ISFDB:Community Portal/Archive/Archive03

From ISFDB

Jump to: navigation, search

This is an archive page for the Community Portal. Please do not edit the contents. To start a new discussion, please click here.
This archive includes discussions from

Archive Quick Links
Archives of old discussions from the Community Portal.


1 · 2 · 3 · 4 · 5 · 6 · 7 · 8 · 9 · 10 · 11 · 12 · 13 · 14 · 15 · 16 · 17 · 18 · 19 · 20 · 21 · 22 · 23 · 24 · 25 · 26 · 27 · 28 · 29 · 30 · 31 · 32 · 33 · 34 · 35 · 36 · 37 · 38 · 39 · 40 · 41 · 42



Contents

I unlinked a pub from a title and now I can't undo it.

I unlinked GRNHLLSFRT1985 from 174851, with the intent of deleting the title record, though now that I think about it, maybe it should have stayed linked? In any case, the publication display is now broken, and I can't attach it to 174851, or to any title for that matter. I had thought I could at least merge it back in, but publication merge is broken, so creating a dummy publication linked to the title and merging them didn't work. I think I may need someone to poke at the DB directly for this. grendel|khan 10:48, 21 Nov 2006 (CST)

Just a note to the effect that Publication merge is not available by design. Bogus Publication records should be deleted with extreme prejudice :) Ahasuerus 12:14, 21 Nov 2006 (CST)
Okay. I wasn't using it to merge publications anyway (I was trying to use it to affix a publication to a title.) Maybe the option shouldn't be there, then? grendel|khan 13:08, 21 Nov 2006 (CST)
Oops, I didn't realize it was still accessible via the Advanced Search option! Do we want to get rid of it, Al? Ahasuerus 13:30, 21 Nov 2006 (CST)
Ridded. Alvonruff 14:56, 21 Nov 2006 (CST)


I unlinked a title from an author and now I can't undo it

The help section “Editing:Making_a_Title_a_Variant#Relink_a_variant_title_to_the_correct_publication.” is empty and so I tried to wing it… Earlier I had made a couple of titles variants but that had the undesirable (to me) side affect of removing the titles from the pseudonymous author’s page and so I figured I would move the titles back. At the moment I have

  • Title record 17394 - Solar Kill – this is the original title record. This record “disappeared” from view when the variant was created.
  • Title record 199021 - Solar Kill - This was created when making the variant but is now “gone” as I tried to merged it with title 17394 to see if that would undo the variant links.

At the moment Title record 17394 is “invisible” or perhaps an orphan in that not showing up on the author’s page though I can locate it via advanced search. Ignore Solar Kill (30670) – that will be merged with 17394 once things get cleaned up.

  • Title record 17401 - Alien Salute – this is the original title record. This record “disappeared” from view when the variant was created though it’s still available from the advanced search.
  • Title record 199031 - Alien Salute – This was created when making the variant.

I edited 199031 to change the author name from "Rhondi_A._Vilott_Salsitz" (the real name) to "Charles Ingrid" (the pseudonym) and at the moment the title is showing on Charles Ingrid's page as "Alien Salute (1989) {as by Charles Ingrid}." Once I understand how to undo a variant I'll fix this record up too. Marc Kupper 02:25, 22 Nov 2006 (CST)

An update on this - Title record 17394 - Solar Kill is only missing from the author's Long Works page. The record is visible on the alphabetical and chronological pages. Marc Kupper 02:37, 22 Nov 2006 (CST)

I think I know how that "The Two Towers" had the doubled publication contents.

Consider THRLLNGSTP2005. I created it as follows:

  1. I created THRLLNGSTN2005 with the Add Publication to This Title tool.
  2. I cloned it and changed the ISBN, noting that it had 1284 listed as contents at that point.
  3. I accepted the edit in the moderation screen, and saw there that it had contents listed. (Specifically, one copy of the title it was created for.)
  4. The resultant new publication was listed twice in the title list, and appeared to contain four copies of itself when viewed, or two when edited.

Hope this helps track it down. grendel|khan 11:38, 21 Nov 2006 (CST)

Created bug 10051 to capture this. Mike Christie 09:42, 25 Nov 2006 (CST)

Legal name vs. Birth name

We have a field for the author's legal name, which works well for simple cases, e.g. pseudonyms or abbreviated names. However, what happens if an author's legal name changes over time? A number of genre writers -- off the top of my head, Andre Norton and Nicholas Yermakov/Simon Hawke -- have changed their legal name to their primary pseudonyms and, besides, people have been known to get married. Do we want more than one field to accommodate these cases? Ahasuerus 19:35, 21 Nov 2006 (CST)

I think we can give it a definition (e.g. most recent legal name) and leave anything else to be documented in the notes field on the author record. Mike Christie 20:55, 21 Nov 2006 (CST)
I know what I want here, I just don't think there's a convenient, all-encompassing label that neatly fits. This should be "the-name-which-would-show-up-in-an-encyclopedia-article" field. It's not exactly the legal name, as in the case of Simon Hawke it leads to loss of useful and interesting information. It's not exactly the birth name, because we get weird results for female authors who married and have a later married name (James Tiptree, Jr. vs Alice Sheldon vs Alice Bradley). I don't think I want three different fields for Tiptree, so I'm in agreement with Mike that we should find some definition that works and document the rest. Alvonruff 09:57, 22 Nov 2006 (CST)

Stray Publications?

As part of ISFDB:Community_Portal#What_it_looks_like_when_I_stick_to_the_rules_I_wrote I took a look at the author named “Not Available” and fixed all four outstanding publications as the author names are available. I then hit refresh on the author page – the four novels disappeared as expected but was replaced by a list of “Stray Publications.” Google finds ISFDB:Community_Portal/Archive#Editing_omnibuses that mentions this but does not address why these only show when there are no long works. Google also finds that there are 580 authors with “Stray Publications” (meaning no long works) – I think many of them are magazine stories. Marc Kupper 03:11, 22 Nov 2006 (CST)

"Stray Publications" are typically Publications whose Author(s) do not match the Author(s) of the Title that they contain. They are only displayed in Long Works if there are no legitimate Titles for the given Author. There is a fair amount of complexity behind the scenes due to pseudonyms messing things up, but whenever Stray Publications appear, I pull them up in Publication Edit and 90% of the time there is a discrepancy between the Pub and the Title that needs to be fixed.
In this case, half the Publications do not belong in the ISFDB, so we can zap them, and the other half (like that MZB anthology) need some TLC and cleanup. Ahasuerus 09:46, 22 Nov 2006 (CST)
(Ahasuerus beat me to the Save Page button). This probably needs to be titled something more straightforward, but the "Stray Publication" warning is intended to show that something needs fixing. In this case the title records have been changed such the author in the title records is now "unknown". But the old publication records still point to the author "Not Available". "Not Available" doesn't have a bibliography, since he doesn't have any title records.
In this case if we want to preserve the author "Not Available" on the publication records (which seems unlikely), then it would need to be made a pseudonym of "unknown". If we want to get rid of "Not Available" altogether (which seems more likely in this case), then the publication records need to be updated to use "unknown". The quickest way would be to do an advanced author search on "not available" OR "unknown", and then do an author merge selecting the "All records" option. Alvonruff 09:49, 22 Nov 2006 (CST)

Bug lists

I see Ahasuerus closing and moving bugs. I'd originally thought when I numbered them that they would stay there when fixed, with a status change, so that the "increment last bug" method of getting the next bug number would work. I realize now that that doesn't sit very well with the "Open . . . Bugs" title of the page. What's a good compromise? Maybe we could rename the "Open" pages to just say "Display Bugs", and "Editing Bugs", and mark the Fixed pages as historical? Mike Christie 14:45, 22 Nov 2006 (CST)

Ideally we would have software support for bug tracking like the previously mentioned Bugzilla, which would give us multiple views of the open/closed/NAB/etc bugs. As long as it doesn't go on a rampage and destroy Tokyo, that is. Barring that, I think it's more useful to have a list of currently open bugs that can be reviewed quickly when reporting a suspected new bug than to have a complete history (which admittedly has its uses, including the ability to more reliably assign bug numbers) in one place. Ahasuerus 01:12, 24 Nov 2006 (CST)

Non genre

What's the intended use of the NONGENRE entry type? I assume it's intended for non-sf fiction. Mike Christie 12:20, 23 Nov 2006 (CST)

That's right. There has been pressure from the writer community to list all books, SF/F or otherwise, for any SF/F writers that are cataloged by the ISFDB. This is certainly understandable from the authors' POV: "If they like my fantasies written as by Matthew Hughes and/or my Wolverine book written as by Hugh Matthews, maybe they will like my mysteries written as by Matt Hughes too?"
As an aside, the opposite can also be true. Writers can become unhappy (for various values of "unhappy") when we link undisclosed pseudonyms to their main biblios or, in some cases, list their legal/birth names. Some writers who have changed gender have been known to become quite upset about the latter practice.
Anyway, since we are not limited by mundane considerations like page counts, we can certainly go along with this expansion of the project scope as long as two conditions are satisfied. First, non-genre works need to be clearly labeled as such since otehrwise our users may order what they think are a bunch of spec fiction books and end up with a pile of mysteries or romances. Second, we don't want things to get out of hand. For example, if a writer like Poul Anderson or Andre Norton wrote 100+ books of genre interest and about a dozen non-genre books, we can easily list the latter. But if somebody like E. Phillips Oppenheim wrote only half a dozen borderline SF/F books and 150 non-genre books, do we really want to list all of them? Probably not. Ahasuerus 00:59, 24 Nov 2006 (CST)

However, there's no way to mark a new novel as non genre as it doesn't show up in the pulldown at the publication level. Mike Christie 12:20, 23 Nov 2006 (CST)

At this point the software only recognizes NONGENRE Titles, but not NONGENRE Publications. Which makes sense in a way since "NONGENRE" can be a subjective thing while "NOVEL" is much closer to the "immutable truth". Ahasuerus 00:59, 24 Nov 2006 (CST)

It does show up for the title; are you supposed to create the book as a novel, then use editpub to change the title entry type to nongenre? Mike Christie 12:20, 23 Nov 2006 (CST)

That's the workaround currently in place. At one point I suggested adding "NONGENRE" to the dropdown list so that the software could then create a NOVEL Publication linked to a NONGENRE Title. Al seemed to be reluctant to go along because, IIRCH, based on prior unfortunate experiences, he wanted submissions to stick as close to the "objective" stuff as possible. But I am sure he will chime in re: this issue as soon as he wrestles that turkey to the ground :) Ahasuerus 00:59, 24 Nov 2006 (CST)

It also appears we don't differentiate between nongenre short fiction and novels; is that correct? Mike Christie 12:20, 23 Nov 2006 (CST)

There are a few grey areas in the current categorization scheme and I believe that this is one of them. Ahasuerus 00:59, 24 Nov 2006 (CST)

Cloning question

The "Clone this Pub" tool displays the following note: "You may add titles, but you cannot delete or modify titles. Modification of a title requires knowing whether you intend to modify the canonical title or wish to create a variant title. If you need to change the canonical title, do so before cloning the publication. If you need to create a variant title, first clone the publication, and then go back and make repairs after integration." Can someone explain what the third sentence here means? Surely, either way, the simplest thing to do for a non-matching title is to clone, then delete the content entry and re-add it, finally manually merging it as either a variant or the canonical version. Is there a simpler way being referred to here? Mike Christie 06:05, 24 Nov 2006 (CST)

You're saying the same thing as the third sentence. The bottomline is that the cloning tool is for cloning, not performing significant (i.e. complicated) editing. Anything complicated requires, as you state, cloning then editing. Alvonruff 11:12, 24 Nov 2006 (CST)

Same-ISBN publication records.

I fetched a database dump and ran select pub_title,pub_tag,pub_isbn,count(*) as c from pubs where pub_isbn like '0%' group by pub_isbn having c > 3;. There appear to be many duplicate publications with same publisher, same ISBN, same pagecount, but different year. This likely refers to just a reprinting with no new ISBN, right? Should these duplicate publications be deleted? I wanted to be sure before I went in and started cutting away at data. (Examples are 5936 and 2803.) grendel|khan 10:32, 24 Nov 2006 (CST)

I'm sure there are occasions when two printings will have the same ISBN. I had a quick look on my shelves and can't find an example, but I don't think we can delete without having further evidence. In the case you cite, two of the publications are complete duplicates, so it seems pretty definite that one can be deleted. There are others with slightly different dates that are likely to be reprints.
I'm putting instructions in the help saying that the printing should be noted if it's not a first printing. That would help disambiguate some of these situations. Mike Christie 11:06, 24 Nov 2006 (CST)

Content Deletion

I just tried deleting content and I'm not sure I should approve the results. The pub http://www.isfdb.org/cgi-bin/pl.cgi?DSTNNOVDEC78 shows four duplicate entries for each of two titles. I tried "Remove Titles from the Pub" and saw only two duplicate entries (instead of the four duplicates) for each of those two titles. I selected the second of both duplicate entries and submitted the edit. When I saw the moderator's approval notes, it appeared that both entries for both titles were about to be deleted. Here I stopped, suspecting that something was strange with the database. The edit is in the queue. PortForlorn 23:58, 24 Nov 2006 (CST)

This is a recently identified issue with duplicate contents-Publication pointers sometimes appearing in the database. Al is aware of the issue, but I don't think he knows what is causing it yet. Whenever I run into this problem, I open a separate window to keep the data handy, then delete the problem contents item and re-add it. This is the only method that has worked for me so far, but I may well be mssing something. Ahasuerus 01:01, 25 Nov 2006 (CST)
OK, I'll do it that way for now. PortForlorn 01:28, 25 Nov 2006 (CST)

Magazines, pb Anthology Series, Awards and James Baen

I looked at the title and pub listings for James Baen, James P. Baen, and Jim Baen (listed as an pseudonym of James Baen). There was a pb anthology series edited by James Baen called "Destinies" (and other, later series called "Far Frontiers" and "New Destinies"). Althought "Destinies" is not listed as one of the ISFDB magazines, that was how the editor/publisher described it. And "Magazine" is the category it won the Locus Poll Award over several years. The awards were listed in the name of James P. Baen - for four years - but only identified the magazine title, "Destinies", not the individual issues each year. I'm not sure how the awards data should be handled when James P. Baen is edited to show that it is a pseudonym of James Baen.

