi?189500) shows the editor as Heinlein but it is supposed to be a varient title of http://www.isfdb.org/cgi-bin/title.cgi?35080 edited by Pohl and Merril. Now I've checked the web biblios for Pohl and Merril and looked through Contento and, as far as I can tell, they have never claimed to have written under the pseudonym Robert A. Heinlein :-). I realize the wikipedia is not an Authoritative Source but that explanation (http://en.wikipedia.org/wiki/Tomorrow%2C_the_Stars) is that "Heinlein wrote a six-page introduction in which he discussed the nature of science fiction, speculative fiction, escapist stories, and literature. None of the stories had previously been anthologized. According to science-fiction historian Bud Webster, however, Heinlein's introduction and name on the book were his sole contributions; the actual selection of the stories, and the work involved in arranging for their publication, was done by Frederik Pohl and Judith Merril. This is confirmed by Virginia Heinlein in Grumbles from the Grave (without mentioning Pohl or Merril) and by Pohl in chapter 6 of his autobiography, The Way the Future Was (Del Rey 1978)." So it seems that this title should be listed without the VT and as edited by Heinlein, Pohl and Merril. It appears the there is an incorrect pseudonym link along with an incorrect VT link. However, I have no idea how to fix these problems. PortForlorn
Yes, this is quite messy. At one point Grendelkhan did a fair amount of work on Heinlein and he may have looked into the issue in depth, but he hasn't been seen around here in some time :( The current situation is not quite as bad as it may appear, though. The Publication record correctly states that the titular editor was Heinlein -- remember, this is one of the big differences between Publications, which reflect what's in the book, and Titles, which try to give a more accurate, behind the scenes picture, revealing pseudonyms, abstracting titles, etc. The fact that there are two Title records is simply a reflection of the (admittedly suboptimal) way pseudonyms are currently implemented.
I figure that for now the easiest way to improve the record would be to add Heinlein to the list of "real" editors and add a Note field to the Title, which may help unconfuse J. Random User. Ahasuerus 01:23, 24 Oct 2006 (CDT)

ISFDB:Operations

ISFDB:Operations has been created. I'll try to update it weekly or thereabouts. If you can think of any other operational issues that need to be addressed or monitored, please chime in :) Ahasuerus 16:53, 20 Oct 2006 (CDT)

WorldCat now available on the Web

Please note [Sources_of_Bibliographic_Information#Aggregate_Library_Catalogs_and_Search_Engines|this change]] to WorldCat's availability. Basically, the whole thing is now available on the Web, which is a huge change to OCLC's modus operandi. There is no free Z39.50 or comparable access for now, but unlike OCLC Fiction Finder, they don't hide URLs or use session-based obfuscations, so we can link directly to their images, records etc. The Web interface is somewhat non-trivial and subject to change since this is a beta project, so I wouldn't be in a hurry to write HTML scraping tools, but this is potentially a big change. They have ALL kinds of stuff (billions and billions of items cataloged!.. well, OK, 1.3 billion), including fanzines, manuscripts and heaven knows what else. Ahasuerus 14:08, 24 Oct 2006 (CDT)

DAW list - discussion

Please note that there is an ongoing discussion of how to best incorporate user-submitted corrections into the master (publisher-provided) list of DAW Publications at Publisher talk:DAW. Ahasuerus 01:01, 25 Oct 2006 (CDT)

Blocking users preventing all editing

I am trying to figure out what it was that I did while blocking user "Zexter", which, as we know, had the unintended side effect of preventing any editing of the ISFDB Wiki. I recall entering the duration of the block as "240" and not "240 hours" -- see Special:Log/block on 10/27/06. Apparently, entering a number without the word "hours" in that field didn't have the desired effect since Zexter vandalized his User page again on 10/28/06. I then blocked him again, this time for "2400 hours", which apparently caused the software to block all Wiki users.

I wonder if there is a (known?) bug in the MediaWiki code that confuses it whenever there is a pure numeric in the block log history? Ahasuerus 11:39, 30 Oct 2006 (CST)

Improving Verificiation support

Now that Verification support is live, here are some useful things that we could do with it:

  • A list of recently verified publications
  • A list of publications verified by individual editors
  • A list of publications that have been explicitly marked "Not Verified"

[Al's ideas to be added to later on]. Ahasuerus 13:44, 30 Oct 2006 (CST)

A Grab Bag of Questions on Adding New Publications

I'm new to ISFDB, wanted to look up a book, saw that ISFDB does not have it, and so I added the thing while reading down the help at Editing:Adding_a_New_Publication

A number of questions that came up while entering the field data for this publication.

Author - For my book it was straightforward but I did not see any mention of how to deal with people who may be listed but are not authors. For example

  • "with" Person's name (usually a co-author but it implies the person is less important)
We have no support for "junior co-authors" at this time, so all authors are created equal, at least for now :) There are also related unresolved questions about assistant encyclopedia editors and about ghosts writers: if the ghost did 80% of the work, was it pure ghost writing or was it a collaboration? 90%? 95%? Ahasuerus 23:56, 2 Nov 2006 (CST)
  • Editor
At this point, the ISFDB software supports Editor records for anthologies, shared worlds, etc, but not for single author works like Collections and Novels. There have been multiple requests to add this field since some houses, e.g. Tor, have been making editor names public the last few years. Al is thinking about it, but for now the best we can do is put the name in the "Notes" field. Ahasuerus 23:56, 2 Nov 2006 (CST)
  • Illustrator
Al added "INTERIORART" to the list of "[Contents] Entry Types" a few months ago and there is support for "Cover Artists" at the Publication level, but nothing else at this time. Ahasuerus 23:56, 2 Nov 2006 (CST)
  • Photographer
No support at this time, I am afraid. Ahasuerus 23:56, 2 Nov 2006 (CST)
  • Translator
This is an interesting questions. Are translations "derived titles" or just one of many Publications of the same Title? It's on Al's list of things to do once the bugs are addressed later this year. For now, I just enter the translator's name in the Notes field. Ahasuerus 23:56, 2 Nov 2006 (CST)

There are other roles such as "Introduction" that are often mentioned on a cover but I can't recall a book that mentions those on the title page.

Introductions are a special case since they are separate "Titles" in our parlance. When entering a "Collection" Publication that contains one (or more) introductions, add the introduction Title(s) as "Essay(s)". The current convention is to specify the name of the book in which the introduction (preface, afterword, etc) is published in parenthesis -- otherwise we will have hundreds and thousands of identical Titles, all called "Introduction".
If you are entering a "Novel" Publication, it's a little trickier since the "New Novel" data entry option doesn't let you enter anything but Novel-specific data. However, once the novel in question has been submitted and approved, you can pull it up for editing and use the "Add Title" button to add introductions, afterwords, etc to the Contents section as needed. We may simplify this two step process later on, but it's not quite as simple as it may appear since the "New Novel" form creates both a Publication record and a Title record. Ahasuerus 23:56, 2 Nov 2006 (CST)

[Add Author] button - the Wiki did not mention this button and I hit it figuring it must be the way to get ISFDB to add new authors...

New Author records are created automatically whenever a submission includes a name that is currently absent in the database. This is both good and bad: good because it's quick and bad because we end up with hundreds of misspelled names, names with missing middle initials, etc. Ideally, there would be a validation step, e.g. "You have submitted a new book by Robert Heinlein. The ISFDB doesn't have this Author name on file. Do you want to: [radio button] use the existing name Robert A. Heinlein or [radio button] create a new Author record for Robert Heinlein?" Not this year, though... Ahasuerus 00:07, 3 Nov 2006 (CST)

Then I got worried that perhaps the book will end up with a second author with a blank name.

The ISFDB software used to have some problems with blank lines, but they have all been resolved and blank lines are simply ignored. Ahasuerus 00:07, 3 Nov 2006 (CST)

Tag - the help page mentions this field but I did not see it and assume it's been merged into ISBN / Catalog #. Would I just delete this from the wiki or ask why I don't see the field and if so, where do I ask? At the moment the publication is in the "waiting to be approved" queue at http://www.isfdb.org/cgi-bin/mod/pv_new.cgi?55194 where I see that there is a "tag" field.