The problem seems complicated since each award was for that year's collection of issues - not for one of them. We could reclassify "Destines" as a magazine rather than an anthology series. But that would remove the list of 11 individual issues from his summary bibliography and substitute the summary magazine/year reference. This seems a bit misleading since each issue was published as a paperback book. We could leave them as an anthology series and edit the awards data so that all the issues for the award year were listed. But that isn't what Locus actually said.

The situation is simpler for "Far Frontiers" and "New Destinies" since the awards data isn't involved. They were published as pb anthology series. But "New Destinies" also has, on its cover, the announcement: "The Paperback Magazine of Science Fiction and Speculative Fact" (just as "Destinies" did). Is it a magazine or a pb anthology? "Far Frontiers" (edited by Pournelle and Baen) was not described as a magazine but it has a similar format: each pb issue has the same title and they are only distinguished by their Volume/date numbers.

As a minor aside, I noticed that http://www.isfdb.org/cgi-bin/title.cgi?29820 shows as edited by Jim Baen AND James Baen rather than Jim Baen AS BY James Baen. And is there any way to get all the titles listed for the pseudonym shown under the canonical name? PortForlorn 01:28, 25 Nov 2006 (CST)

I struggled with this when writing help text recently; here's all I could come up with:
"It can be difficult in some cases to determine if something should be regarded as a magazine or a book. Some magazines were published in book format; some books were published as series with letter columns and regular dates of publication. Borderline cases should be discussed on the magazine or book wiki pages, but generally a magazine must have a common title from issue to issue, and an enumeration or dating system of some kind. This still leaves anthology series such as New Worlds Quarterly as judgement calls. In these cases, look for a consensus on the publication bibliographic wiki page. If no discussion exists, use your best judgement and document the decision on the wiki page."
In this case I think I'd go ahead and call Destinies, at least, a magazine, despite the drawbacks you note. Probably all three are best listed as magazines. Mike Christie 08:00, 25 Nov 2006 (CST)
With regard to your question about pseudonym listing, doesn't it already work that way? See Megan Lindholm for example, which lists the Robin Hobb works. It appears to work correctly for short works too; see A. Bertram Chandler for example. Do you have an example where it doesn't do this? Mike Christie 08:46, 25 Nov 2006 (CST)
Pseudonyms are generally displayed "as by", but sometimes Author data in the database is messed up. In this case, some Titles/Pubs point to two Author/Editor records, i.e. both "Jim Baen" and "James Baen". I have done some work on Baen's bibliography, but there are still quite a few issues. I'll try to clean things up at the end of December when I have access to my collection which includes all Destinies, New Destinies and Far Frontiers. BTW, what would you say the canonical name for Baen should be, "Jim" or "James"? Ahasuerus 12:50, 25 Nov 2006 (CST)
Nicholls gives "Jim Baen" priority, so let's go with that. Mike Christie 12:58, 25 Nov 2006 (CST)
I'd agree with Mike because he used Jim since the 80's on his books. As for the pseudonym titles not being included with the titles in the canonical name's bibliography, I noticed it with Jim Baen and James Baen. Jim is currently shown as the pseudonym for James but Jim's titles don't appear in James' biblio. I'm not sure I understand Ahasuerus's comment about two author records. PortForlorn 15:19, 25 Nov 2006 (CST)
In Jim Baen's case some Titles list "Jim Baen" and "James Baen" as co-editors, e.g. http://www.isfdb.org/cgi-bin/title.cgi?29820 There are rare cases where it would be legitimate, e.g. John Wyndham once published a books "co-authored" with his own pseudonym, but in Jim Baen's case it's just a database error that needs to be cleaned up. Ahasuerus 18:52, 25 Nov 2006 (CST)
I thought that the pseudonym author always had his own author record
That's right. Ahasuerus 18:52, 25 Nov 2006 (CST)
but that it linked to the cannonical name record thus forcing the AS BY. PortForlorn 15:19, 25 Nov 2006 (CST)
Well, sort of. Pseudonyms are currently handled on a Title by Title basis, not on an Author by Author basis. This is primarily due to the fact that there can be dozens of writers behind a "house name" as well as other bizarre combos like the John Wyndham book mentioned above. Full software support for pseudonyms is hard -- if it wasn't, then everybody would be doing it! Ahasuerus 18:52, 25 Nov 2006 (CST)
If Jim is currently known as the pseudonym of James, shouldn't any title listed for Jim result in an AS BY James? PortForlorn 15:19, 25 Nov 2006 (CST)
Ideally, it should. However, only half the "Jim Baen" Titles are currently marked as pseudonymous Titles by "James Baen", the rest haven't been touched yet. Of course, now that we have decided to go with "Jim Baen" as the canonical name, it will all have to be redone anyway. The fun never stops :) Ahasuerus 18:52, 25 Nov 2006 (CST)
By the way, what kind of bribe would it take to get Al to recreate the old ISFDB1 capability that listed an author's pseudonyms on the summary biblio page? It would also be useful in the author search results to see any pseudonym annotated with the author's cannonical name. PortForlorn 15:19, 25 Nov 2006 (CST)

Outstanding Questions

I am close to completing a draft rewrite of the whole help system, and in the process I have discovered some areas I don't fully understand, which of course hampers my ability to write the help text. Here are three questions I'd appreciate some help with.

1. Awards. How is award data entered? I can see the tables in the ER diagram, but I haven't found any data entry screens for any of this information. Is it done behind the scenes? I'd like to know both how award nominees and winners are marked, and also how new awards are created.

Can't do it yet. I'm still mulling over the best UI paradigm to accomplish that. I'll probably put that in over the Christmas break. Alvonruff 10:35, 25 Nov 2006 (CST)
OK, glad I didn't just miss a screen that was right in front of me. I'd say this is quite low priority, since it is not very volatile. There are lots of higher priority bugs and features -- the ISFDB watchlist would definitely be top of my list. Mike Christie 10:46, 25 Nov 2006 (CST)
It's true that ISFDB award data is not very volatile, but keep in mind that there is a bug in the Title Merge code, which results in the Title-Award association being lost 50% of the time when you merge Titles. Until this bug is fixed and/or there is a way to edit Awards, we will be slowly corrupting our Award data as we review and fix Titles :( Ahasuerus 18:58, 25 Nov 2006 (CST)
Good point. I think the bug is more important than the UI: we need the bug fix regardless -- dataloss bugs are top priority. Mike Christie 08:05, 26 Nov 2006 (CST)

2. Author Merge. The Advanced Search form, when used for authors, permits Author Merge. Here's the existing help text. The discussion refers to "Robert Heinlein", the canonical form is "Robert A. Heinlein". The way I've written the help text now, a story credited to "Robert Heinlein" (as several early stories were) would be listed that way in the publication, and "Robert Heinlein" would be a pseudonym. You would definitely not want to merge these. Alternatively, if the name were entered incorrectly, you would simply correct it. So what's the use for Author Merge? It occurred to me that if a significant number of titles were entered incorrectly in a way that led to a spurious split into two authors, this might be useful. E.g. "S.P. Somtow" (kerned) vs. "S. P. Somtow". If a big tree of titles were below each of those, it would be easier to merge them than to edit each publication. Is that the intent?

Merge is for doing bulk conversions. As we move forward, this feature will become less useful as we'll have found most of the author names that need tweaking. It has been very useful up to now. Alvonruff 10:35, 25 Nov 2006 (CST)
Makes sense; I also found a reference earlier on the portal discussions to using this to merge "Not available" and "Anonymous". This gives me enough to write the help text. Mike Christie 10:46, 25 Nov 2006 (CST)

3. Find Stray Titles. When looking at an author's bibliography, there's a link to "Find Stray Titles". What does it list? I tried it on Ted Chiang and it just listed everything he's written. Incidentally, after displaying stray titles, the other bibliography links cause bizarre errors; I'm about to go report that as a bug.

The stray title finder was a new tool invented a couple of days ago. Ahasuerus had deleted the titles and publications for an author, but the author didn't get deleted from the database. There were still book reviews that referred to the deleted book, but those author references were not detectable. Hence the stray title tool. It would be nice if it only showed up in the navbar whenever there are no works for an author, but that's a bit complicated. I'll look into the bibliographic links - it's probably not passing the author info to the navbar properly. Alvonruff 10:35, 25 Nov 2006 (CST)
So the tool should only display titles that are referenced but do not exist? Right now it's displaying everything, as far as I can see. Mike Christie 10:46, 25 Nov 2006 (CST)

Apart from the above three points, I think everything else is documented. I've incorporated all of Al's help text, sometimes with a bit of rewriting to fit it into the new format. I plan to replace the link from Help:Contents that goes to Editing the ISFDB with a link to Help:ScreenList, with the same label. Let me know if you have a better idea.

I still have to go through the last couple of dozen community portal sections to find discussions of odds and ends that may need to be included. Please go ahead and edit the help wherever you can see a way to improve it. Mike Christie 08:00, 25 Nov 2006 (CST)

Excellent job! I will try to review the new Help pages over the next few days and comment on them. I am currently nursing my wrists after a 500 edit marathon session the other day -- fixing data is addictive :) Ahasuerus 18:58, 25 Nov 2006 (CST)
The Help screen now points to Help:ScreenList for the "Editing" link. I think the old help pages can all be deleted, once the new pages are reviewed. I'll leave them for a week or two in case anyone wants to refer to them for comparison; I think I got most of the useful text out of them and reused it, but no reason to delete them right away. I'm also going to delete the templates I put in them; I've reorganized the templates by source table -- there's a list on my user page.
I still need to add text for "Find Stray Titles", which I don't understand yet, and for "Publishers", which I haven't really looked at yet -- someone else can add that text if they wish. Mike Christie 08:04, 26 Nov 2006 (CST)
Since the Stray Title tool is confusing, I've removed it for now. I'll retool it to make it's meaning more intuitive. Alvonruff 09:39, 26 Nov 2006 (CST)

Using Wiki's Project pages

While reviewing various discussions over the last few days, it occurred to me that the table that Mark posted in this section is a good candidate for addition to ISFDB:Author Names Cleanup, one of our Project pages. We may want to start moving some of our other discussions (Help pages? Pseudonym support?) to Project pages as well if we want to keep them manageable. Wikis are great for some things, but sometimes you wish you had old-fashioned thread support a la Usenet or even Web forums... Ahasuerus 19:12, 25 Nov 2006 (CST)

This is an area that’s been bugging me a little but I don’t have a “great” solution in mind. My problem is that with wikitalk is that it’s rather painful to catch up as I do want to read all of the new comments for each thread/sub-thread plus an running into that “conversations” are starting to happen on multiple talk pages. Does the wiki system have a message board like interface that’ll organize posts into threads by date, etc? I’d suggest something like Google or Yahoo groups though that creates some pain in that we can’t just copy/paste good ideas, tables, etc. right into the wikitext of ISFDB help pages. Marc Kupper 17:05, 26 Nov 2006 (CST)

Wiki Site Map Generator

I just popped over to add a new question and Mike’s comment of “I am close to completing a draft rewrite of the whole help system” is relevant. The question is “Does the wiki system have a site map generator?” Lately I’ve been finding help pages by going to ISFDB, pulling up a random publication, clicking Edit, and then clicking on the link for the help page that I wanted. There must be a bunch of helpful stuff out on the wiki that I don’t know about as they are “orphan” pages that only get linked to from places like ISFDB. If a site map generator is available then setting up a page like Help or Help:Screen would be easy. Marc Kupper 17:12, 26 Nov 2006 (CST)

Related to this is that I did not understand what Ahasuerus meant by "Project pages" in #Using_Wiki's_Project_pages and so clicked on the link to discover that "oh yes, a long time ago I saw something about ISFDB projects - I wonder where?" The problem is the project page does not have a link back to ISFDB:Projects or similar page so that someone could see what the current projects are. Marc Kupper 17:23, 26 Nov 2006 (CST)

The master Project page is Bibliographic Projects in Progress, which is linked from the Main page, but I am all for adding more and better links :) Ahasuerus 18:28, 26 Nov 2006 (CST)
When browsing about I found the MediaWiki AllPages thing which does exactly what I was looking for. It did bring to my attention that we have been referring to the “Author” and “Magazine” namespaces but have not actually set up things. Here are a few relevant articles:
  • Manual:Using_custom_namespaces – Talks about how to define/add a namespace and some ramifications of this. Someone must have spotted this at one time because there is already a custom namespace called “ISFDB.”
  • Manual:Flat_namespace – Discusses pro/con of “flat” vs. “deep.”
  • Project_talk:Namespaces – It seems ISFDB used to be “flat” and “developer centric” which has lead to confusion on my part as a new person trying to figure things out. I think much of the structure is already in place for a three tier “users,” “system administrators,” and “developers” model but could be better defined/documented including in places such as the Main_Page (which I had to link to as “Main_Page” rather than “Main Page” because the main page is in the generic “all pages” namespace and this page (Community Portal) is in the ISFDB namespace…). It’s also not clear to me in that while we talk about the “Author” or “Magazine” namespaces if such things should be formally defined. It would certainly clean up the Special:Allpages view and perhaps would expose that some pages belong in the currently unused Category namespace. Marc Kupper 13:29, 8 Dec 2006 (CST)

Uncredited material

There's been some discussion about this already, but I want to suggest a resolution. How would people feel about a rule that said that publication attributions have to reflect only what is apparent from the publication itself? For example, an editorial by John Campbell, that's well-known to be by him, but which is signed only "The Editor", should be listed with author "The Editor" in the publication. Then a variant title is created to assign that work to him, and he gets a pseudonym of "The Editor".