It's a bit more involved, actually. "Tags" served as important unique identifiers of Publication records at one point (back in the dark ages). Now they are less important, but still serve as anchors for magazine records and link Publications to their respective ISFDB Wiki pages. They are automatically generated at submission time, as you have already noticed. Ahasuerus 00:42, 3 Nov 2006 (CST)

Year - The field title is mis-leading and I really had to stare at and think about the help to understand it. I would rename the field "Pub Date."

Thanks, that's one of the blind spots that we have developed over time :) It's another artifact from an earlier era when the database supported publication years, but not dates. Ahasuerus 00:42, 3 Nov 2006 (CST)

I'm really surprised you don't ask for a printing number as it may make people more aware that the "Pub Date" and "Pub #" are related and that you are not asking for the first printing date.

Sort of related is I'm surprised you don't have a First-Edition checkbox so that if a publisher states "First Edition" you can check it. I guess that can go in the notes and would also allow for statements like "A DAW Books Original - Never Before in Paperback."

Some type of "first edition checkbox" has been mentioned a few times over the last year. However, not all first editions state that they are first editions and many ISFDB editors will probably try to guess (there are different algorithms for different publishers), which may result in a mess :( Ahasuerus 00:42, 3 Nov 2006 (CST)

ISBN / Catalog # - My book says UQ1064 on the front cover, 451-UQ1064-095 on the spine, and no other coding. While 99% of book sellers would list this as UQ1064 I chose to go with 451-UQ1064-095 as one of the things I track in my personal DAW list is the code on the spine and there may be others that care of a book was published with "UQ1026-095" vs. "451-UQ1064-095." Again there were no guildelines about which Catalog codes would be regarded as "better." Related to those are those codes where the formatting is ambiguous. For example, a book may be coded "S 309" with a tiny widget of a space between the S and 309. I suspect most people would use "S309" some would use "S-309" and others "S 309". For eyballing your book vs. what's in isfdb all of these will work but it'll matter if someone is searching for a book.

That's a good point, although I am not quite sure what we can do about it. My primary concern is that we are putting ISBNs and catalog numbers (for the pre-ISBN era) in the same field, which makes it hard to validate ISBNs and you no good place to enter catalog numbers from the early 1970s on. Ahasuerus 00:42, 3 Nov 2006 (CST)
It should be easy to check if it seems like someone tried to enter an ISBN. For example, I've got code that scans strings looking for ISBNs. The code’s main loop scans the string looking for a numeric digit, when it finds one it calls _CheckIsbn() and then skips past all consecutive digits. The intent is to find ISBNs embedded within junk and so there is no other parsing other than the scan for digits I just mentioned. _CheckIsbn() expects nine numeric digits (ignoring hyphens) and computes the checksum. If _CheckIsbn() sees something other than a digit or hyphen while it's collecting nine digits it assumes whatever the scanner found is not an ISBN and returns. _CheckIsbn() then skips past any hyphens, compares the computed checksum against whatever is in the string, and then checks to see if the following character the end of the string, white space, or punctuation. If that's the case _CheckIsbn() assumes it has an ISBN. At the moment I accept ISBNs with invalid checksums but will change that soon, or maybe I'll code a special check because many phone numbers have ten digits and are often written as 201-555-1212. Marc Kupper 04:46, 3 Nov 2006 (CST)

Price - My book was priced at 95¢ but the directions did not talk about books priced under $1.00. I entered it as $0.95 as that's what other people seem to have done though would have preferred either 95¢ or if you insist on a leading currency symbol then ¢95 :-).

Well, the de facto standard is to use the dollar sign, but I am not sure if the ISFDB software does anything special with the first character or if it's just force of habit. Al? Ahasuerus 00:42, 3 Nov 2006 (CST)

Artist - While not applicable to my pb I did notice when looking up another book that isfdb had a separate record for a hardcover book's cover. If a separate record for a hc's cover is an isfdb standard practice then it should be mentioned in the wiki.

If a hc edition and its paperback reprint have different cover art, then they should have two different associated "cover art" records, but if it's the same cover art, then it should be the same record as well. I am afraid you will discover that the ISFDB software tends to prefer to err on the side of caution and not merge records automatically (except in certain carefully defined cases), thus leaving many duplicate records in its wake. It's usually up to human editors to follow up and merge the duplicates. Also, Cover art records are generated automatically at Publication entry time and are somewhat messy at the moment. For example, if you delete a Publication, which happens to be the only Publication associated with a particular Cover Art record, should the software auto-delete the latter as well? Ahasuerus 00:42, 3 Nov 2006 (CST)

Image URL -

  • Is there a way to upload scanned images or would I first need to upload them to a web site and then point isfdb at it?
At this point, the ISFDB database consists of textual data and pointers (URLs) to images hosted elsewhere. ImageS present not only man-hour problems, but also potential legal problems, disk space problems and bandwidth problems. Ahasuerus 00:42, 3 Nov 2006 (CST)
  • If I upload to a 3rd party web site does it have to be a permanent URL or will isfdb be taking/caching a copy?
The ISFDB doesn't cache any images at this time, so it would have to be a permanent URL or else we will have a broken link shortly. Ahasuerus 00:42, 3 Nov 2006 (CST)
  • Am I allowed to provide a URL to an image that's not "mine?" For example, a cover might be on Amazon but presumably that's a copyright image.
The ISFDB has thousands and thousands of URLs pointing to Amazon's images, so you would be in good company :) Ahasuerus 00:42, 3 Nov 2006 (CST)
  • What if I scan an image, send it to Amazon, and then point isfdb at it?
Sounds like a plan! Ahasuerus 00:42, 3 Nov 2006 (CST)
  • Are there any file name guidelines? My personal practice has been "Title (Publisher, ISBN or code, printing #).jpg" but another site that I contribute to likes the file names to start out with the ISBN and so for them I put the ISBN or ASIN first.
As long as it doesn't break the URL (we have seen some bizzare URLs that broke our software), any format should be fine. As far as the ISFDB is concerned, it's just a string in the database. Ahasuerus 00:42, 3 Nov 2006 (CST)
  • Last but not least - what are the size guidelines? I normally scan at 300dpi resulting in 300K to 400K byte files that are on the order of 1300x2100 pixels. In the past I have cropped images but am thinking of changing this to always show the edges of a book so that things like artist names don't get cut off.
Since we don't host any image at this time, size is not a consideration. Also, Al resizes some images to fit a certain part of the screen, but he would be in a better position to comment further. Ahasuerus 00:42, 3 Nov 2006 (CST)

Notes - As one of the informal keepers of the DAW book list I would like to start a practice with those publications of putting the printing #s in the notes here using a fairly consistent format of "Printing: 1st, 2nd, 3rd." or maybe "1st, 2nd, 3rd printing" would be better as that's how most book people write it. --Marc Kupper 23:58, 1 Nov 2006 (CST)Marc

Sounds reasonable, but conventions are hard to enforce in free text fields when you have (potentially) hundreds of users :( Ahasuerus 00:42, 3 Nov 2006 (CST)

Idea/suggestion: using the wiki for collecting blurbs

Speaking (over there in ISFDB Feature List) of future uses for the wiki, I was reminded of an old idea of mine. I've long been fascinated by blurbs (defined for my purpose as laudatory quotes on SF book covers by other writers, or sometimes other kind of personalities). I don't know of any website that collects them. Several years ago I toyed with an idea of founding one myself, but gathering a substantial amount of material, let alone putting it online and handling new additions that people would - hopefully - send me, seemed like way too much work. Then I discovered wikies and realized that this would be an ideal format for such a project. I'm not sure whether it would fit into the Wikiquote's scope; and while that would provide a reliable hosting as well as exposure and (possibly) collaboration, there's too much noise from other stuff there. A smaller, (semi)dedicated system would be better; but I have no way of setting it up myself, and the option of trying to start one at Wikia doesn't look that great either.

However, here I come across a wiki with a reasonable guarantee of permanence, associated with a project dedicated to collecting bibliographic data on SF - seems like a match made in heaven. I think it might be a useful symbiosis: blurbs in the wiki would provide a supplement to the content of ISFDB proper (which already lists reviews, so blurbs aren't that far off) with more flexibility than trying to have them within the database (like Notes); on the other hand, as blurbs for any given work (and/or author) are strongly dependent on a particular edition, I've always planned to document their provenance in detail, and being able to lean on ISFDB's Publication data would be good too.