Does this seem reasonable? I have been assigning editorial material in Fantastic Universe on a few occasions, where I felt fairly safe doing so. However, I find I'm uncomfortable with creating publication records that don't accurately reflect the publication. I've been adding notes to the publication, saying what deductions I've made, so technically I think it's OK, but I think it might be better if the rule for author and artist attribution was "if it's not in the publication, it doesn't go in the publication record". That rule might not apply to publisher, or price, since there's no equivalent to the "variant" mechanism for recording alternative sources of information. In those cases we'd just enter the data and make a note. But given that we do have a pseudonym mechanism that works on the title level, should we use it to this extent? Mike Christie 20:34, 26 Nov 2006 (CST)

In the spirit of sticking to "objective information", I would be inclined to agree. Genre history suggests that just because an essay is signed by "The Editor" and page 1 proclaims that John Smith is the current editor of this fine magazine doesn't necessarily mean that the article was written by John Smith. There has been so much ghosting (and plain delegation) over the years that the assumption "The Editor = John Smith" strays too far away from the "objective information" standard for comfort. It's more work for us, but I think the extra effort can be justified. Ahasuerus 21:21, 26 Nov 2006 (CST)
I would agree on sticking to "objective information" information but am nervous about how when you make the variant titles that works will disappear from the pseudonym’s author page. See #Pseudonyms. For example, I just added the cover artist name for a publication. The publication stated the cover as “Romas” and that’s what I entered. I’ll assume this is Romas Kukalis who routinely did covers for DAW at the time but decided to not make the title record a variant title to link it to Romas Kukalis as it’ll yank that title away from the “author” page for “Romas.” The main downside is there are now two author records for the same person. Marc Kupper 02:32, 27 Nov 2006 (CST)
I agree this is not ideal, but it's a display issue and I think should be fixed as such. I don't want to compromise the source data for that reason -- I think we can get the display improved and in the meantime enter the information correctly, as you did for Romas.
I'm going to change the help files to correspond to this approach, and will also go back through the Fantastic Universe issues to make them conform. Mike Christie 19:33, 27 Nov 2006 (CST)

Publications: cloning and splitting

The Publication cloning tool is very nice, but there are two related scenarios that we currently have no way of handling aside from re-entering and merging all Contents data manually.

The first scenario occurs when you find a Publication that shouldn't be under its current Title. This happens fairly frequently with collections when you discover that a later reprint edition contained more (or fewer) fiction Titles. You would then want to create a new Title and make it a variant Title of the original Title plus "(expanded)" or "(abridged)" added, but the only way to do it would be to delete the existing Publication and re-enter/merge all of its Contents titles, which is a pain. The obvious solution would be to add a "Make this Publication into a New Title" option on the navbar, which I think Al briefly mentioned some months ago.

I don't see this on the feature list; it probably should be added. Do we need feature numbers, like the bug numbers? Mike Christie 10:18, 27 Nov 2006 (CST)

The second scenario happens when you find a Title with numerous Publications associated with it, e.g. http://www.isfdb.org/cgi-bin/title.cgi?13556 , where only the first Publication (edition) (http://www.isfdb.org/cgi-bin/pl.cgi?THWLKLNG311900 in this case) has Contents data. The only currently supported way to propagate Contents data from the first edition to the rest of them is to clone the first edition, then update the newly created Publication with Publication-specific data from the existing Publication, and then delete the current Publication. Doable, but somewhat time consuming. One possible solution would be to add an "Existing Publication ID" field to the Cloning tool. When this field is populated, the cloning process, instead of creating a new Publication, would copy the contents of the current Publication to the one that was identified by the specified Publication ID. If the destination Publication already has Contents data, the cloning process would presumably fail since merging the Contents data of two Publications is potentially too messy (?).

Does this sounds reasonable/desirable? Ahasuerus 21:40, 26 Nov 2006 (CST)

Sounds worthwhile to me. I don't think either of these is needed for go-live, but they should be in the roadmap. Mike Christie 10:18, 27 Nov 2006 (CST)

Adding contents to an omnibus

I decided to take a spin with adding contents to an omnibus. The process itself seemed to go well but now when I look on the author’s page I see the two novels (The Spell Sword & The Forbidden Tower) contained in this omnibus floating down in the novels. Does this mean that each time I add the contents to an omnibus that I’ll need to then merge the new title record with the existing one? Marc Kupper 02:00, 27 Nov 2006 (CST)

Unfortunately, that's the way it currently works. There have been various discussions about improving the process (ultimately going to AJAX, etc), but nothing definite so far. It does raise the bar for current and especially potential editors, but in absence of a better data validation process, enabling auto-merging of newly created records is asking for trouble :( Ahasuerus 09:32, 27 Nov 2006 (CST)

One minor complication, and maybe this is why I got a new title record, is that the title (and the omnibus) are part of a series.

A thought that came to mind is that when I was adding the contents for this I was wondering why I could not just enter the title-ids for the existing novels as I was entering data I knew was redundant (title, author, copyright, etc.) and also needing to look up the dates on the copyright page. Marc Kupper 02:00, 27 Nov 2006 (CST)

That would be one way to do it and may well be the easiest one for now. Ahasuerus 09:32, 27 Nov 2006 (CST)

Watchlisting Requirements

We need to start accumulating requirements for the watchlisting feature, as there is a general belief we need to have it prior to going live. Requirements should be of two forms:

  • What changes trigger an event? Adding a book to a watchlisted author is a no-brainer. Modifying a title of which the watchlisted person is an author is a no-brainer. What about changes to titles in an anthology or magazine, which was edited by the watchlisted author? Are we only watchlisting authors, or are we watching other datatypes as well?
  • What happens when a watchlist event occurs?

Alvonruff 06:47, 28 Nov 2006 (CST)

Of course I support the watchlist, but I don't know that everyone does -- Ahasuerus is unconvinced, I think, for example. I've started a separate topic below about what should be done by beta, but here are my answers to the questions above.
  • I think we need to be able to watchlist titles; I think authors are lower priority, and publications may not be needed at all. Watchlisting awards isn't necessary because we can't edit them; series can be handled via titles; and publishers are free text so that would be very hard to do well, and shouldn't be in the beta.
  • For a watchlisted title, the triggers are:
    • changes to the title record
    • adding or deleting it as a parent of a variant title record
    • changes (add, modify, delete) to any of the content records for that title
    • title merges that involve it
  • For a watchlisted author, the triggers are
    • changes to the author record
    • changes to any of the author's title records
    • author merges that involve it
    • addition or deletion of title records that refer to the author
So the answer to your question is yes, if you watchlist Campbell, you're notified of any changes such as an edit to the title record that is part of the contents of an Astounding he edited. Future enhancements could include the ability to fine tune watchlists to exclude magazines or anthologies, for example, but I don't think we need that now -- and as I said, I don't think we even need author watchlists first time out. An author project could manage with just titles.
When a watchlist event occurs, nothing happens. When you click "My watchlist" in the ISFDB, a list of approved submission records is displayed, sorted by title, that meet your watchlist criteria. Clicking on the submission record displays the same screen as moderator approval, but without the approve/reject button.
The other change needed would be an "Watch this Title" and "Unwatch this Title" link in the navbar, and similarly for authors if we implement that. Mike Christie 07:45, 28 Nov 2006 (CST)
The reason I am not convinced is that I am not sure what the scope of this task will be and how long it will take to implement, although Mike's requirements are a good start and I am currently in the process of trying to digest them. I wonder if we could start with the low hanging fruit and update the Recent Edits section by linking the Subject field to the updated Publication/Title/Author record as well as adding a new field -- call it Edit # -- that would be hyperlinked to the same "Before and After" form that moderators see when they approve submissions? Ahasuerus 13:13, 28 Nov 2006 (CST)
I'd personally can the project for now. It's a good abd useful idea but having coded several watchlist style systems in the past I can say it's a big messy thing. Marc Kupper 21:33, 30 Nov 2006 (CST)

Where to talk about the beta

Where are we going to talk about what's needed for the beta? I'm mindful of Ahasuerus's comment that we need to stop having every single conversation on this one page. Some thoughts:

  • A feature list roadmap showing what's in and what's out for the beta seems necessary. Personally I don't think we need the pseudonym editor, for example; I'd rather have, e.g., a resubmit capability for rejected edits. I think the feature list talk page seems the right page for that discussion.
  • How to recruit beta editors, how many to recruit, and what the goals and exit criteria for the beta are. Maybe this should be a new page; perhaps Project:Beta.
  • What bugs need to be fixed before beta. Possibly a section on Project:Beta, or perhaps on the talk pages of the buglist pages?
  • What further page editing and organization is needed to the Wiki itself before beta. E.g. cleaning out the old help files; improving the project page links and organization, adding new help pages for any new functionality, and anything else we think we might need. Parts of this could go on Project:Beta, but perhaps some of it could go on the relevant talk pages -- perhaps Talk:Bibliographic Projects in Progress would be a good page to organize projects from, for example.
  • Do we want to assign tasks amongst ourselves? E.g one of us takes on project organization, another works on beta planning? I know Al is going to have to do all the ISFDB work, unless someone else here knows Python and the task is farmable.

-- Mike Christie 07:45, 28 Nov 2006 (CST)

Good questions, Mike. I am trying to think this through, but at the moment I am a little groggy and overmedicated due to some pebbles stuck in the system , so I am not making much progress. I'll post something on the new Project page once I am borderline coherent. (No sympathy posts, please: It's all my fault - I should have gotten an extended warranty for this hardware!)Ahasuerus 20:41, 29 Nov 2006 (CST)

Unmerge?

I see comments on "unmerge" in the notes on 10006; what's an unmerge? Is this a feature I missed writing help for? Mike Christie 19:12, 28 Nov 2006 (CST)

There used to be two unmerges: publication and title. The publication unmerge was a horrible mistake and support was revoked. This lead to the widespread purge of all things unmerge, which unadvertently removed the title unmerge function, which I just found while looking at the bug Ahasuerus filed.
Let's say that someone mistakenly merged two titles together and the submission somehow got through our top-notch moderation process and was integrated. Rather than deleting publications from the existing title, and re-entering them into a new version of the old title that was lost after the merge, an unmerge will take all of the publications under a title, and break each one out into it's own title. Example: let's say there is a title that looks like this:
 Title: Darwin's Blade
 Author: Dan Simmons
 Year: 2000
 Publications:
   * Darwin's Blade (2000 , William Morrow, 0380973693, $25.00, 368pp, hc)
   * Darwin's Blade: A Novel of Suspense (Oct 2001 , HarperTorch, 0380789183, $7.50, tp)
   * Das Schlangenhaupt. (Mar 2002 , Goldmann, 3442451051, tp)
   * Bisturi de Darwin (Jul 2002 , Diagonal -Grupo 62, 8495808072, $60.90, tp)
Performing an unmerge on this title (not that you would, but as an exanmple) would bust up these publications and create titles, such that the following titles would now exist:
   * Darwin's Blade (2000)
   * Darwin's Blade: A Novel of Suspense (2001)
   * Das Schlangenhaupt. (2002)
   * Bisturi de Darwin (2002)
Any duplicates would have to be remerged by hand. Alvonruff 19:32, 28 Nov 2006 (CST)
(This time Al beat me to the Save button). At this point, the option gives you no choice but to convert all Publications for a given Title into separate Titles, but that Submit button makes me wonder if Al may be planning to convert it into a form with checkboxes that will let you select which Publications you want to Unmerge from the Title. That would save a lot of re-merging time when there are only half a dozen erroneously merged Publications out of a few dozen (see the Bramah example above). Ahasuerus 19:39, 28 Nov 2006 (CST)
That's a good idea. For later. Alvonruff 20:27, 28 Nov 2006 (CST)

Reviving Bibliographic Rules

FYI, in the spirit of moving different discussions to different pages, I just tried to revive Bibliographic Rules with a post on Multiple Authors with Identical Names. Ahasuerus 23:54, 29 Nov 2006 (CST)

A related FYI: I'm going to go through Bibliographic Rules and update help with any conclusions. I'll add a note to each Bibliographic Rules section about what impact it will have on help text.
I also think there's a discussion to have about the relationship between help and bibliographic rules; I'll start that at Talk:Bibliographic Rules. Mike Christie 19:04, 30 Nov 2006 (CST)

Anonymous, uncredited, etc.

(Salvaged from an archived section as it may be useful for ongoing cleanup. Feel free to move it to an appropriate project page if there is one. Mike Christie 08:37, 2 Dec 2006 (CST))

Isfdb seems to have the following: Marc Kupper 14:10, 16 Nov 2006 (CST)

AuthorLongShortAwardsComments
Anonymous~150~2506
Anonyous030Should get merged with Anonymous [Fixed Marc Kupper 02:47, 22 Nov 2006 (CST)]
Not Available400This sounds like our "not stated" [Fixed - I checked each of the four titles and for each was able to dig up the author name. Marc Kupper 03:01, 22 Nov 2006 (CST)]
uncredit010Should get merged with uncredited. [Fixed Marc Kupper 02:47, 22 Nov 2006 (CST)]
uncredited0~1,0000
Unknown~1,050~550172Used for many cover/interior art instead of "unsigned"
unknown~1,050~550172Both "unknown" and "Unknown" get used but search returned same list
unknownAfghan010No explanation in the story notes about this
unsigned4500All 45 long works are for Interior Art
Unknown Unknown Listed on http://www.isfdb.org/DIR_U.html but not found
Various6900
Various Authors200Should get merged into Various [both were bad entries and have been deleted. Ahasuerus]


Reorganization of Bibliographic Rules and Help

I was going to start this at Talk:Bibliographic Rules, but I think it's more general than that. I think that BR has a lot of overlap with the help files, and we need to regularize this. For instance, the detailed help file templates contain a lot of information about what should be entered in every field. Any relevant information about the data entry has to end up in the help files, either explicitly or by reference. I've opted for putting it in there explicitly, so I think that reduces the utility of the BR page.

Discussions of the help rules should, I think, be conducted on the talk pages of those templates. For example, if you want to debate whether a ToC entry for an author should be higher or lower priority for spelling than the entry on the heading for the story itself, then Template talk:TitleFields:Author is where you should start, since Template:TitleFields:Author is where that rule is given.

There are some things that could legitimately be included by reference. For example, there's a section on scope that is probably better left as a separate policy statement. It's already been duplicated in ISFDB:Policy, in fact, and that's a better home for it.

I suggest I archive each of the current BR discussions once I've gone through and made sure that their conclusions are reflected in the help file, and then post a heading to the BR page saying that it is for debates on bibliographic rules, but only where they do not fit on the talk page for an appropriate help file or help template. If we find no debates occur there, we can close the page down. Mike Christie 09:11, 2 Dec 2006 (CST)

Well, back when I first created Bibliographic Rules, we had no other place to discuss these issue aside from the Community Portal and we had no Help pages. Now that we have more formal and more detailed Help pages that describe how data should be entered intho the systen, we can certainly move low level discussions to individual Help/Talk pages. We just need to make sure that everybody can find these discussions easily :) Also, there will probably always be a need for high level discussions re: policy issues that are too generic to be handled on individual Talk pages within the Help namespace. Ahasuerus 18:19, 2 Dec 2006 (CST)
Agreed on both counts. I'll go ahead and archive the page some time today, and put up a draft notice about where to go for certain kinds of discussion. Mike Christie 08:02, 3 Dec 2006 (CST)

Using the Storylen / Length field to note series/volume numbers

On Tanya Huff’s page I saw

I was curious as to how the “/1,2” etc. was added to the omnibus indicator and on editing the title record learned it's was done by setting the “Storylen” field to “/1,2”. This use of the “Length” (aka “Storylen”) field is not documented and I'm wondering if it's in error or should be documented as a standard ISFDB practice. I certainly found the indicator helpful though it does mean a little double entry of data. Marc Kupper 15:45, 2 Dec 2006 (CST)

The field has been doing double duty for a long time. The notation itself comes from John Wenn's biblio lists back in the early 1990s. "Storylen" is also used for things like "jvn" (YA novels), so it's currently heavily overloaded. I think that Al's main reason for not splitting it into multiple fields is the desire to keep the number of fields per form manageable. And yes, we should certainly document the current usage. Ahasuerus 17:33, 2 Dec 2006 (CST)
I've added a description of this to the help page. I didn't cover the "jvn"; where does that get displayed? Mike Christie 08:10, 3 Dec 2006 (CST)
To quote an ancient Database Schema document:
title_storylen - This column has been overloaded to mean numerous things. For shortfiction, nv=novella, nt=novelette, ss=shortstory, sf=shortfiction (unknown length). For novels jvn=juvenile, nvz=novelization. This is totally messed up and needs a new mechanism.
Emphasis added :) Ahasuerus 12:58, 3 Dec 2006 (CST)

Determining the source(s) of ISFDB data

Copied from Author:Franklin_W._Dixon: Something I wondered about a long time ago is when I’m looking at an ISFDB record (author, title, publication, etc.) is how can I determine the source of the information and its update history? Sometimes knowing the source, a dope smoking penguin for example, will help me decide if something is just garbage data and safe to change or if the data is from a typically reliable source (Locus Magazine for example) then I’m better off creating a new or variant record rather than overwriting the existing data. Marc Kupper 03:13, 3 Dec 2006 (CST)

This is a known problem and the driving force behind the Verification system which has taken shape over the last few months. In absence of Verification flags, you can check the usual suspects listed in our Sources of Bibliographic Information, although it can be quite time consuming. Ahasuerus 14:09, 3 Dec 2006 (CST)
My question was aimed more at that when I'm verifying today I will look and may see 20 publications and I see one that's close to my copy but not perfect. I'm still unsure if it's safe to edit that publication to make it a perfect match with my publication or if I should clone it and edit the copy. If the record is from an “unreliable” source then everyone is better off with the publication record getting edited. With a DAW Books publication I'm familiar enough with what should or should not be out there that I can make this judgment call without knowing the source but for something like a Macmillan reprint from the 1960s I'll have no idea if a record is one that a person has carefully edited or if it got scraped together by the dissembler. Marc Kupper 00:29, 7 Dec 2006 (CST)
That's precisely the "known problem" that I was referring to above :) Prior to the implementation of the Verification flag, there was no way of telling where the data had come from. Hence it was impossible to determine by looking at the data whether the difference between your edition and the edition(s) listed in the ISFDB was due to a real difference between two real printings or whether it was due to bad data in the ISFDB.
The good news is that once most Publications cataloged by the ISFDB have been Verified, we will have a high(er) degree of confidence that any differences are real. The bad news is that it doesn't help much while the percentage of Verified Publications is still very low, but things should improve over time. For now, the only recourse that I am aware of is to consult the "usual suspects" mentioned above and see if the ISFDB data is obviously in error. Not a great solution, but at least there is light at the end of the tunnel... Ahasuerus 10:36, 7 Dec 2006 (CST)
My recommendation would definitely be to edit the copy you see to match the copy you have in hand, and mark it as verified. I don't think unverified records will have much value to the ISFDB, eventually, and a verified record from you is far more valuable than an unverified one of indeterminate origin with no verification and no promise there will ever be verification. If the data is close enough to your copy that you think it's reasonable to believe it's your copy, edit it. Mike Christie 18:42, 7 Dec 2006 (CST)

Order of authors in a collaboration

The help text currently says to make the order of authors alphabetical. I did these because I'd seen some problem caused by having different publications of a title entered with the authors in different orders. I now can't think of what the problem was; perhaps just that reviews don't connect correctly if the order differs? If anyone has seen this issue, can you tell me what it is? Mike Christie 21:44, 6 Dec 2006 (CST)

The only issue that I am aware of is that the Title Merge screen will ask you to select the preferred order if you are merging "A+B" with "B+A", but AFAIK it's harmless. Ahasuerus 23:03, 6 Dec 2006 (CST)

The related question (which could be posted at Template:PublicationFields:Author, but I'm here so I might as well ask it) is whether we should make the order reflect what's on the title page or not? The "alphabetical" instructions will lead to reversing some well known pairings, in some cases making the "with" author come first -- e.g. "Donald A. Wollheim with Arthur W. Saha", who edited some of the DAW anthologies. Any preferences? Mike Christie 21:44, 6 Dec 2006 (CST)

If we decide to capture the stated order of author names in a collaboration, then I assume that we would have to record any reversals in subsequent reprints, making them, mm, variant titles, I guess? I am not sure it's worth the effort. Ahasuerus 23:03, 6 Dec 2006 (CST)
I personally am picky about the author order which is why I engaged in a battle with ISFDB (Open_Editing_Bugs # 10054) to get the order to show the way it's stated. But, I also care more about the order at the publication level than at the title level as I see the title records as a "merged consensus" of the publications. Hopefully if someone researches just what the consensus seems to be then ISFDB will reliably preserve whatever the person enters and not change the order of things. Related to this are the collections or anthologies where page numbers have not been entered yet – the order ISFDB uses by default seems baffling and it leads to a rather long hunt-n-peck session just to get the page numbers entered so that I can say “yes, this record now matches my publication.” Marc Kupper 00:50, 7 Dec 2006 (CST)
There is NO author order in the ISFDB. There is only the set of authors associated with a book. The author order as returned by the SQL query will depend upon the order of records in the author/title mapping table, and can subsequently be shuffled by the Python scripts as they collect/translate the data into HTML. The same is true of title order in a collection - there is no way for the ISFDB to know what order the titles appeared in unless you tell it - with page numbers. You can't depend on the order that they appear in the database - because that is EXACTLY what the ISFDB is doing now. If you first have collection A, with stories in order 1, 2, 3, and 4; and then have anthology B with stories in order 5, 3, 2, 6 - then the order in the title table will be: 1, 2, 3, 4, 5, 6. Displaying the anthology without page numbers would then give an order of 2, 3, 5, 6. There's simply no way to preserve the title order without specifically stating that order, and we're using page numbers to do that.
The only way to implement title ordering without page numbers would be to put a hidden ordinal that isn't displayed. That would lead to other problems. For instance when I enter a magazine I first enter the data in the table of contents and submit, and then I carefully thumb through the magazine page by page to find errors in the TOC, enter interior artwork, and enter book reviews. Without page numbers, but using ordinals, this would result in the wrong order unless I could do title INSERTS instead of APPENDS, which greatly complicates an HTML form (maybe in AJAX that could work). The author order could theoretically be done with ordinals in the author/title mapping table. The bottomline, however, is that there is currently no order preservation in the ISFDB. Alvonruff 07:14, 7 Dec 2006 (CST)
Isn't it a temporary problem, though? As ISFDB data improves, we will have more and more Publications with page numbers, so eventually the issue will fade away. The only scenario where the problem is likely to persist is when Publication data comes from a secondary source that doesn't list page numbers, e.g. a Mongolian library catalog, and it may be hard to Verified physically. Granted, it may take a while to get to the point where a vast majority of Publications have been physically Verified, but eventually the problem should be alleviated, shouldn't it? Ahasuerus 10:45, 7 Dec 2006 (CST)
I agree that the hidden ordinal for titles will be a pain and so let’s not try that path. A simpler idea for the edit-publication (and probably the display-publication) page(s) would be to sort unpaginated titles in some sort of order. By author or title will do, by date is probably less useful. That way when someone is looking at a list of 30+ titles in an anthology trying to assign page numbers or look for a story the visual search process will be a tad more efficient. Or, you could duplicate the title and/or author in the display HTML along with what’s in the edit field and someone can use the Ctrl-F find to search the page for the field that needs the page #.
Something that could be done as a 1-time pass would be to search for publications that have unpaginated titles, sort the list, and to assign page numbers starting at 10,000. Then the main display code would not need to change (assuming it’s ok with five digit page numbers).
A visible ordinal for the contributor names does seem needed. People care about whose name goes first and secondary contributors are nearly always at the end of the list. The ordinal would be visible on the edit page so that if there’s a need to change the order you can just change the numbers. Marc Kupper 14:12, 7 Dec 2006 (CST)
Some version of a sort order would be handy, but I don't think we need it for Beta. This would be a good feature suggestion. Mike Christie 18:44, 7 Dec 2006 (CST)

Prioritizing Data Entry/Verification

I will have access to my collection in late December - early January as well as a fair amount of free time to enter/verify data. What would you say the most efficient way of spending this time will be? I am thinking that I could start with entering the stuff from my "special" boxes which contain rare pulps, obscure 1940s/1950s British digests, early self-published pamphlets by people who eventually became professional writers (e.g. Lin Carter), etc. Alternatively, I could do physical verification of a bunch of pulps, which would go faster but wouldn't yield any new data for the database. Ahasuerus 12:33, 8 Dec 2006 (CST)

That's a really good question. Last week I found the boxes with my Analog collection (1965 to some relatively recent date), and have started systematically going through them 1 by 1. It's fairly tedious, and I find the need to mix it up with other work (like implementing reject reason support). I think that there is a ton of validation that can first be done against Contento - it's the first step I take with the magazines I do have. When I'm off I can manage to verfiy about a decade per day against Contento, which means something like F&SF will take the better part of a week to validate. So there's easily a few years worth of work to do against the stuff we already have entered.
I think I'm leaning toward the more obscure stuff as a priority, as it may not have been indexed anywhere else to date, and it's obscure - meaning fewer people will have that data on hand. Alvonruff 12:48, 8 Dec 2006 (CST)
Well, trying to think like a resource optimizer, I'd say that as moderators we'll be doing just fine if we never enter a single thing and spend time making things easier for editors. I'm working on Fantastic Universe right now, kind of as fill-in, while waiting for us to go live. I hope that there'll be enough data entry going on by editors that I'll be busy reviewing it and won't have time to enter much. Of course it's fun to enter data, so I won't give it up completely, but I think my best value to the project is going to be validating old sf data entry, particularly magazines; and perhaps also communication and new user training.
Al, for you the highest value has got to be bug fixing and enhancements. I bet you feel the same way about having fun entering data, but I don't think there's a lot of point in any of us doing much data entry right now -- it doesn't bring editing closer, and we don't (yet) have the tools to ensure the data doesn't get corrupted. Marc's data entry is different -- his data source is outside the ISFDB, so he has an audit trail for his data.
Ahasuerus's access to rare pulps etc. is special enough that I do think there's a lot of value in entering that data, more so than the pulps. There should be a fair number of people with decent magazine collections; I'm not going to bother doing verification on the digests, for example, until I see where the gaps are after some other editors have got involved.
So I'd say if Al didn't want to have fun, he should do nothing but code till we go live. That's the way to maximize value for time, I think. Al, do you want to have fun?  :o) Mike Christie 13:08, 8 Dec 2006 (CST)
I have personally flip-flopped at times on this with the conflict being generated because I see value in both having good data to act as a model/guide for new people but that attention also needs to be spent on making sure the overall system works well. Mike Christie recently verified a bunch of DAW books and that has been helpful to me as it exposed holes in my personal verification methods and how/what the DAW List and ISFDB should be showing (or not showing). It’s also triggered additional questions about ISFDB which I’ll mull over and post. Marc Kupper

Beta bugs?