So, what is the opinion of the Powers That Be? --JVjr 11:31, 3 Nov 2006 (CST)

Well, we already have a free text Synopsis area for all Titles, which we sometimes use to enter "Publisher's description" one liners. I can't think of any reason not to add blurbs to the same field unless they contain Evil Spoilers, which they admittedly do on occasion. Ahasuerus 12:18, 3 Nov 2006 (CST)
re “I don't know of any website that collects them.” Dave Wands at www.fantasticfiction.co.uk collects blurbs and shows them on his site. For example, see the “Arthur C Clarke recommends” section of www.fantasticfiction.co.uk/c/arthur-c-clarke/. Each of those is a blurb that also links to the book that Clarke recommended. Marc Kupper 15:57, 10 Nov 2006 (CST)

Question about a typo in a review title

I just entered data for the Oct-Nov 1953 Fantastic Universe, and found that the review of E.E. Smith's "Second Stage Lensman" mis-types the title as "Second Stage Lensmen". I've entered the title as it is in the magazine, but since I believe reviews are matched lexically this will not match the title record for Smith. What's the best fix for this? Also, where should I put the note saying that the title is indeed incorrect, so that someone else does not change it? I've put it on the pub record wiki page for the magazine; I think that's the best place. Mike Christie 07:43, 4 Nov 2006 (CST)

I would enter the review title so that it matches the book's title. The note concerning the title error should be placed in the notes section of the magazine entry. Alvonruff 08:15, 4 Nov 2006 (CST)
Keep in mind, folks, that the software doesn't auto-merge newly entered magazine/collection/anthology/etc contents with pre-existing data. All newly entered stories need to be reviewed and merged manually when appropriate, which is a bit of a pain, but much better than an auto-merge messing things up. I have merged a few stories entered today, but it's something to keep in mind going forward :) Ahasuerus 17:46, 4 Nov 2006 (CST)
Right; I'd forgotten that. I'll check through what I've entered later this weekend. Thanks for the reminder. Mike Christie 18:35, 4 Nov 2006 (CST)

Variant titles question

I noticed that J.T. McIntosh's "The Broken Record" shows up here for early publications as James McGregor (his real name), and also here for its later republication under his usual name. These evidently need to be made variants (not merged, I assume, as that would lose the information about the different author attributions for the different publications). Does it matter which variant is the parent? I would assume J.T. McIntosh should be attached to the parent record, since that's the usual name. That is, I'd display the title record for 93176, and then click on "Make this title a variant title or pseudonymous work", and enter 96845 in the title field. Is that correct? Thanks. Mike Christie 07:35, 5 Nov 2006 (CST)

That's exactly what I have been doing for the last 6 months. The only catch that I have run into is determining which one of the two Author names is the "official" name and which one is the pseudonym, e.g. S._P._Somtow, Megan_Lindholm or Simon_Hawke, who had a non-trivial career under their birth names and then switched to another name and had an even more notable career. Ahasuerus 15:05, 5 Nov 2006 (CST)

Another variant titles question. I just merged several records for Eric Frank Russell's "Boomerang", which was originally titled "A Great Deal of Power". One of the "Great Deal of Power" records pointed to a vt record of 57167, which has now disappeared via a separate merge. If I understand the bug reports correctly, this is a known bug; the vt should have repointed to the merged target. Second issue is that having done this there is no way to delete the reference; is that correct? I have changed it so that the vt is pointing the other way, since "A Great Deal of Power" is the original title. That appears to have worked successfully, but the result is a circular pair of vt pointers. Is there any way to remove the ref to 57167 on 187373? Mike Christie 07:52, 5 Nov 2006 (CST)

The only way to get rid of circular pointers (that I am aware of) is to create a bogus Title/Publication via "New Novel" and then merge the new Title with the Title that needs fixing. When given the standard list of discrepancies between the two Titles, select the correct Title's data, but ignore the pointer. Once the merge goes through, don't forget to delete the bogus Publication record that will be still associated with the merged Title.
These are exactly the kinds of bugs that prevent us from going live tomorrow -- circumventable, but not something that J. Random Editor can figure out on his own, especially if he can't approve his own submissions. Ahasuerus 14:54, 5 Nov 2006 (CST)
OK, I think I fixed this. However, Boomerang shows up twice in Russell's short works list; once in its own right; and once under the main title of which it is a vt. Should it be indexed in the short works list in that way?
Nope, it's a known bug. Variant Titles are displayed twice in Short Works. As an aside, one of our problems is that there are different sections of ISFDB code that are responsible for displaying Long Works, Short Works, Alpha, Chrono, Series data, etc. When Al makes adjustments and corrections to one section of the code, they don't get automatically propagated to the other sections, hence the discrepancies. Ahasuerus 18:14, 5 Nov 2006 (CST)
Also, it's odd that if you click on "A Great Deal of Power" you get six pubs, whereas either of the "Boomerang" links (they are the same link, in fact) only show the five pubs with the "Boomerang" title. Is this intentional? I think it's OK this way, just wanted to be sure it was deliberate. Mike Christie 16:33, 5 Nov 2006 (CST)
Actually, that's intentional. The parent title shows all publications for itself and all its child titles. Child titles only show their own publications. Ahasuerus 18:14, 5 Nov 2006 (CST)

Leading Articles

Another question: I've seen several cases where the word "The" is the only difference in a variant title. For example, "The Life Watch", by Lester del Rey, shows up under that title in the September 1954 Fantastic Universe; it was reprinted in "Gods and Golems" as "Life Watch". Are these true vts? I've been keeping the earlier form of the title and simply merging rather than creating variants in these cases. Any reason to make them vts instead? Mike Christie 08:09, 6 Nov 2006 (CST)

This is a tricky area. Most of the time, the leading article is something that editors add or delete at will, e.g. see the history of David Gemmel's publications in the UK vs. the US. On the other hand, articles (or lack thereof) can also be indicative of substantive differences, e.g. see Orscon Scott Card's Maps in the Mirror vs. Maps in a Mirror, although the article in question is not a leading one. Other times, authors add or substract leading articles when expanding short works into longer ones. Finally, Amazon.com has been known to add or drop articles (or otherwise butcher titles) at will, which can lead to phantom differences between publications.
Personally, I am inclined to err on the side of caution and create variant titles whenever there is a confirmed difference between two publication titles, although it's admittedly a pain. Ahasuerus 10:34, 6 Nov 2006 (CST)

Fillos

How minor does interior art need to be before we ignore it? Fillos do not illustrate a story, and are often uncredited. I'm inclined to ignore art unless it meets at least one of the following criteria:

  1. credited to artist
  2. illustrates a story
  3. large (half page for digest)

There are some fillos in the May 54 Fantastic Universe, for example; one (p. 105) is about a third of a page, and is a skiffy picture of a rocket port. No credit, nothing to do with the story on that page. It's definitely interior art but it's hard to enter into the database with so little info or relevance. I think this is borderline but does not go in. On the other hand, Fantastic Universe had several squib ads for sister magazines that were preceded by little scribbles, of the Saint logo, or other very simple drawings. These are clearly just fillos and don't need to be entered. Mike Christie 08:05, 5 Nov 2006 (CST)

I think the first two criteria you list would be the primary ones. In modern magazines, the interior art of Analog almost always supports a story, is found on the title page for the story, and has a credited artist. On the opposite side is Asimov's, which hasn't really done interior art for a few years now, but will occasionally use canned art around poetry and sometimes as... well, filler.
There are novels that contain a small number of illustations (Stephen King's Gunslinger comes to mind), but we so far haven't been tracking each individual illustration in that case - a single note in the publication describing who did the interior art suffices. As such, I'd say:
  1. An interior art work must be associated with a particular story before we find it interesting.
  2. An artist credit is not a prerequisite (although obviously more useful). In fact, a listing of unknown art works across a particular magazine could serve as a todo list for someone to create some research around.
  3. Size is probably a good indicator of whether or not something is really a pen and ink doodle, but what we're really interested in is whether it's original artwork of interest to the field. If the artwork is by an artist like Vincent Di Fate or Virgil Finlay, then I'm probably going to enter it - even if it is only a doodle. Alvonruff 21:28, 6 Nov 2006 (CST)
Let me just add that this is one of the reasons why I have my doubts about an earlier proposal to disable editing once the Verification Flag has been set across the board for a Publication. On the other hand, perhaps we could disable changes to the core fields like Title, Name, Year, etc? Even if we discover that the Publication was published pseudonymously later on, it would only affect the associated Title record(s), not the Publication record. Ahasuerus 18:19, 8 Nov 2006 (CST)