I added some notes about showstoppers to ISFDB Talk:Beta#More on showstoppers. Nobody else has commented to agree or disagree; is there any problem with me changing the main beta page to show just the three showstoppers I list there? Mike Christie (talk) 07:58, 11 Dec 2006 (CST)

Well, I responded wrt the specifics the other day, but I'd like to hear from Al re: 10027 (which he reported originally) and also to see how long he thinks it will take to fix these outstanding bugs. I suspect that some of the uglier ones may require only a few minutes to fix. Ahasuerus 08:22, 11 Dec 2006 (CST)

Adding support for split publications, translations, etc.

Moving forwards from How do you record split publications? it seems like a design change to ISFDB is needed to deal with split publications, foreign language translations, etc. In thinking of what’s needed I tried to write up a complete list of what authors/publishers can do to make the lives of bibliographers interesting. The list follows. Marc Kupper 14:42, 12 Dec 2006 (CST)

  • Works can be given new titles. (The existing variant title mechanism handles titles well and the only thing missing is a way to undo a vt link).
  • Works can be given new authors. (The existing pseudonym support seems to work well but has gotchas such as it removes the works from the pseudonym’s bibliographic displays.)
  • Works can be split.
  • Works can be serialized. (This is similar to splits though the existing serialization mechanism is tied to magazines.)
  • Works can be translated to new languages.
  • Derived works can be created such as young adult, graphic, etc. versions. This could be handled like foreign language translations though often times the “translator’s” name is not stated and the resulting works are in same language as the original.
  • Works can be included in omnibuses, anthologies, collections, and magazines.
  • Works can be organized into series. (The existing series mechanism seems to handle this well and the only things that seem to be missing are a way to assign a work to multiple series, and a notes field for the series itself.) Marc Kupper 14:42, 12 Dec 2006 (CST)
When it comes to making bibliographers' lives interesting, I think there are an infinite number of spanners that books can throw in the works. Here are a couple of my favourite problem children, from when I was writing a bibliographic database for my own books back in the eighties:
  • Names that are sometimes pseudonyms and sometimes not. "Henry Kuttner & C.L. Moore" sometimes meant what it said, and sometimes meant Kuttner, and sometimes meant Moore. Sometimes even they couldn't remember which it was.
This was one of the reasons for reworking pseudonym support a year+ ago. The old method didn't handle these permutations very well while the new one seems to support them adequately. Ahasuerus 15:40, 21 Dec 2006 (CST)
  • Stories that are known to exist but which can't be identified. At least one moderately well known author has said he published a story back in the pulp days, but can't remember the title or the pseudonym he published it under.
I doubt there is much anybody can do about this one. A note on the Author's Bibliographic Notes page is all we can hope for IMO. Ahasuerus 15:40, 21 Dec 2006 (CST)
  • John Wyndham and Lucas Parkes, as authors of "The Outward Urge", are both pseudonyms of one person. I believe there's another instance of this somewhere, but I can't recall it.
Not only that, but different editions credit different authors too! Again, this was why pseudonym support has been redone and it seems to handle these bizarre cases much better. At the cost of extra complexity, regrettably. Ahasuerus 15:40, 21 Dec 2006 (CST)
  • Revised versions of stories; often to restore cuts made by editors. Are these the same "title"?
Typically one will be a Variant Title of another with either "abridged/expanded" or simply "revised" added. As we were discussing a few days ago, the approach is not perfect and may be revised in the future. Ahasuerus 15:40, 21 Dec 2006 (CST)
  • Fixups with additional linking material written but little or no change to the text of the stories fixed up; e.g. Simak's "City". Are these stories or part of a novel?
Different bibliographers draw the line in different places. Contento seems to think that almost any fixaup is a collection, but then his emphasis has historically been on collections. Ahasuerus 15:40, 21 Dec 2006 (CST)
  • Novels where the collaboration is known to have been done chapter by chapter, such as on a spoof "worst sf book" the name of which escapes me but which had a separate pseudonymous author for almost every chapter.
I think we have decided that round robin novels are considered collaborative novels rather than collections. The problem here is that we don't support more than 10 Author records per Title or at least we don't support them very well. Something to bring up with Al, probably. Ahasuerus 15:40, 21 Dec 2006 (CST)
  • Short fiction in series. Series titles that are announced but never published (e.g. some of Heinlein's original "Future History" outline). Retconning to change series numbers; e.g. Asimov fitting almost everything he wrote into a mega-series.
The current threshold for including vaporware title is their fame or notoriety, e.g. Last Dangerous Visions. I'll be the first one to admit that it is an awfully nebubulous standard. Ahasuerus 15:40, 21 Dec 2006 (CST)
Obviously this is not the right time to focus too much on all these things, since we're still trying to get the Beta out of the door. However, I do think the current paradigm is correct: publications are the root, and titles and authors come next. Everything else is built on top. For example, a connected set of derived works (graphic novels, YA, etc.) could be handled by some set of data that defines connections between titles. I suspect that we'll end up with independent architectures to handle different things: series, awards, title revisions, pseudonyms, serializations, translations. But not for a while yet. Mike Christie (talk) 17:36, 12 Dec 2006 (CST)
That's entirely possible as we continue betaing our heads against the wall, as it were :) Ahasuerus 15:40, 21 Dec 2006 (CST)

How to I move a publication to another title?

I suspect this one will end up in the FAQ but on Man in His Time is a publication that belongs under another title record but I don’t see a way to control the linkage between titles and publications. Marc Kupper 03:01, 14 Dec 2006 (CST)

Looking at the data, I'm guessing that you've already fixed this. However, if I understand the question, Unmerge Titles is the way you would do this. Mike Christie (talk) 06:50, 14 Dec 2006 (CST)
Eventually we will have the ability to Unmerge selected Publications from a Title -- see the list of requested features -- but for now you have to Unmerge all Publications for a Title, which creates a bunch of new Titles, one per unmerged Publication. You can then merge the newly created Title of the offending Publication with the Title that it belongs under and merge the rest of the unmerged Publications with the original Title. This approach is preferred when the Publication in question has a lot of Content items that would be time consuming to re-enter and re-merge. However, if (a) there are many Publications associated with a Title and (b) you want to unmerge only one and (c) that Publication has no Content data, then it may be easier to delete it and re-enter under the right Title. We should probably document these options somewhere. Ahasuerus 11:43, 14 Dec 2006 (CST)
Thank you . I ended up going with the unmerge/merge route as all of the titles involved were large collections. To add to the challenge it turned out inspection to be three separate titles all with the same name and one of them also having a variant title that was the same name as one of the short stories in the collection. To make it more apparent on what’s what to others I renamed the titles to have “(1965 edition)”, “(1971 edition)”, and “(1988 edition)” plus put them in a series. Overall, ISFDB handled it well and the tedious aspect was that there no direct way to change the linkage of a publication to a title much like how you can now change the vt linkages. Thus you need to unmerge/merge while keeping many notes to keep track of the pieces and to make sure stuff gets reassembled. Marc Kupper 19:00, 14 Dec 2006 (CST)
One of my fears is that these types of complicated reorganizations can get so involved that, like you said, they may require keeping notes. It may be tedious and error-prone for people with moderatorial privileges, but what happens to regular editors who can't approve their own submissions? They will have to wait for each individual submission to be approved before they can move to the next step and by that time they may no longer remember all the details of what they were trying to do. And how can a moderator tell that a particular Unmerge is legitimate/advisable if he doesn't know the logic behind the whole multi-step mini-project? I am thinking that at some level of complexity we are almost better off having editors describe what they want to accomplish in the Wiki and have moderators make the changes for them :-\ Ahasuerus 00:14, 15 Dec 2006 (CST)
I was thinking exactly that during my editing plus as I approved changes I was looking at the screen and realizing that I’d have a very hard time trying to understand what the editor was trying to do. Mike has already brought up a free format notes field to allow editors to pass notes to the moderator. Unfortunately, trying to clean up a large bibliography takes a lot of steps. How about this, and this is very much down the road, is to allow editors and moderators to take “snapshots” of an author and that editing with real-time (no human intervention) approval would be allowed for specific authors by specific editors. The editor should be allowed to make snapshots too so that if they are about to trudge into a particularly challenging minefield of changes that they would have an “undo” or “rollback” button. Marc Kupper 02:15, 15 Dec 2006 (CST)
Well, if we are talking "very much down the road", by then we will (hopefully) have more robust data entry capabilities which will let our editors perform multiple operations in one submission. For example, one of the biggest pains with the current setup is that entering a new anthology/collection with a few dozen stories/essays in it requires going back and manually checking which entries need to be merged and then merging them. If we could add a validation step during the submission process, e.g. an extra screen that would list all Titles by the same Author that match each new Title in the Contents section (along the lines of the Title Merge page), then we could eliminate well over 50% of all potential submissions.
Ideally, we would have some kind of on the fly validation using Web 2.0 techniques (AJAX, etc), but that may take a while. For now, multi-screen submissions may be the way to go because they are relatively easy to code and don't make the form look busier than the cockpit of a Boeing 747, which seems to be Al's threshold :) Ahasuerus 19:22, 15 Dec 2006 (CST)
Yes - the 747 cockpit is my threshold. However, if I can accomplish complex submissions through the judicious use of dials, then I'm there. Alvonruff 19:27, 15 Dec 2006 (CST)

Codes for Foreign currency

I’m looking over the old ISFDB files to better understand the background and found http://www.isfdb.org/isfdb.xls which seems to be part of the old prototyping and says “US prices: $24.00 UK prices: L24.00 Canadian: C$24.00. Australian: A$24.00 Only 1 price please.” Does ISFDB still support currency codes C and A? They are not mentioned on Help:Screen:EditPub. Also, rather than L can people use Unicode price codes such as £ or ¥? Marc Kupper 00:17, 17 Dec 2006 (CST)

The first question is for Al, I think, unless Ahasuerus knows the answer. I hope the answer to the second question is yes, because I've been using £. Mike Christie (talk) 00:34, 17 Dec 2006 (CST)
IIRC, the answer to the first question is "Yes, currency codes like C and A are still supported". The answer to the other question -- again, assuming that IIRC -- is that "L" is the character to be used for the British pound. I suspect that a database conversion (£ -> L) wouldn't be too hard to write, though. Of course, Al is the ultimate authority on these issues and he may have other ideas now that he is wading through what he calls "Unicode hell". Ahasuerus 01:26, 17 Dec 2006 (CST)
Rule number 1 is still one currency only (others can be put in notes if desired). The use of 'L' comes from the old days when there was no unicode support. I use £ now. C$ and A$ are notations also used by Locus. At any rate, people should be able to enter the unicode currency code as desired, and if there are problems, we'll fix em. Alvonruff 07:27, 17 Dec 2006 (CST)

Title date on first entry of a book

Since most of the data entry for books I've done has been for existing titles, I hadn't noticed until now that entering a publication record for a new title won't correctly record the date on the title record. I just entered a copy of Le Guin's "Catwings", and realized when I looked at Long Works that it was showing up as 1998. What would be our recommendation to editors -- please go and update the title record after you've entered it? Or don't worry about it; it will get cleaned up when more editions have been entered? There will be cases where it's not clear what the date of first publication was, so asking them to update the title record might be risky. Mike Christie (talk) 08:22, 17 Dec 2006 (CST)

Well, this goes back to the issue of "objective data vs. subjective data". "Publication date" is more "objective" while "Title date" is more "subjective". For example, in this case you derived the Title date based on what the 23rd printing of the book stated, but that date (1988) may have easily been the date of first magazine publication or the copyright date, which is not what we want.
We could conceivably add a new field to the "New" form and call it "First date of book publication if known and different from this Publication's date", but that would go against the spirit of limiting the "New [Publication]" forms to objective data. If we open this can of worms, then we might as well add Series data and other Title-specific fields to this form. Not that I am necessarily morally opposed to this change of direction -- fewer submissions mean less work for moderators, faster turnaround and less overall frustration while the potential harm is not that great (IMO) since editors will enter the same Title data via the Title edit form anyway -- but first you would have to convince Al, who still has nightmares from the last time :)
As long as we are sticking to the current approach, I think we should tell our editors that the preferred way of handling this issue is to find the true first edition (using secondary sources) and enter it in the ISFDB. If that is impossible for some reason, then they should do what Mike did and enter the source of information in the Notes field.
P.S. We will likely need a Python/Perl script to find all Titles whose date doesn't match the first Publication date for the Title. I'll add it to one of our Projects. Ahasuerus 14:57, 17 Dec 2006 (CST)
As it is – I was going to propose a feature request pretty much exactly in line with what Ahasuerus suggested where the new-publication form would have title related fields so that the extra-fixup of the title record is not needed. Could you chew a bit of Al’s dreamgum and explain what the nightmares were about? Marc Kupper 02:27, 19 Dec 2006 (CST)
One of the problems that ISFDB1 had back when it accepted user submissions was that editors submitted unreliable and unverified data. Think "Gee, I am pretty sure I have seen a Simak collection not listed in the ISFDB and the title was... um, City and the Stars? I guess I'll add it real quick" and you'll see why Al is unwilling to allow editors to enter anything but "objective data" from the book itself, at least not on the first pass. Ahasuerus 17:18, 21 Dec 2006 (CST)
This is also an area that has been bugging me in that the db field name is title_copyright but the help files are being edited to move away from describing it as the copyright and towards it being the “first published” date. What triggered this change? Marc Kupper 02:27, 19 Dec 2006 (CST)
I believe that the original intent of the field was to capture the date of the first publication. I suspect that at the time the field was labeled, Al didn't realize just how often "first publication date" is not the same as the "copyright date", hence the subsequent change. Ahasuerus 17:18, 21 Dec 2006 (CST)
Related to this is I did not think the title’s “first published” date should be something editors need to worry about at all but rather it can be computed from the publication records. Marc Kupper 02:27, 19 Dec 2006 (CST)
Well, ideally, we would list the first Publication for all of our Titles, at which point deriving the Title date from Publication data would become trivial. Unfortunately, there are cases when the only source of first edition information is a secondary source and, what is worse, some of them, e.g. encyclopedias, do not list any Publication data except for the Publication year. Given these harsh realities, I think it makes sense to have an editable Title-specific date field. As the data in the database improves over time and we have more and more editions on file, we may reach a point where we have no (or very few) of these cases, and then we can make the field non-editable, derived, or what have you. I'll make sure that we have a script looking for Title-Publication date discrepancies added to our Project area so that we could keep an eye on this. Ahasuerus 17:18, 21 Dec 2006 (CST)
Maybe we need a “Post Beta Feature Request” page but I personally would like to see a “First published” field added to the publication record with the caveat added that it be the first published for that publisher (or maybe imprint). For example, if the printing history in your copy of a book says “G.P. Putnam’s Sons 1969 / Berkley Medallion Sep-1971 / Nineteenth printing May-1981” then “First Published” for your Berkley edition’s publication record would be “1971-09-00” and “Year” (which should be renamed to “Published”) would be “1981-05-00”. If there is not a publication record already for the G.P. Putnam’s Sons edition then the editor should add one using the information available (put something in the notes so that others know it’s a placeholder and someone with a copy of that first edition can fill in the details) so that the computed “First Published” will be correct.
Unlike “Published” the “First Published” date is available in most paperback books and is objective data that can be copied into ISFDB with little guesswork meaning that even if someone has a later printing where Published is n.d. we will still be able to extract useful printing history data via First Published. Marc Kupper 02:27, 19 Dec 2006 (CST)
Sadly, "First Published" is not always "objective data" either :( I have run into quite a few cases where "First Published" meant "First Published in the US" or "First Published in the UK" or even "First Published by this Publisher since we don't care about other, inferior publishers". But at least it means that there was a prior edition and we could probably encourage our editors to capture that edition's data as well. Ahasuerus 17:18, 21 Dec 2006 (CST)

Novel/short story

I noticed that "The Word for World is Forest" has two title entries: one as a novel: 32494, and one as short fiction: 40773. I believe the text is essentially the same in both. Am I right in thinking that these titles should be merged? It looks like a long-standing division, so I'm checking here in case I'm missing something, but as far as I can tell the title is a NOVEL and should be recorded that way in all publications. Mike Christie (talk) 08:22, 17 Dec 2006 (CST)

It's a common problem, especially with older books before the bloat set in (think the Ace Doubles, "complete novels" published in the pulps, etc). What seemed to be a bona fide novel 40-50 years ago can (and will) be reprinted as a novella in 2006 in one of those monster "Best XYZ of the 20th century" or what have you volumes. I tend to think that something that was published and widely seen as a "novel" at the time of its first appearance should remain a "novel" forever even if our standards change, but I don't recall if it has been codified anywhere in the Wiki. Regardless, the two titles should be definitely merged. Ahasuerus 15:06, 17 Dec 2006 (CST)
Done. Mike Christie (talk) 16:25, 17 Dec 2006 (CST)

Updating incorrect titles

Marc asked a while back how to approach old data that seemed wrong. I have decided to have a crack at the Le Guin biblio, since I have a fair amount of her work, and I ran into an example of this. There is a Puffin edition of "A Wizard of Earthsea", listed here, which is clearly misdated as 1968 -- I have a 1975 Puffin and it's priced at £0.45, whereas this one is at £1.95. I think in this case I will update the date to be "0000-00-00" and put a note on the Le Guin biblio page explaining what I did. The alternative is just to update it with the data from my copy, but it does record the existence of a £1.95 copy, so I don't want to get rid of it. Mike Christie (talk) 12:45, 18 Dec 2006 (CST)

As it's an issue regarding a single publication I went ahead and put a note in the publication’s bibliographic record that detailing why you believe the date is wrong as people are more likely to see the publication record rather than the author’s bibliographic notes. Marc Kupper 19:08, 18 Dec 2006 (CST)

Titles as descriptions of works

In the past have there been conversations about having title records have a table of contents to list the child works for omnibuses, collections, anthologies, magazines, etc?

With standard single story works the title record gives the appearance of describing a work and the publications are physical instances of that work. To help reinforce this model ISFDB already has support for variant titles. Where this breaks down are the titles that contain more than one piece of content and the work description is instead shifted to the publication level and duplicated for each publication under a title.

I keep forgetting that title records don’t describe works and will have an omnibus in hand, pull up the title record, see that the child-works are not listed, and start wondering how to add them. I almost filed a bug/feature request yesterday asking how to convert a title record from the old style version that does not have contents to the current one that does before I accidentally noticed the work descriptions are hidden at the publication level and then remembered that ISFDB defines/describes works at the publication level.

Thus I’m wondering what the ramifications are if titles were allowed to list contents and that a publication would be used to “fill in the blanks” on things like the cover artist, page numbers, etc. Marc Kupper 14:19, 21 Dec 2006 (CST)

Followup - After I wrote the above I realized that part of the problem is the title records are doing double duty to describe both individual works which can appear in many different formats and to act as a placeholder for, but not description of, collections of works which are also sometimes included in other works (collections/anthlogies being included in an omnibus or dos-a-dos for example). Marc Kupper 15:46, 21 Dec 2006 (CST)
I am afraid this is a known and thorny problem. On the one hand, there are many Titles (typically novels, but also some collections and anthologies) that go through endless editions without a single word being changed. You could certainly argue that in their cases we would want to have the table of contents at the Title level and not at the Publication level. On the other hand, there are also numerous Titles whose ESSAY content (introductions, prefaces, etc) changes from edition to edition. For example, the first edition of John Myers Myers's Silverlock had no introductory material while the 1979 Ace reptint contained 3 (!) introductions by Niven, Anderson and Pournelle. If the Contents data was at the Title level, then the only way to capture these differences would be by creating a new Title record every time ESSAY content changed. After mulling it over, we decided that it would be too confusing and pushed Contents information down to the Publication level. (Of course, if fiction content changes, then we create a Variant Title).
Having said that, there is something to be said for a compromise solution: list fiction content at the Title level and ESSAY content at the Publication level. However, that would split content information in an arguably non-intuitive way and possible confuse our users and editors. Ahasuerus 23:52, 21 Dec 2006 (CST)

Should HTML tags be allowed/encouraged in the ISFDB notes fields?

I just added a rather long note to a publication and formatted it using HTML <ul> and <li> tags. This worked well but I saw nothing in the help that indicated of this would be a good or bad practice. Marc Kupper 18:25, 21 Dec 2006 (CST)

Since we don't support structured text like the wiki, it seems to be okay with me. Alvonruff 20:30, 21 Dec 2006 (CST)
Thank you - I'm not sure if I want to advertise it via the help pages though it can help a lot in making the longer notes readable. For example, in the first edit I forgot the closing </ul> tag meaning my <ul> list sucked up the "Bibliographic Notes" <li> that you output after the notes. Marc Kupper 21:08, 21 Dec 2006 (CST)
We may want to edit it to destruction before we advertise it :) Ahasuerus 21:19, 21 Dec 2006 (CST)
Yes. The second sentence to my reply was going to be "I'm sure I'll regret saying this." Alvonruff 06:26, 22 Dec 2006 (CST)

Wiki editing problems

Please note that editing a secrion of the Community Portal page may result in strange behavior: an attempt to save the modified section will bring up the Preview page and when you click the Save button enough times (?), it will file the section but delete the rest of the page contents. We may want to archive some earlier discussions to see if helps. Ahasuerus 17:24, 21 Dec 2006 (CST)

Did you get an editing conflict message? I've gotten those if someone's editing the page at the same time as I am. To reduce the odds of conflicts I normally compose replies off line and than to a fast edit, paste, preview, save cycle. Marc Kupper 18:27, 21 Dec 2006 (CST)
Nope, no conflicts today. Seems to be back to normal now, so I'll just write it off as an occasional Wiki weirdness. Ahasuerus 21:03, 21 Dec 2006 (CST)

Title referencs

What does it mean when a publication has no "title reference" displayed? For example, this title shows five pubs; this one has no title reference displayed, but this one does. Overall, two of the five have a title reference record. What's going on in the database that's different between these? Mike Christie (talk) 16:42, 22 Dec 2006 (CST)

I recall something about if the title is a publication that it will not show in the contents list. The first is a title (flagged as a collection) and the second is one of it’s child publications and it does not show the title (nor do any of the children) as the title is a collection. Marc Kupper 17:29, 22 Dec 2006 (CST)
correction - As you noted - one of the five publications does show the title. Thought #2 is someone did a remove-titles from the other four pubs to clean things up. It does bring up that ISFDB does not attomagically include the link to the parent title. Marc Kupper 17:33, 22 Dec 2006 (CST)
The thing to keep in mind is that there isn't a clear-cut link to the "parent title" (there's no bibliographic requirement for such a link - the Title Reference is supplied to help aid our work). In this case we have 5 pubs, and 14 titles that may or may not refer to all of those pubs (9 stories, and 5 other titles that should be collections, but in this case aren't). There is no specific "parent title" link.
The previous algorithm for determining the "title reference" was fuzzy, and could sometimes pick the wrong title. The new algorithm is quite precise, and will not pick the wrong title, at the expense of sometimes not picking a title at all. As such, I put in some clues to help track down problems like this one. If you look at the title, you will notice the line that says "Type: COLLECTION". Now scan the 5 publications attached to this title, and look at the last attribute in each line. Two of them have "coll" (collection) in their attribute list. If you click on either of those two "coll" publications, you will find that they DO have a Title Reference. The other three publications are of type NOVEL, and if you check them you will see that they DON'T have a title reference.
If you look at the 2nd publication, the system has 9 titles to chose from as a potential Title Reference. 8 of them are short stories, and one of them is the collection title itself, which is what we want to be the Title Reference. The problem is that this publication is of type NOVEL, and it is looking for a NOVEL title match. Not finding a match, it prints nothing. The fix is to change the types of the 3 errant pubs from NOVELS to COLLECTIONS, and they will all have Title References.
Bottom line here is: if a title reference is missing, check the type. We've already tried a guessing heuristic, and didn't care for the results. We can come up with some other tools that help put the pubs into the right type, or make type mismatches more explicit. Alvonruff 17:58, 22 Dec 2006 (CST)
I've put a new section at the bottom of the title bibliography called "Bibliographic Warnings". It only has one heuristic at present; we can add more as time goes on. This should call attention to problems that need to be addressed with a title. Alvonruff 18:11, 22 Dec 2006 (CST)
I really like this and have already used it when looking up a title to see that one of the publications unrelated to the one I was verifying needed to be fixed. One request is that when there are no bibliographic errors to drop the section entirely rather than saying “None” and it may help to highlight the “Bibliographic Warnings:” header with red so that someone will immediately notice it. Marc Kupper 21:49, 22 Dec 2006 (CST)
I think I follow the explanation. Here's a restatement; can you tell me if this is right? Each pub has an unambiguous list of content titles, linked in the db via pub_content. That is always intended to include a title record that refers to the whole pub, whereas the other title records will refer to partial contents such as art and shortfiction. This is the "parent" title record. It doesn't have to be unique to this pub, because of title merges, so multiple pubs can point to one title. However, there is no marker in pub_content that says this title record is the parent. So when scanning contents, if there's a title that matches the pub author, pub title, and pub type, it's assumed to be the parent title and treated as such. That includes displaying it for editing in the editpub screen, and displaying it as a title reference link in the pub display. Is that correct? Mike Christie (talk) 07:15, 23 Dec 2006 (CST)

(undent) Last note on this... I've added three heuristics so far to the title.cgi app:

  1. If the title is a COLLECTION, and the pub is not a COLLECTION nor an OMNIBUS, a warning is printed.
  2. If the title is an ANTHOLOGY, and the pub is not an ANTHOLOGY nor an OMNIBUS, a warning is printed.
  3. If the title is a COLLECTION, ANTHOLOGY, NOVEL and the author list of the title doesn't match the author list of the publication, a warning is printed. This rule will need some sprucing up to handle pseudonyms, so watch those. Alvonruff 18:59, 22 Dec 2006 (CST)
You added an ISFDB:Missing_Titles page though I don’t understand the db structure well enough to follow the code vs. your explanation above. I grabbed a title from near the bottom of the list; No title: 69322 (City of Illusions), and did a title search on it to find a title record. My complication/question is the missing-title list shows publication by a numeric ID while the title page links to them using that stripping of vowels and adding the year thing meaning I needed to inspect all of the pub records to see if it looked the same as the “missing” one. One record was close (different price/binding). The question becomes – how do you link the pub back to the title? Marc Kupper 00:22, 23 Dec 2006 (CST)
This may be a bug report but the title The Secret Galactics reports under Bibliographic Warnings:
I believe this should not get reported as an error/warning as Science Fiction Special 23 is an anthology (the editor happens to be anonymous) and The Secret Galactics is just one of the stories within it. I believe your consistency checks should allow that when type: NOVEL titles reference type: ANTHOLOGY publications that the author names may not match. Marc Kupper 19:23, 23 Dec 2006 (CST)
Actually, it's an omnibus edition of two novels, but changing the book from ANTHOLOGY to OMNIBUS didn't make the warning disappear. Ahasuerus 19:31, 23 Dec 2006 (CST)
I suspect this needs to go on a page Al pays more attention to but I entered anthology, the title page looked, ok, I did a minor change to the pub-record, and the previously reported issue about ISFDB flipping author names cropped up and now the title page reports
Bibliographic Warnings:
  • Author Mismatch (Pub=Rosalind M. Greenberg, Martin H. Greenberg, should be Martin H. Greenberg, Rosalind M. Greenberg): Phantoms Marc Kupper 17:14, 26 Dec 2006 (CST)
Well, since there is no way to enforce co-author (or co-editor) order in the ISFDB, this looks like a pretty obvious bug. I'll cut and paste it to the bug section, thanks! Ahasuerus 17:19, 26 Dec 2006 (CST)

Unapersson

I was doing some cleanup and noticed that for a while “Unapersson” was doing a bunch of unmerges/merges that were getting approved though now they are starting to queue up. My question is – is this a person and who normally approves his/her/it’s submissions (which are coming in pretty fast)? Marc Kupper 17:35, 22 Dec 2006 (CST)

They're our first editor who is not a moderator. Go ahead and approve anything that looks OK; I've been doing a fair amount of that today. Mike Christie (talk) 17:48, 22 Dec 2006 (CST)
"Their" name is Ian -- see ISFDB:Beta recruitment and the Web site that he has linked :) Ahasuerus 23:20, 22 Dec 2006 (CST)

Statement of projects

While cleaning up a series I was on the mod-queue and approved a new-pub for User:Pagadan but then did a double take and hunted down the title & pub to learn it’s seems to be free web based non-genre children’s story by an unknown author though it eventually links back to Joy V. Smith. It appears she’s adding many web-base stories. I personally don’t have a problem with that but have a couple of ideas

  1. That editors put on their wiki home page a statement of purpose on what areas of ISFDB they are likely to be updating. Some editors will be adding/verifying books from their personal collection and it’s quite likely all of those will be sf as long as the editors understand this is sf site. Other editors may be editing/updating all of the works for an author, series, etc. where it’s possible there will be occasional non-genre material.
  2. In the moderator queue that the submitter name be a link to the user’s page so that moderators unfamiliar with a particular editor can quickly get an understanding of what they may be trying to do. The Recent Edits list already has such a link.
  3. That the Subject field of the Recent Edits list be a link to the results or at least link to the search page to make it easy to get back to the result. Marc Kupper 20:13, 22 Dec 2006 (CST)

Marc,

My children's story, Sammy's Beautiful Tail (non-genre unless you consider it a fantasy)isn't exactly free or necessarily web-based, though it's available at a website as a download or on a CD. (Childrenzbooks is complicated and changes...) The CD (Childrenzbooks Volume Two; it's a collection of stories) has an ISBN: 0-9748989-1-0. Feel free to do with it as you choose.