Accented characters in authors' names

What's the right approach to names such as Philip José Farmer? I have been using the accented form of the e, but I see that's causing problems -- see this page and click on the Farmer link. Is there support for special characters? Mike Christie 10:34, 5 Nov 2006 (CST)

The ISFDB software does support Unicode -- most of the time. Unfortunately, there are still some issues with Unicode support, which mess up non-ASCII characters and even apostrophes (since Unicode has two different apostrophe characters). I would simply add this incident to the Open Editing Bug page and hope that Al gets to it in late November when he concentrates on bug exterminations. Ahasuerus 14:59, 5 Nov 2006 (CST)
OK, I'm confused. I decided that pending any other information I would change all the occurrences I had created of "Philip José Farmer" to "Philip Jose Farmer" so that things would work, figuring I'd fix them later if needed. I discovered that "Edit this Pub" would show a change for "José" to "Jose", but on approving the change nothing happened -- the accented version was still there. So I tried changing it to "Philip Farmer", approving that, and then changing it to "Philip Jose Farmer", thinking the double change would get around whatever it was that was ignoring my changes. The first change worked, but the second change restored the accent! So I'm giving up for now. I will enter other issues without the accent until I hear otherwise. Mike Christie 17:53, 5 Nov 2006 (CST)
While Joel Spolsky likes maintain that unicode is easy (and it is for integrated vertical applications), it's a bit more difficult in our case, as it depends on:
  • The user's browser settings.
  • Unicode support in the the http server (which passes strings to applications)
  • Unicode support in Python (which gets strings from the server)
  • Unicode support in MySQLdb (which maps python strings to MySQL strings)
  • Unicode support in MySQL (which stores and searches for strings)
In changing "José" to "Jose", I would argue that it actually does the right thing - It prevents someone from forking a new author name, when the correct one is in the database. Where things went wrong was how you got "Jose" to get entered in the first place.Alvonruff 21:39, 6 Nov 2006 (CST)
OK. In some of the magazines I am now entering, there are stories by Dorothy Madlé. I have entered them as by Dorothy Madle. If I understand your comment correctly, I really should have used the accent. To avoid forking, is the correct sequence to edit the author data to change the name to include the accent, then enter subsequent stories with the accent in her name? Or are there other places I need to edit as well? Mike Christie 07:45, 8 Nov 2006 (CST)

Problem with bad entry in Fantastic Universe

This is mostly a note to Al to request a cleanup behind the scenes, but others might want to know about the mistake I made and what happens if you make that mistake. I pasted in an old tag for the April 55 Fantastic Universe, and apparently didn't edit it. Hence the April and March 1955 Fantastic Universes both have the same tag: "FANTUNIVMAR1955". As a result the April pub record can't be accessed. I guess this is a bug, in that the system should probably not permit the entry of duplicate tags, though of course it's really my error.

Al, can this be easily salvaged with some updates? Or should I just delete both publications and reenter them? -- Mike Christie 07:39, 8 Nov 2006 (CST)

OK, fixed it myself via a sequence of tag edits. Not a problem anymore. Mike Christie 13:40, 9 Nov 2006 (CST)
Keep in mind, folks, that Al is in Oulu, Finland at the moment -- see his schedule at User:Alvonruff :-) Ahasuerus 17:27, 9 Nov 2006 (CST)

Fiction versus essay

I've come across a slightly odd situation in Fantastic Universe. There are one page pieces inside the front cover of most issues, almost all by Frank Belknap Long. The early ones are simply nonfiction descriptions of the inspiration for the cover art. However, later on, they start to become word-portraits of the fictional situation being depicted; the text reads as if it came from a story. I'm going to treat these as shortfiction rather than essay types, though I think they're borderline. Mike Christie 17:56, 8 Nov 2006 (CST)

Now that I am thinking about it, there were quite a few Startling Stories editorials written by "Sergeant Saturn" (aka O. J. Friend in his YA-friendly mode) that read almost like fiction. Hm... It may be difficult to come up with clearcut guidelines. Then again, some purported poetry reads like prose these days, so there may be other areas of overlap. Ahasuerus 18:13, 8 Nov 2006 (CST)

"Note" vs. "Bibliographic Comments" in the Wiki

Let's see if we can come up with some guidelines re: the use of these two fields. Would it be fair to say that permanent comments that reflect something indisputable about a Publication, e.g. this one, belong in the "Note" field in the ISFDB database proper? And that temporary comments that support the bibliographic effort but will be eventually deleted (once all issues are resolved), e.g. this one, belong in the "Bibliographic Comments" field in the Wiki? Ahasuerus 18:53, 8 Nov 2006 (CST)

Agreed. Actually I need to clean up the Aug/Sep 53 notes; I completed those fixes. I'll also remove the biblio note version of the Jun 55 one, and modify the editing help to reflect your defs. Mike Christie 19:10, 8 Nov 2006 (CST)
I updated the template help; see what you think. One additional kind of data that might go into the Bibliographic Comments field is information about conflicts amongst the sources. This would be permanently recorded, but I think does not belong in the notes field on the pub record. Mike Christie 19:20, 8 Nov 2006 (CST)
Sounds reasonable since the numer of potentially conflicting secondary sources is unpredicatble and I see no reason to add a second level discussion to the database proper. Ahasuerus 00:38, 9 Nov 2006 (CST)

Cataloging non-genre publications

One thing that I have been struggling with is the rather high number of SF stories originally published in various obscure and/or non-genre places like newspapers, fanzines, foreign language publications, etc. For example, here is where some of Forry Ackerman's stories collected in Science Fiction Worlds of Forrest J Ackerman & Friends first appeared (http://contento.best.vwh.net/t1.htm#A1):

  • 40 • The Atomic Monument [revised from “Memorial” Astounding Apr ’46] • Theodore Sturgeon & Forrest J Ackerman • vi Fantasticonglomeration
  • 43 • Confessions of a Science Fiction Addict • ar After Hours v1 #4 ’57
  • 61 • Burn Witch, Burn! • ss Sex and Censorship, 1958
  • 114 • Tarzan and the Golden Loin • ss V, 1948
  • 122 • Count Down to Doom • Forrest J Ackerman & Charles Nuetzel • vi Famous Monsters of Filmland, 1966
  • 135 • The Lady Takes a Powder • Tigrina & Forrest J Ackerman • ss Inside, 1953
  • 169 • The Record • Forrest J Ackerman & Ray Bradbury • ss Futuria Fantasia Sum ’39
  • 213 • The Radclyffe Effect [revised from “Mundo de Soledad” Los Cuentos Fantasticos July 8 ’48] • ss, 1969

Or check out the stories in Bruce Boston's She Comes When You’re Leaving & Other Stories (http://www.locusmag.com/index/t54.html#A6233):

  • 7 • Broken Portraiture • ss Gallimaufry #5 ’75
  • 13 • Interview with a Gentleman Farmer • ss City Miner #1 ’76
  • 17 • The Monster and the Moon • ss Fiction Magazine #4 ’73
  • 22 • Limb Still Kicking from a Stillborn Novel • ss Gallimaufry #7 ’76
  • 27 • Sunday Review • vi City Miner #3 ’76
  • 30 • The Poets’ War • ss City Miner #7 ’77
  • 37 • Doctor’s Dozen • ss Portland Review, 1981
  • 44 • She Comes When You’re Leaving • ss Berkeley Poets’ Co-op #9 ’75

And if you peruse a reasonably complete L. Sprague de Camp bibliography, you will see hundreds of publications in all kinds of unexpected places.

I doubt that we want to include a complete or even partial bibliography of every anthology, magazine and newspaper that every SF writer has ever written for, but the only alternative that I have been able to come up with was simply recording information about previous publications (when available) while entering data for bona fide collections, anthologies, etc. And if the story has never been reprinted in a genre publication or a single author collection, then we will simply have a Title with no Publications pointing to it and a Note field with its publication history. Any other ideas?