Cataloging Old Pulps

I am back at the secure undisclosed location where my library lives. The place is a mess: the graviplane needs new ditherium crystals, the force field is sagging, the ansible needs some TLC, etc. Still, I'll try to enter as many old and unusual pulps as I can over the next 9 days or so, possibly at the expense of other moderatorial activities. Take a look at the first submission, the February 1941 issue of Strange Stories. Note the problems listed on the associated Bibliographic Notes page. Keep in mind that obscure pulps => obscure publications => obscure attribution problems :( Ahasuerus 00:20, 23 Dec 2006 (CST)

It looks good and I'm glad you are still able to tunnel through to this universe from time to time. Are you doing image scans too of the cover and perhaps table of contents too? Marc Kupper 00:29, 23 Dec 2006 (CST)
Not at this time since these pulps are post-1923 and I don't know what the Paratime Police may have to say about it. I can always go back and do the scans, of course, assuming I can take a break from my wanderings and spend some time with my collection before I get reassigned to the next, more ethereal universe. Ahasuerus 00:57, 23 Dec 2006 (CST)
Another one is in and it comes with a bag of problems all its own. Take a look and see what you think re: "semi-fictions" and unreadable signatures. Is there anything special that we want to do about letter columns aside from marking them as such in their Titles' Note fields? Ahasuerus 12:02, 23 Dec 2006 (CST)
I think you've done the right thing on all the dubious choices. I'd suggest leaving the semi-fictions as ESSAYs, as you have done. Mike Christie (talk) 12:08, 23 Dec 2006 (CST)
As an aside, I found a folded piece of paper within the 1932 Strange Tales issue. It turned out to be a handwritten budget for a 1936 Oklahoma fanzine co-edited by Jack Speer. I don't think he will mind if I disclose that at that point their total expenses (paper, hecto fluid, etc) were $1.00 and their total income from subscriptions and advertising was $1.05. Who says you can't get rich by publishing fanzines?! Ahasuerus 12:41, 23 Dec 2006 (CST)

Should rejected edits be visible to at least moderators?

I was tired and when I saw all the red on the mod screen I mistakenly rejected an author update. A few minutes later I got to thinking and it took a bit of jiggling to locate the previous submission and to get it displayed so that I could decide on what to do. I’m thinking it would be handy to have the rejected edits visible – I don’t think there’s a need to add a way to resubmit them at the moment as viewing should be enough. Marc Kupper 03:15, 23 Dec 2006 (CST)

There is now a "Recent Rejects" option in the moderator navbar. Alvonruff 06:22, 23 Dec 2006 (CST)
Thanks! Is that big chunk of empty space up top intentional or some kind of fluke? Ahasuerus 12:43, 23 Dec 2006 (CST)
Perfect and thank you Al! I suspect the “Approver” column header should be renamed to “Rejector” or “Moderator.” Thank you for calling what looks like “Recent Edits” page “Recent Integrations” as I had been wondering what status code “I” could mean. Would the “Reason” string for Integrations ever have anything in it? Marc Kupper 17:24, 23 Dec 2006 (CST)

Observations from the first 60 hours of beta

Firstly, we need a "Hold" checkmark box in the Moderator screen and we need it badly. Some submissions are clearly well-meaning, but so badly messed up that a moderator would need additional information to decide what to do about them. If one moderator sends an e-mail to the editor or leaves a message on the Talk page, you don't want other moderators to be doing the same thing or, even worse, approving a submission that is still "in progress".

Yep - hold with comments would be great. I just approved a User:Tjacksonking submission wondering if anyone was around that was already dealing with it and so added a thing to the user's talk session about the item. Marc Kupper 03:34, 25 Dec 2006 (CST)

Secondly, one thing that has come up during a discussion with User:Tjacksonking on his Talk page is that authors may (not entirely surprisingly) be interested in their own stories, but not necessarily in all those other stories that appeared alongside theirs in magazines/anthologies. Our current model emphasizes entering complete information about physical publications, but that's not necessarily the emphasis of individual authors and we may have a bit of a conflict here. Not sure what to do about it, but it's food for thought. Ahasuerus 00:09, 25 Dec 2006 (CST)

I've drafted a hold feature definition at Feature:90068 Add hold flag to submissions; it doesn't currently include a comments feature. I'm not sure we need one -- after all, the hold is not visible to non-moderators, and the other moderators don't need to know why something is on hold, do they? Mike Christie (talk) 05:39, 25 Dec 2006 (CST)
The specs look good; I don't think it's vital to have a free text field explaining why something is on hold, although I can see how it might be useful in some cases, e.g. "This submission may be on hold for a while since I am trying to communicate with the submitter who is not responding". Ahasuerus 10:28, 25 Dec 2006 (CST)

Voting

I see that voting showed up. A couple of observations

  • The default is "1 - Extremely poor. One of the worst titles ever written. Couldn't be any worse." I suspect the default should be "0 - no vote."
I've modeled the scale on the IMDB scale, as that's a popular, well-known voting system. It scales from 1 to 10.
I was not commenting on the scale but rather the default button value. IMDB defaults the vote button to "Your Vote" and not "1 (awful)."
  • When you click on the select button the order shown is 1, 2, 3, 4, 5, ... 10 which is opposite of voting scale which is 10, 9, 8, 7, ... Can the order in the button select get reversed? (it reminded me of a local gas stations' pumps that has "No / Yes" buttons just below the display and after a recent software upgrade the pump asks questions like "Debit card Yes / No?" and you both need to mentally translate what that means for a credit card and mentally swap the yes/no that's just above the No/yes buttons.)
I'll think about the reverse the scale order. At present it's identical to the IMDB.
I took a look at IMDB - one difference is they don't have an explanation of the scale. With ISFDB someone is likely to be looking at the scale at the same time as they vote while they decide which score to give the title. Another option is instead of a drop-down list is to radio button the explanation list. Marc Kupper 19:08, 26 Dec 2006 (CST)
  • After you vote you are not thanked nor taken back to the title record. I suspect most people are not interested in knowing "insert into votes(title_id, user_id, rating) ..." Marc Kupper 03:31, 25 Dec 2006 (CST)
Fixed. Alvonruff 18:33, 26 Dec 2006 (CST)

Italics in titles

As an experiment I included italics tags in this title, which shows there, and which also shows here, but which doesn't show here (because it's already italicized. Is this something we should support? I think it could be handy, but could also be handled by a note on the title record (which I'm also going to add). Mike Christie (talk) 11:45, 25 Dec 2006 (CST)

One problem that I just realized this causes is that a search for "lady was" doesn't find the title. You have to search for "lady</i> was" All by itself this probably makes italics undesirable. There's only one other record out there with an italics tag.Mike Christie (talk) 11:50, 25 Dec 2006 (CST)
I have been using the Wiki format ('') for italics in the Note field with some success, but it doesn't sound like it's worth the aggravation in searchable fields like Titles. Ahasuerus 12:00, 25 Dec 2006 (CST)

Pagadan and ezines

Pagadan has been entering ezines -- see the subs list. I assume we don't take these and have left her a talk page message to say so, but was unwilling to reject them till I got confirmation. Mike Christie (talk) 20:38, 25 Dec 2006 (CST)

Thanks for taking care of this, Mike! I was trying to puzzle it out myself when you posted on her page :) I think that the standard that you spelled out, i.e. that there should be a very good chance that the URLs will still be around in a few years, is a reasonable one. BTW, remember our discussion re: the kinds of contributors that we may expect? For now it looks like the profile hasn't changed much vis a vis what we had earlier. We'll see what happens once the word gets out. Ahasuerus 21:10, 25 Dec 2006 (CST)
Thank you Mike - Yes, I saw that they were eZines too which is what triggered ISFDB:Community_Portal#Statement_of_projects as I wanted to get a better understanding of what is being catalogued and if it's something that fits within the ISFDB mission. Marc Kupper 13:21, 26 Dec 2006 (CST)
Yes, I'd forgotten that you'd posted that note. My feeling is that these should not be in the ISFDB, though -- I was thinking some more about it, and I just think that it'll be impossible for us to draw a firm line between web-published professional material and vanity material. I am aware this is going to lead us to omit stuff, but I don't see a way around it. The web is not a durable medium. So I'd suggest we delete the material she is adding, or at least not accept any more of the new publications. I also think we should update the rules of acquisition to say that web material is not eligible for entry at the moment. Mike Christie (talk) 16:11, 26 Dec 2006 (CST)
In thinking about it I keep changing my mind on if these should be included or not. I believe I now know how the editor of a publishing house feels where you see thousands of corner/edge cases and you need to draw the line somewhere on what eventually gets published. Of course, in their case it’s driven by “can we recoup our investment?” The submission queue finds 144191 which is an interview of Ken Rand by Joy V. Smith. Google finds the article and I see that writers must be quite used to rejection notices meaning they should be ok with ISFDB setting its policy so that it’s an index only of printed material or works derived directly from printed material. Marc Kupper 17:03, 26 Dec 2006 (CST)
With CDs and audiobooks and eBooks added, I'd agree -- basically anything that could or did realistically get an ISBN is legit. I'll wait to hear agreement from Al or Ahasuerus before I reject the one I've got on hold (nice work on implementing that so quickly, Al!), since I had to hit that reject button without consensus on the policy. I'll also hold off on updating the rules of acquisition till I hear consensus. Mike Christie (talk) 19:46, 26 Dec 2006 (CST)
Well, as I wrote above, "I think that the standard that you spelled out, i.e. that there should be a very good chance that the URLs will still be around in a few years, is a reasonable one". In other words, anything ISBNable is in, while things that are not ISBNable need to plead their case and prove that they are "special". For example, if the National Archives decide to make an SF piece by Mark Twain available on the net, we may want to consider linking to their URL since it is likely to stick around, but ezines are generally out. I would post a note on the editor's Talk page explaining what's going on before rejecting the submissions, though. Ahasuerus 19:58, 26 Dec 2006 (CST)
OK, it's a deal. I'll reject the one I have on hold, and leave another note on her talk page. I don't see a need to post the note first; I'll just make sure it's a friendly explanation. I'll also mention that Al is likely to reject the other ones. I'll also update the rules of acquisition. Mike Christie (talk) 20:03, 26 Dec 2006 (CST)
Doesn't Al know Joy from RL and/or prior encounters? Perhaps he would be a better candidate to break the bad news :) Ahasuerus 20:06, 26 Dec 2006 (CST)
Oops. Too late. Al, feel free to overwrite/edit/follow up my note to her. Mike Christie (talk) 20:12, 26 Dec 2006 (CST)

Normal vs. advanced search

I had started to enter an anthology last night and this morning wanted to pick up the record so I could finish it. The title is “Christmas Bestiary” and when I searched for Bestiary using the search box it comes back with 6 matches but not the title I was working on. Advanced search though returns 47 titles including the one I wanted. I was not aware the regular search was so restricted. Is something available that explains what sort of results I should expect from the standard search? Marc Kupper 13:30, 26 Dec 2006 (CST)

A regular search on "Bestiary" returns 47 matches for me. Did you enter "Bestiary " (with a space at the end), by chance? When I search on it, I get 6 matches back. Ahasuerus 13:46, 26 Dec 2006 (CST)
Thank you – I had wondered a while back if trailing spaces mattered… (maybe I should go back to drinking coffee in the morning) Marc Kupper 14:04, 26 Dec 2006 (CST)

Attribution of letter columns

How do we attribute letter columns in magazines? I have been using the following rule: if there is an attribution in the table of contents or at the beginning/end of the column, then I attribute the ESSAY to him; otherwise I enter "various" as the author since sometimes the lettercolumn is just a collection of letters. Or should it be "uncredited"?