And then there is the (even more frustrating) case of periodicals like The Blue Book Magazine, which published a respectable amount of SF over the years, but it never added up to to more than a fraction of the total page count... Ahasuerus

Agreed; these are problems that we can't deal with right now. I think they're also fairly low on the priority list, though if we are going to become an authoritative bibliographic source they will eventually have to be dealt with. Your suggestion that we use the notes field seems the best one for now. Mike Christie 07:29, 9 Nov 2006 (CST)
One problem that I see with the proposed approach is that our verification scheme is Publication oriented. If we don't have a Publication record in the database, we can't mark it as Verified even if/when you have the Publication in question in your hands. That's not a very attractive proposition, but the alternative appears to be the creation of partial bibliographic records for any Publication (anthology, magazine, newspaper, fanzine, etc) that has an SF-related piece in it. Is that sustainable in the long run? Ahasuerus 10:31, 9 Nov 2006 (CST)
No, it's not. A couple of options occur to me. For a publication like Blue Book, we might choose to only enter those content records that are relevant to the ISFDB. Then the publication notes would record that the contents are incomplete, and why; verification could then proceed. Alternatively we could enter everything, and have a way to mark whether a title is of interest or merely "filler", in the ISFDB's eyes. That seems rather timeconsuming, but would jibe with extensions or replications of the ISFDB to cover other genres -- not something we should plan on. So the first option seems more likely to me. Mike Christie 13:31, 9 Nov 2006 (CST)
Well, as the current ISFDB:Policy page states, the original intent was to avoid incomplete Publication bibliographies:
In - Otherwise unrelated works found in any publication that will be cataloged based on other criteria. This is done to avoid creating incomplete biblio records for magazines, anthologies, etc.
...which is a laudable goal, but what happens when you have a single supernatural story in one of those monster Norton anthologies of 19th century mimetic fiction? Do we list the non-SF stories as well (dozens of them) and have the software create Author records for every 19th century author? The more I think about it, the less attractive it sounds. Ahasuerus 10:55, 10 Nov 2006 (CST)
The more I think about it, the more convinced I become that the current policy quoted about is not sustainable. There are just too many Publications with random SF/F stories or related non-fiction -- from Playboy to The New York Times -- to make it viable. And since we need to have a Publication record if we want to be able to use the Verification flag, I guess there is little choice but go with partial bibliographic records. Which saddens the completist in me, but oh well.
If we all agree with the proposed approach, then the next question is how do we support partial Publications? Do we add a new "partial" checkmark to the form or do we rely on our editors to add appropriate free text comments? Ahasuerus 23:17, 11 Nov 2006 (CST)

Prioritizing Open Bugs

Welcome back, Al! Mid-November is fast approaching, so I guess we should start massaging the Open Bugs lists and prioritizing them. I'll see what I can do over the weekend. Ahasuerus 10:55, 10 Nov 2006 (CST)

I will also try to take a pass at this. I think reviewing all the help files for completeness would be good too, and I'll try to do that as well. Mike Christie 12:16, 10 Nov 2006 (CST)

Common titles such as "Introduction"

I noticed that Hans Stefan Santesson has introductions listed with titles that include the name of the publication they are from. This is out of step with the title help, which I think I wrote all or part of. Which is wrong? I think these should just be titled "Introduction". Mike Christie 12:16, 10 Nov 2006 (CST)

On the one hand, we strive to make our data as accurate as possible, which would suggest that in this case we would want to make the title "Introduction". On the other hand, a prolific editor may be responsible for dozens and dozens of Essays titled "Introduction", which would make using their biblios difficult. Neither approach sounds particularly attractive.
I wonder if the problem can be solved via software. Suppose we have a list of known "ambiguous titles", i.e. "Introduction", "Foreword", "Afterword", etc. The software that generates ISFDB HTML pages could then check whether a particular title is on this list and display the associated Publication title (the first Pub title if there is more than one?) in square brackets. Ahasuerus 12:33, 10 Nov 2006 (CST)
Sounds like a smart fix; not for this first release, though, right? Mike Christie 12:41, 10 Nov 2006 (CST)
Interesting. There are loads of other column titles from magazines (in fact I used to have a tool to go and find them) which would also enjoy such an enhancement. I'll look into it. Alvonruff 12:46, 10 Nov 2006 (CST)

Author names and punctuation

I was cleaning up Sam Merwin's biblio the other night and noticed that we have two Author records for him:

  • Sam Merwin Jr.
  • Sam Merwin, Jr.

Running a slightly modified version of the Perl script at ISFDB:Author Names Cleanup, I find a number of duplicate Author records where the only difference between the two records is that the suffix is spelled "Jr." in one case and "Jr" in the other case or there is an extra comma between the name and the suffix as per the example above. I suppose some of the discrepancies may be attributable to data entry problems, but in other cases they may reflect a bona fide difference in publishers' standards.

I wonder if the difference between these two forms of an Author's name is significant enough to justify two separate Author records in the database? Ahasuerus 13:04, 10 Nov 2006 (CST)

One thing I would like to do in the help text is to make these distinctions crystal clear. Here are some examples of what I think the help text should say; I'll use my own biases as to the answers to these questions, so that I have content to quote, but it's the level of detail that I feel is important here.
  1. Name suffixes: Al's help already specifies that these are preceded by a comma. I think we can add the period at the end as required, for abbreviations. Names differing by case and punctuation in these respects are translated to canonical form.
  2. Titles: capitalization should be standardized: "and", "the" and a list of other words gets no initial cap; everything else does. Titles listed in other forms are translated to canonical form; if the punctuation is significant to the title (e.g. Don Marquis) then exceptions can be made and should be noted in the pub notes page.
  3. Fonts: special fonts (e.g. Davidson's "Selectra Six-Ten") are not of interest, though they are relevant to the story.
  4. Characters such as em-dashes are entered in Unicode; a list should be given.
  5. Title changes such as the addition of a hyphen, or a space ("Hellfire" vs. "Hell Fire") or a leading "The", are significant and should be recorded as variants.
And so on. I know Bibliographic Rules holds a lot of this sort of stuff; I think the help texts are a good final repository for a lot of these decisions. Mike Christie 13:21, 10 Nov 2006 (CST)
Sure, help pages need to be as authoritative as we can make them. Also, there is no reason to have more than one authoritative page since duplication can easily lead to bifurcation and confusion. Thus it stands to reason that our help pages should be the final repository of whatever bibliographic wisdom we end up with here. Ahasuerus 23:46, 11 Nov 2006 (CST)
Semi-related to this is just now I was doing a scan of the DAW list for all upper-case titles and converting them to standard title casing if appropriate. To ensure reliability I was doing this by doing a case-insensitive scan for the title in my personal list and then copy/pasting the title to the DAW list. This was working fine until I hit “TRAITOR’S SUN” and the scan did not find it in my list. I shrugged, looked at the book on Amazon, and did another search on the lists for Amazon’s title (“Traitor's Sun”) and found it in my list but not DAW’s. I stare for a moment before realizing that character 39 (apostrophe or -'-) is different than character 146 (right single quote or -’-). User:Marc Kupper
Oh yes, apostophes are evil and the ISFDB software still has trouble with them, e.g. Short Works biblios are messed up if the Author's name has embedded apostrophes. Also, the two different types of apostrophes create problems in Series names. Ahasuerus 22:59, 11 Nov 2006 (CST)

Bug tracking suggestion

I was just looking through the bugs list to see which ones I thought needed to be fixed prior to going live with a beta, and I realized that the lack of IDs for the bugs makes it hard to discuss them. Short of implementing something like Bugzilla, a simple solution might be to implement a "hard" enumeration system. E.g. all bugs have five digit ids; Editing Bugs start at 10001, Display Bugs at 20001 and so on. Bugs are recorded like so:

10001 OPEN No review option on default content pulldown. [followed by all the usual text]

A fixed bug just has the status changed to "FIXED", so that the bug number is still visible. New bugs continue to be added at the top. Then we can have concise discussions about which bugs need to be fixed.

Is there a better quick way to deal with this?

Unless Al can pull a pre-cooked WikiMedia-provided rabbit out of his bag of tricks, I don't think we can do better than use the proposed "hard numbering" approach. Ahasuerus 23:52, 11 Nov 2006 (CST)

Separately there is the question of what needs to be done to the Wiki to get it ready for beta. As I said earlier, I think the help text needs upgrading and I'll try to work on that this weekend. Mike Christie 21:13, 10 Nov 2006 (CST)

Once the help pages are reasonably stable and have been cross-reviewed, I don't think there is much else that needs to be done within the Wiki prior to go-live. One "nice-to-have" feature would be to automatically add a link back to the related ISFDB page whenever you start editing a blank page in the Wiki, but that's just gravy. Ahasuerus 23:52, 11 Nov 2006 (CST)

Help text

I have done some expansion of the help text for publications, focusing on the title and author help fields. Comments and improvements/expansions welcome. Take a look at Editing:Adding_a_New_Publication for the new version. Mike Christie 01:23, 11 Nov 2006 (CST)

The first thing that comes to mind is that the following paragraph:
When a book has a section with Roman numeral page numbers for introductory material, followed by Arabic numerals for the main text of the book, the introductory material is ignored for page numbering purposes.
appears to diverge from the standard bibliographic format, i.e. "viii+233". Not that I am a huge fan of Romans ;) but do we really want to discriminate against them? Ahasuerus 10:34, 11 Nov 2006 (CST)
I think we're okay with Roman numerals - I even have support to print roman numeral page numbers in the correct sorted order. I've also been using the "viii+233" format for a book's page count. Alvonruff 10:57, 11 Nov 2006 (CST)
Oops, you're right. Fixed in the text. Mike Christie 12:06, 11 Nov 2006 (CST)

Ace Doubles

I suggest we assert that Ace Doubles are omnibuses. The ones I've seen are entered as novels, which seems quite wrong; the only alternative is anthologies, and that seems wrong too. We could create a new type, DOS, per the Nicholls/Clute coinage, but I think there's no need. Any disagreement? I haven't made this change to the relevant help template, but I would like to do so if no one objects. Mike Christie 10:10, 11 Nov 2006 (CST)

I think I'm okay with that, but we need to decide what to do about title records. For a normal omnibus, we have a title and publication record for that omnibus, such that the omnibus shows up in the author's summary bibliography. In this case, if we just change the pub record from NOVEL to OMNIBUS, we won't see the Ace Double specifically listed in the summary bibliography. I don't have a problem with that, but we should definitively decide which way it should work. Alvonruff 11:04, 11 Nov 2006 (CST)
I think this is OK. The result could be an Ace Doubles page that looks rather similar to the magazines directory pages, from which an interested browser could see all the Ace Doubles.
I don't plan to go around changing all the data any time soon, but I will go ahead and change the help text to say this. What about author data, though? If it's an omnibus, typically one enters nothing, or "Anonymous", right? Mike Christie 12:02, 11 Nov 2006 (CST)
I believe most ISFB omnibus records list the Authors that are displayed on the title page. If the omnibus reprints novels by a single author, then it will be just that author's name, but if it includs works by multiple authors (which is common in the world of futuristic romances), then the Publication and associated Title list all names that are given on the title page. See Christine_Feehan's bibliography for various examples.
As an aside, these "futuristics" (as "futuristic romances" are known to romance fans) can be tricky since many "novels" are borderline novellas and vice versa. Of course, the same was true in the days of Ace Doubles, Belmont Doubles, etc -- there is nothing new under the sun -- but it tends to be worse with futuristics since they commonly publish 3-4 short novels by 1-4 authors per volume.
One rule of thumb that has helped me distinguish between omnibuses and anthologies is whether the book lists the editor's name on the title page. It breaks down when you have a particularly thick reprint anthology which includes one or more novels, but it seems to work 95% of the time. Ahasuerus 22:47, 11 Nov 2006 (CST)

More thoughts on this. I'm not sufficiently familiar with the ISFDB to be clear on the ramifications, but I found in writing up the help text that I wasn't clear how an anthology or collection within an Ace Double would work. For example, take D-205, which is Wollheim's anthology "The Earth in Peril", backed with Lan Wright's "Who Speaks of Conquest?" I first thought that the natural way to do this is to have the publication be "The Earth in Peril/Who Speaks of Conquest?", type OMNIBUS, author blank; and have the novel as a content element, and each story as a content element. The problem with this is that there is nowhere for the anthology title to go: "The Earth in Peril". I don't believe that particular anthology was ever republished, but if it were, this Ace Double should show up as a publication of that title, and under this approach it would not.

Yup, this is a tricky one. Logically, we have an omnibus Publication whose contents happen to be an anthology Title (with an Author/Editor record associated with it) and a novel Title (with its own Author). Then the anthology has contents items of its own, namely a bunch of stories. I think the software currently supports this approach, but I am not sure if that's what we want. There are similar complications with reprint anthologies that consist of two+ previously published anthologies. Ahasuerus 00:00, 12 Nov 2006 (CST)

Currently this appears to be entered as two separate publication records. I think this might be the best workaround for now. Ideally the physical volume should correspond to a single publication record; that's the intended meaning of "publication" after all. But I don't see a better way to handle the anthologies. Mike Christie 19:39, 11 Nov 2006 (CST)

Minor content items

Another borderline case came up while I was verifying the October 1956 Fantastic Universe. Most issues of this magazine had a one page "Story Behind the Cover . . .", written by Frank Belknap Long, on the inside front cover. This issue has an advertisement for the Worldcon there instead, and a note at the bottom says "Turn to page 93 for Hannes Bok's own Story Behind the Cover!" On page 93, about a third of the page is given to Bok, who has written a fictionalized account of the cover, as if it were a story. I have been entering all of Long's "Story Behind the Cover . . ." entries, and making them "SHORTFICTION" if they were completely fictional without any narrative paragraph framing the fiction. This just seems too short to me to enter as SHORTFICTION, but I feel I'm being inconsistent.

Similarly, I realized as I was writing up the content help that there are going to be many items that aren't easy to classify. I think we do want to record the existence of letter columns, "In Times to Come" columns, AnLab, and so on. These can be very short. Sometimes there is just a sentence of filler at the bottom of a page: "Don't miss Arthur C. Clarke's great new story 'The Pacifist' in next month's issue of 'Fantastic Universe'!" I don't think that deserves an entry.

I will try to formulate some rules and put them at the top of the content help page. I think an entry in the table of contents always warrants a content entry. After that, I think it's going to be a judgement call -- is this something you would regard as a separate item in the magazine, or is it just a line of filler or advertising that would not exist if there was not a space to fill on that page? Pure advertising -- subscription calls, back issue ads, and so on -- are never entered. I'm sure there will be more borderline cases I haven't thought of yet. Mike Christie 07:32, 12 Nov 2006 (CST)

I've now put some rules at the top of the contents help page. Take a look and see if anything needs changing. Mike Christie 10:39, 12 Nov 2006 (CST)
The rules up top look reasonable, but the following paragraph:
Use the lower case form of Roman numerals, for pages in introductory material. Sometimes there are pages without a given number in between the last numbered page with a Roman numeral, and the first numbered page with an Arabic numeral. In these cases, you can count backwards to determine the "page 1", and count forwards to determine prior page numbers in terms of the Roman enumeration.
...confuses me, which makes it likely that at least some other editors with less ISFDB experience will be confused as well. A clarification, perhaps? Ahasuerus 21:35, 12 Nov 2006 (CST)
I removed the text; it was confusing, as you say, and on reflection I think it's unlikely to be necessary. I was trying to specify what to do for those pages in between the last Roman numeral page and the first page with Arabic numerals. Mike Christie 05:32, 14 Nov 2006 (CST)

Cover art

Now that we have "Cover art" as a content type, do we need the cover artist field in the pub record? I have just realized that I haven't been entering COVERART, so I need to go back and do that for the Fantastic Universe records. I think it's harmless entering both, but I don't see a reason we need them. Mike Christie 11:18, 12 Nov 2006 (CST)

Hmm. Looks like I've misunderstood something. I went and looked at some COVERART records via advanced search, and found e.g. [this record]. The parent title doesn't appear to have any content records, though, so is this record coming solely from the artist field on the publication? Mike Christie 15:07, 12 Nov 2006 (CST)
I think Al has some special magic brewing for Cover Art records, but I'll defer to him in case my understanding of the Cosmic All as it pertains to Cover Art is incomplete. Ahasuerus 21:38, 12 Nov 2006 (CST)
Theoretically, the two different methods you describe should have identical results. Entering an author in the Artist field of the publication will result in the generation of a coverart title record. Entering a coverart title record as part of the publication will also do the same thing. When editing a publication, the coverart record will show up as part of the publication metadata, and not as content with the titles. I think the only outstanding problem we have in the ISFDB with cover art at present is with back covers. Alvonruff 10:48, 22 Nov 2006 (CST)

Non page number locations

I don't think we need to do this right away, but I'd like to suggest that we establish conventions for the page number field to provide information in that field for everything entered, regardless of whether there is a page number or not. For example, I'd suggest we use the following abbreviations:

  • fc -- front cover
  • fep -- front end paper, or inside front cover of a magazine
  • bp -- unpaginated pages that precede pagination
  • ep -- unpaginated pages that follow pagination (although generally we would expect people to count forward to find a page number)
  • bc -- back cover
  • bep -- back end paper
  • ## (actual count) -- actual handcounted pages, not counting the front cover or front endpaper, for a book without page numbers
  • <text> -- descriptive of the location in some other way. E.g. "Inset artwork on poster inserted with this magazine".

Sorting is an issue, but not an urgent one. Any comments? Mike Christie 14:27, 12 Nov 2006 (CST)

I think the biggest issue will be the "actual handcounted pages" proposal. It is nice to have, but handcounting pages can be tricky and the distinction between page numbers and "actual handcounted pages" may also confuse some editors. Still, it would be a plus if we could document the rules clearly enough.
As an aside, I am a little concerned that having very detailed Help pages may turn off casual editors who would be disinclined to read about all the details and special cases. Perhaps we could have a synopsis/Reader's Digest version at the top of the page and then the details further down, the latter to be consulted only if necessary? Ahasuerus 21:45, 12 Nov 2006 (CST)
I was thinking along the same lines; a "How to Get Started" page, perhaps? Straightforward entry of a novel would be a good way to get started. Another option would be to split the help pages up by field and have help links for each field; each help text could then have the summary text at the top, and the details below for those with questions. Mike Christie 21:50, 12 Nov 2006 (CST)

Canadian Editions in the DAW List

In Miscellaneous DAW List Corrections I had written

  • A row existed for UE1951 at $2.25. It turns out I have two copies of this publication, a 1st printing at $2.25, and another 1st printing (!) at $2.50. I added a new row to handle this. Marc Kupper 03:13, 12 Nov 2006 (CST)

Don Erikson queried me off list about this asking if I had compared the ads in the back to help date these publications. I had not but to help compare the pages I took photos of a bunch of pages and threw them up side by side. The ads were identical but I did spot a difference on the copyright page I had not noticed before in that at the very bottom of the $2.25 copy has “Printed in U.S.A.” and the $2.50 copy has “Printed in Canada / Cover Printed in U.S.A.”. I had heard of Canadian editions when looking at DAW book listings on the Internet but had not seen one. This complicates the DAW List a little bit and I’d like to solicit some ideas on how to deal with it. One thought is that in the Verified column instead of “p” that people could use “c” for the Canadian printings and we’d color code them. A minor hassle with that is I’m hoping to do a future version of the DAW List that’s driven off isfdb’s database and presumably the Canadian printings would need to be flagged in the Notes field. Marc Kupper 19:06, 12 Nov 2006 (CST)

I think ideally you would have a field on the publications record indicating country. I've seen it done this way on other bibliographic databases. Mike Christie 19:29, 12 Nov 2006 (CST)
Canadian editions and Canada-related artifacts present at least two separate challenges. First, there are pesky Canadian editions that only an expert can tell from their American counterparts. I am thinking about one notorious Thorne Smith case in particular. Second, some books have not just two prices, but two distinct ISBNs listed, e.g. Glen Cook's The Dragon Never Sleeps. Since we have only one field for ISBNs, there is no place for the Canadian ISBN to go to, which can be a problem if somebody tries to do a lookup based on that ISBN.
I decided to take a step back, reviewed Concept of Operations and believe that implementing logic to deal with the country would be expanding isfdb’s scope. The existing publication notes field, while not perfect, seems like a good place to document details such as that the book was printed in Canada and how differs from the U.S.A. version. As it is, I’ll add a low priority note to myself to re-verify my DAW books to flag those that are printed in Canada. While I knew that DAW printed some books in Canada until today, I had no idea that those books could be identical to the U.S.A. editions and that they would have their own printing # sequencing.
It’s a low tech solution but I decided for now to deal with the DAW List by noting the country in the Printing field and to not color code the rows or to add any other special logic. For the book that started this discussion the rows now say “1st U.S.A.” and “1st Canada”. I’m happy with how that looks and will also update the description of the Printing field to note that Canadian printings exist and what to look for when verifying a DAW book. Marc Kupper 23:05, 12 Nov 2006 (CST)
In the first case, i.e. a different (albeith barely distinguishable) Canadian printing, we could, theoretically, enter two Publication records and a free text Note explaining the difference between them. In the second case, I am not quite sure what could be done aside from making the ISBN field a repeating field a la the "Web Site" field, which doesn't sound very attractive. Ahasuerus 21:24, 12 Nov 2006 (CST)
Again, it may be expanding isfdb’s scope but my own list I made the ISBN field repeating and it ended up being pretty attractive as I also have logic that allows for explanatory comments in parentheses. Here are some examples of ISBNs from my list. “0-812-53076-4; 0-812-53077-2 (Canada)”, “0-523-48572-7; SBN 523-48572-7-295 (on spine)”, and “0-425-03717-7 (on copyright page); 0-425-03052-0 (on spine).” I’ve done this quite a bit for books that provide conflicting information or where the information was in an unusual spot and have found it to be pretty handy to have the note right there next to the data rather than off in my comments/notes field. Marc Kupper 23:05, 12 Nov 2006 (CST)

Dealing with attached excerpts and other questions

I just compared the pub listing (http://www.isfdb.org/cgi-bin/pl.cgi?GHSTZDLJCG2006) for John Ringo's "Ghost" with the Baen pb release. I've updated the pub data to match the details of the actual release. The existing pub record's date was too early by a month

This is a tricky area. Keep in mind that most paperback publishers follow the rules established by magazine publishers decades ago. Thus a "December" publication will most likely hit the stores in November and a "January 2007" publication will be in the stores long before Christmas 2006. A mess, isn't it? We definitely need to make a decision wrt to this thorny problem one way or the other, though. Ahasuerus 22:06, 12 Nov 2006 (CST)
I’d personally vote for going by what is stated in the books. I used to go by what Amazon listed and found that often Amazon would have a date somewhere in the month preceding the date in the book or would show a reissue date while I would have an earlier printing with the same ISBN. Marc Kupper 23:22, 12 Nov 2006 (CST)
One thing to keep in mind is that on rare occasions a Publication would have, say, 1991 printed as the time of publication, but in reality it would't be released until 1994 or even 1996, i.e. 3-5 years later, due to various publication delays. Granted, this is mostly a problem with small presses and such, but it's something to consider. Ahasuerus 23:41, 12 Nov 2006 (CST)

and the page count was too high by 12 pages. However, the book itself claims to be the "First Baen paperback printing" and the dates are too close for this to be a second printing with an unchanged title page. I'm not sure of the souce for the original pub entry but if someone can explain how to determine to origin of a pub record, I can approve this update or reject it and reenter it as a second pub. PortForlorn 21:43, 12 Nov 2006 (CST)

The last 3+ years worth of ISFDB data was mostly entered by Dissembler, Al's faithful webbot. Dissembler gets its data from Amazon.com and similar places, which are notoriously dirty: lots of data entry errors, delayed publication problems and, very frequently, bogus page counts. AFAICT, the main reason for bogus page counts is that publishers send the data to Amazon et al before they know the actual page count and Amazon doesn't update its database when it becomes known.
As a general rule, it is very unlikely that two separate mass market paperback editions will appear almost simultaneously. Hardcovers are a different story since special "library binding" printings are commonly issued within a few weeks or months of the first hardcover edition. They usually have a separate ISBN and often a different price, which helps distinguish them from first editions. Ahasuerus 22:06, 12 Nov 2006 (CST)

The more interesting feature of this pub is that the publisher has appended an excerpt of another novel. The excerpt is only twelve pages and the novel won't be released until December. I've included the excerpt as another title in the contents for this pub record with the format "new title excerpt" since I can't seem to find any guidance on the handling of attached excerpts and there is no story type for "excerpt". I've seen this happening recently and expect we should have some established guidance. I've left the original pub record in place and the update in the moderator queue. I'm open for any advice. PortForlorn 21:43, 12 Nov 2006 (CST)

I thought I'd written some draft notes on this in the help text, but now I can't find them. Anyway, I looked at your submission and it looks pretty good to me. I might suggest that we make the format "(excerpt)" at the end, rather than "excerpt", just to make it more clearly separate from the title. I would think in some cases the excerpt will have a title that makes it clear what is going on, so the addition of "(excerpt)" won't be necessary -- e.g. the section might be titled "Chapter One of Joe Author's Forthcoming Book, Fantasy Blockbuster". I think we capture actual title on the page in those circumstances. Mike Christie 21:56, 12 Nov 2006 (CST)
The de facto convention has been to add " (excerpt)" to the title and file it as "shortfiction". I don't recall seeing this convention documented anywhere, though. Ahasuerus 22:06, 12 Nov 2006 (CST)

Non-genre for contemporary adventure novels?

The title in the above pub (http://www.isfdb.org/cgi-bin/pl.cgi?GHSTZDLJCG2006) is actually a contemporary adventure novel. While I could see classifying it as speculative fiction, I could also call it non-genre. Are there any criteria for discriminating between contemporary sf and contemporary adventure titles? Since the author has already written a bunch of sf, should a contemporary adventure be considered speculative fiction? PortForlorn 21:53, 12 Nov 2006 (CST) advice. PortForlorn 21:43, 12 Nov 2006 (CST)

There will always be borderline cases, but ISFDB:Policy currently states that:
Speculative fiction is defined to exclude:
  • techno-thrillers, political thrillers and satires set in a future indistinguishable from the present (?)
so my first instinct would be to enter contemporary adventure novels as "non-genre" rather than "novels". On the other hand, the ISFDB software doesn't display Non-genre series data at the moment and I have been guilty of entering Non-genre series works as Novels to preserve Series data. Hopefully, Al will fix the Series display issues in the foreseeable future and we won't have to worry about it as much any more. Ahasuerus 22:41, 12 Nov 2006 (CST)

Changing Policy

Based on the discussions above, I am about to change ISFDB:Policy from:

In - Otherwise unrelated works found in any publication that will be cataloged based on other criteria. This is done to avoid creating incomplete biblio records for magazines, anthologies, etc.

to:

Out - Works that are unrelated to speculative fiction and are not written by speculative fiction authors that are found in any publication that will be cataloged based on other criteria. This will result in the creation of incomplete bibliographic records for non-genre magazines, anthologies, etc. and they should be marked as such in the "Note" field.

Does this sound about right? Anything else we want to change in the Policy statement? Any ideas re: quantifying the certain threshold statement, i.e.:

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 (see below). This will include any non-genre works published as standalone books as well as non-genre short fiction, but exclude non-fiction that was not published as a standalone book. Thus, Poul Anderson's mysteries and his non-fiction book about thermonuclear weapons will be included, but Gregory Benford's and Robert L. Forward's professionally published scientific articles will be excluded.
Out - Works that are not related to speculative fiction by authors who have not published works either of or about speculative fiction over a certain threshold. This "certain threshold" is hard to define, but we need to draw the line in a way that would exclude Winston Churchill, who published at least one work of borderline speculative fiction. The goal here is to avoid cataloging everything ever published by James Fenimore Cooper, Robert Louis Stevenson, Honore de Balzac and other popular authors. Instead, we would want to catalog their speculative fiction works only.
I agree with your revised text. I would say it's going to be darn difficult to draw a line for the other two, though. Perhaps we should just suggest that doubtful cases should be reviewed and a decision documented on the author's wiki project page? Mike Christie 18:14, 13 Nov 2006 (CST)

ToC vs. story title priority

In looking over the help text to get ready for go-live, I noticed that I wrote that the table of contents has priority over the story itself when picking a form for the title or author name. I think this is wrong, and that it should be the other way round; the ToC is a form of index and is not the primary source. See the title help and the author help. I will change this round unless I hear opinions to the contrary. Mike Christie 20:52, 13 Nov 2006 (CST)

Sure, sounds reasonable. Ahasuerus 22:05, 13 Nov 2006 (CST)

Stated Publication Date vs. Actual Publication Date

Expanding on the "date discussion" above, here is what I wrote in an e-mail to Marc earlier tonight:

[Potential discrepancy between the stated publication date and the actual publication date] is a known problem and different bibliographers have used different approaches in the past, but I am unaware of one definitive way to handle it. Locus, which receives copies of many SF/F/H books, gives both the date printed in the book and the actual publication date when known. When the date is unknown, they say something like "Supposedly published in 1991, but not seen until 1994", which may mean any number of things. They also note alternate editions/ISBNs that they haven't seen, e.g. limited editions. The Library of Congress generally goes with whatever the book says -- including "n.d." for "no date" -- but if the actual date is known, they will add it in square brackets, sometimes accompanied by an explanatory note.

As far as the applicability of these different approaches to the ISFDB goes, well, we have one advantage, namely software support for the "Publication -- Title" dichotomy. I wonder if we may be able to take advantage of this and enter the "physical date", i.e. the date printed in the book, in the Publication record and the "actual date" in the Title record. That way the Library of Congress' "n.d." would become "0000-00-00" in the Publication record and the "real date", e.g. 1907, would be entered in the Title record as 1907-00-00. My biggest concern with this approach is that it may confuse some of our users/editors.

The question then is whether this is something that we may be able to exploit or will it be more confusing than elucidating? Ahasuerus 00:46, 14 Nov 2006 (CST)

I think that would probably be too confusing. I do think both should be tracked, where possible. I also think that the publication record should hold the date printed in the book, no matter how inaccurate. I'd suggest that a note on the publication take care of any discrepancy between actual and printed date. If there's no printed date, we are free to use the actual date in the publication record if we want. That's a reasonably standard bibliographic approach, though as you say there is variation. Mike Christie 05:19, 14 Nov 2006 (CST)
Doesn't this issue go back to what is "knowable"?
Very much so :) Ahasuerus 15:21, 14 Nov 2006 (CST)
We know the publication date printed in the book, so the stated publication date can be determined from the primary source (so long as they actually printed one). The actual publication date can not be determined from the primary source - it requires either access to the publishers shipping records (which would give the ship date), or eyewitness testimony - like when Locus says they received the book (which gives the receive date).
This can be very time consuming in certain cases, especially for older editions. I have seen 19th century books dated by carefully revieweing their authors' letters and reconstructing the chain of events that led to the publication, which is clearly something that we are not likely to do here. More recently and more irritatingly, small presses have been the biggest offenders. Not only do they commonly send their projected publication data to Amazon.com, Library of Congress Online Catalog, etc months and years prior to actual publication (if it ever happens), but they have also been known to release books months and years after the "print date" stated in the book. Grrr! Ahasuerus 15:21, 14 Nov 2006 (CST)
I don't see any way that the actual publication date can be derived from the primary source, so it's always going to be derived from a secondary one, which will always call into question the veracity of the data.
Very much so :( Ahasuerus 15:21, 14 Nov 2006 (CST)
I can see where the difference between the stated and actual dates would be interesting if they differ by years (1991 vs 1994), but if the difference is between March 93 and April 93... is that of any real significance that people care about? In the real world, a book distributed on March 30th will experience statistical delivery variation that will cause it to be received in March or April, depending on where your store is located.
Well, the biggest problem is with books published in, say, December 2006 and stamped "January 2007". Are they 2006 books or 2007 books? And if we go with what the books claim, how do we explain to our users why these book were widely read, reviewed, etc in 2006? Ahasuerus 15:21, 14 Nov 2006 (CST)
I think this is just something the users are going to have to live with and learn about. It's a bibliographic fact of life. A few years ago I raised a similar question to the OED about their database (which I do a lot of data entry for); the answer was that this is just how publishing dates work, and that's what gets recorded. I see no problem with notes field entries, but to stay "knowable" we can't go beyond that. Mike Christie 16:24, 14 Nov 2006 (CST)
Bah, what would the OED folks know about bibliography? :-) Actually, OCLC has very detailed guidelines for publication dates (and everything else), but there are some things about them that I really don't like and, besides, I am not sure we want to get that specific for fear of confusing occasional editors. For reference purposes, here is what they use for dates (field 260 in the MARC-21 standard) (http://www.oclc.org/bibformats/en/2xx/260.shtm):

The date of publication. If the date of manufacture is substituted for the date of publication, use subfield ‡c. If there is a date of publication and a date of manufacture, use subfield ‡g for the date of manufacture. 

260 Stony Brook, N.Y. :