Contento uses "The Readers". I used to put "uncredited". Lately I've been putting the editor, as he/she picks the letters to print, and writes answers to them. Newspapers put an Ann Landers byline on her column, even though it's essentially a letters column (for advice). Ask me tomorrow and I'll probably have a different answer. Alvonruff 18:16, 26 Dec 2006 (CST)
I'd suggest "uncredited" unless the letter column actually provides a credit; and then if they just say "The Editor" I'd use that for the attribution rather than the editor's name. Letter editing is one of those things that often did get farmed out to an assistant, so without attribution I think we should be scrupulous about leaving it as "uncredited". "various" would be OK too, but I think "uncredited" is better -- the letter column would never be attributed to the letter writer, after all, even if it just consisted of a single letter. It's the "work" of the editor who selects (and usually responds to) the letters.
I've been doing something similar with the editorials in "Venture"; they're signed "RPM", which is obviously Robert P. Mills. However, I've been entering them as RPM and creating an undated "canonical" title with "Robert P. Mills" as the author, and making them a vt of that. I think this is best as it preserves the actual credit in the magazine.
Makes sense; I'll make the change for Doc Lowndes' column (signed RAWL) in Bizarre Fantasy Tales. Ahasuerus 20:01, 26 Dec 2006 (CST)
A slightly different case came up with the artists. Venture was very inconsistent about crediting; they failed to credit, or credited "Giunta", "J. Giunta", or "John Giunta". When they credited, I used the form they credited, even though I could just as well have put "John Giunta" every time, because "Giunta" is what they put. But when they don't give a credit at all, I used "John Giunta", since that's his name and I could identify his work since it's signed. I just had to add a note at the pub level explaining that it was a deduction from the signature, rather than an attribution in the magazine. Mike Christie (talk) 19:54, 26 Dec 2006 (CST)

Verifying that Bug 20021 is still there

Could someone with a smaller monitor try to recreate Bug 20021, please? (<curses oversize monitors that don't support low resolutions any more>). Ahasuerus 13:25, 27 Dec 2006 (CST)

(duplcated from bug 20021 reponse) Is grendel|khan still around? The reason I ask is that I’m confused by the bug report. When looking at pages I don’t think I have seen ISFDB use any CSS absolute positioning or Z-layering that would cause stuff to get hidden. I viewed the test page at 640x480 and 800x600 with both FireFox and IE and also in a 25x80 text window using Lynx. I did not see anything to that looked like a problem and also did not see anything in the HTML that could potentially cause a problem. Marc Kupper 14:05, 27 Dec 2006 (CST)
Actually, I recall being able to recreate the problem on a smaller screen, so it looks like it has been fixed, especially since you have reviewed the HTML code. Moving to FIXED. Ahasuerus 18:47, 27 Dec 2006 (CST)

Sorting out "David Alexander"s

There are apparently at least 3 people who have written specfic as by David Alexander.

David Alexander number 1 was a fairly prolific mystery writer (aka "David Alexander (1907-1973)") who published at least 18 mysteries in the 1950s and 1960s. He apparently also wrote at least one specfiction story, The Disposal Unit Man, which was published in Window on the Future. However, I don't know if we can be sure that it was the same guy (even though Contento seems to agree), so we may want to double check.

David Alexander number 2 started writing as David M. Alexander in the late 1970s, publishing an associational thriller in 1978 and a genre SF novel in 1981. Then there was a gap; when he returned to the genre in the 1990s and 2000s, his stories were published as by "David Alexander". The only exceptions were two recent reprints done by Hayford Peirce, his co-author: since he didn't realize that the middle initial was no longer being used, the stories were reprinted as by "David M. Alexander", . These days David Alexander number 2 concentrates on non-SF, which he writes as "David Grace". We have his complete biblio on file directly from him (via Hayford Peirce).

David Alexander number 3 contrinuted to a few "men's action adventure" and post-holocaust series ("C.A.D.S.", "Phoenix", "Nomad") in the late 1980s and early 1990s. They were published as by "David Alexander" and as by "John Sievert" (a house name) and we have most (but not all) of them cataloged. He -- or possibly David Alexander #4 -- has also written a series of novel about Special Ops starting with Marine Force One in 2001 as well as similar near future and technothriller novels/non-fiction. We don't have any of the latter books listed, which is probably OK.

Numbers 2 and 3 currently share the same Author record record, which makes "David Alexander" look unusually active for someone who has been dead for 33 years.

The question for the panel is what the heck are we going to do about his motley crew of David Alexanders? My best guess is that:

  1. The original 1950s/1960s David Alexander the mystery guy should be changed to "David Alexander (active circa 1963)" since he is quite obscure in our field and we can't be even sure that he is the same person as the author of that UK story
  2. "David M. Alexander" aka number 2 should keep his middle initial (even though he no longer uses it) to avoid confusion
  3. David Alexander number 3 can remain "David Alexander" unless somebody has a better idea
  4. David Alexander number 4 (assuming he is not the same person as David Alexander #3) doesn't need to concern us unless he commits some specfic in the foreseeable future.

Suggestion welcome. (Sheesh!) Ahasuerus 21:51, 27 Dec 2006 (CST)

Libraries help the disambiguation by including birth/death dates along with the authors name and a check with my local library finds
  • Alexander, David
    • My real name is Lisa
  • Alexander, David, 1943-
    • Star trek creator
  • Alexander, David, 1956-
    • My mysterious world / by Margaret Mahy ; photographs by David Alexander.
My own book list has the mystery guy's (1907-1973) books (I ran across him a few years ago via Hitchcock's "Hangman's Dozen"). His book writing career seems to run from 1951 to 1966.
Or - to get silly
  • One was a mystery and so he's David Bloodstain Alexander.
  • One of them wrote "Star Trek Creator" and so he's David Trekie Alexander.
  • One of them wrote a book called "My Real Name is Lisa" and so let's change that one to David Lisa Alexander. --Marc Kupper 02:08, 28 Dec 2006 (CST)
Well, birth/death years are indeed the industry standard, but, unfortunately, they are not always known. For example, we are not sure whether the 1907-1973 mystery guy was responsible for the 1963 story, which is why I was thinking that "(active circa 1963)" might be a safer thing to say. BTW, "My Real Name Is Lisa" was published by David M. Alexander writing as "David Alexander". The Star Trek guy (David Alexander #5) should be easy, we can call him "David Alexander (1943-)". I'll make the changes shortly, thanks! Ahasuerus 11:31, 28 Dec 2006 (CST)
My own quick research shows the Trekkie Alexander to actually be "David Alexander Smith," sometimes shortened to "David Alexander" -- this is, admittedly, infuriating. Hayford Peirce 21:56, 28 Dec 2006 (CST)
Another issue has come up in our discussions with Hayford Peirce and I suspect that it has wider applicability. In this case we have "David M. Alexander" who also writes as "David Alexander". We also have another "David Alexander" who writes as, well, "David Alexander". (We'll ignore the other David Alexanders for the purposes of this sub-thread). At this point we believe that we have sorted everything out and the respective pages are accurate and as up to date as we can make them.
Now, suppose a user looks up David M. Alexander in the ISFDB and notices that some stories were written "as by David Alexander". He then follows the "David Alexander" link to the "David Alexander" page and doesn't find any of the stories that he just saw as written by "David Alexander". Instead he finds a whole bunch of other novels. Is this desirable behavior or is this confusing? And if this behavior is confusing, what should we do about it? Ahasuerus 19:54, 28 Dec 2006 (CST)
Well, I don't see *how* it could possibly be "desirable". If you go to my page, say, and see a story co-written by me and "David Alexander" and you click on the Alexander link and immediately find yourself at a page of a guy who has written dozens of techno-thrillers, that is just plain wrong and/or confusing, and certainly not the goal, I would say, of bibliographic scholarship or preciseness. I'm not trying to just carp -- I wish I could offer constructive help about how to fix this; but I do think it ought to be fixed. An instrusive, and non-esthetic, solution, I suppose, would be to go back throughout the entire database and find the six (or seven or even more) characters who have *ever* used "David Alexander" as part of their name and then rigorously rename them a la Wikipedia conventions, which will have articles about, and links to, people like [*[Jack Kramer (tennis player)|Jack Kramer]] and [*[Jack Kramer (baseball player)|Jack Kramer]] and [*[Jack Kramer (eminent Victorian winemaker)|Jack Kramer]], etc. etc. This would obviously necessitate a tremendous amount of work at some point, unless bots could be designed to do some of it, so that all of the "David Alexanders" (and other authors of similar names) could be replaced throughout the database.
Update -- Well, it isn't as bad as I thought and as I wrote above: if you go to my page and find a story written by me and Alexander, it *does* say that the co-author is David M. Alexander and that the story appeared as by David Alexander. So the astute reader ought to be able to figure out who is who here -- that the bibliography of DMA is the guy that he should be interested in, not the other one. So maybe the problem isn't as acute as I thought it was. I could, I suppose, be convinced that no real problem actually exists. But even so....Hayford Peirce 22:39, 28 Dec 2006 (CST)
That's kind of what I was thinking. On the plus side, we supply our users with a pointer/URL to the right person's (in this case DMA's) Summary Bibliography page. And then, on top of that, we ever so helpfully tell them that this particular book was published using a byline that is different from what we call the author's "canonical name", be it a variant form of the person's name, a personal pseudonym or a house name. And that's fine. The problem appears to be rooted in the fact that the "as by" name/pseudonym/house name (in this case "David Alexander") is a link to that name/pseudonym/house name's Summary Bibliography, which may be shared by other authors. And since at this point we don't have robust support for "pseudonym bibliographies", the results may not be pretty as we can see in this case.
Given all of the above, is it really useful for us to hyperlink the "as by" name as long as our support for "pseudonym bibliographies" is as rudimentary as it currently is? Wouldn't it be safer to change these "as by" names from URLs to simple text so that our users can't go to places that are not quite ready for prime time? Ahasuerus 22:57, 28 Dec 2006 (CST)
This last suggestion (ie, remove the links on the "as by" names) is so simple that it borders on genius! At least in the DAM/Hayford Peirce short stories case it would seem to work perfectly. A brief thought experiment: Julia Smith, student at Yale, wants to know about the stuff I've written. She finds my biblio article. She sees I've written 3 shorts with someone named DAM -- she clicks on his link and is taken to the David M. Alexander page. She goes back to my page, looks at the same story, Finder's Fee, sees that it was "as by" David Alexander and, just out of curiosity, clicks on *that*. And nothing happens. She remains on the Hayford Peirce page. She scowls and thinks for a moment. "Hmmm, that's funny, I was *sure* there was a David Alexander who writes stuff. Well, lemme do a "Search" for him." So she types "David Alexander" in the "Search" space over on the left side of the screen and is given a list of 3 or 4 or 5 names in various permutations to choose from. A smart Yalie. Okay, I think that this would be a perfectly acceptable way of doing it. The work involved would be to de-link all inappropriate mentions of "David Alexander", as in, to begin with, the Hayford Peirce biblio mentions. Go for it! Hayford Peirce 09:59, 29 Dec 2006 (CST)
A question though: what would the results be for a major writer like John Dickson Carr, who wrote roughly half his stuff as Carter Dickson? I can't think of an SF writer off-hand like that, but he/she probably exists. Would delinking, or relinking, so many references cause enormous problems? Hayford Peirce 09:59, 29 Dec 2006 (CST)
No worries, no relinking would be required. If we decide to do it, Al would have to change a few lines of code not to create the link for pseudonyms and that would be it -- ah, the wonders of modern science! :) The only useful functionality that we would lose (that I can think of) would be the ability to quickly see additional information for house names. For example, if you see a story by Edmond Hamilton wriiten "as by Will Garth", you can click on Will Garth and immediately discover that it was a house name used by Hamilton, Henry Kuttner, Mort Weisinger, and Otto Binder. It may be worth losing this ability, though, until we sort out the issue of pseudonym bibliographies.
Are there enough of them that it would be worth while setting up another category -- House Names -- to go with Name, Title, Series, etc.? Hayford Peirce 14:11, 29 Dec 2006 (CST)
House names can be tricky in all kinds of devious ways. For example, Captain Future "novels" were written mostly by Edmond Hamilton except in a few cases when he was unavailable. Thankfully, we already have fairly robust support for house names as part of the overall "variant title/pseudonym" support. What we do not have at this time is a way to display all works written "as by" that house name. Instead we just have a list of people who have ever written using the house name.
Having said all of the above, I'd really like to hear from Al before getting any deeper into this jungle. He is the only one who has the whole picture in his head and who can clarify what can and can't be done in a reasonable amount of time. Ahasuerus 17:21, 29 Dec 2006 (CST)
An alternative approach to the DMA/David Alexander mess would be to add a big warning (like the one that we have for single person pseudonyms, e.g. David_Grinnell) at the top of "David Alexander"' Summary Bibliography to the effect that "this name has also been used by David_M._Alexander. Food for thought... Ahasuerus 13:03, 29 Dec 2006 (CST)
Yes, that would also probably do the trick. But suppose there are *3* David Alexanders? Which one would be shown at the top as the true name?Hayford Peirce 14:11, 29 Dec 2006 (CST)
Well, the warning would presumably list them all just like the Will Garth page does currently. Ahasuerus 17:21, 29 Dec 2006 (CST)
Generally we link to authors by name but ISFDB does have internal numbers associated with each author. At present it’s #133 is David Alexander, #134 is David M. Alexander, and #44691 is David Alexander (1907-1973). If the system started using author numbers for internal links then all three of those author records could have the same name. If someone asks about “David Alexander” the system will come back with a page saying there is more than one and give a list of records matching “David Alexander” along with details so the person can decide which one they want. Marc Kupper 23:40, 30 Dec 2006 (CST)

(unindent) It seems like what's really going on here is that there are two kinds of name -- names as used on stories in publications, and names that are used to identify writers. We always have the first kind of name if we have the publication in our hand, and the pseudonyms are our attempt to document the relationship of the first kind to the second kind. What we'd like to do, when someone clicks on a "David Alexander" link in Hayford's biblio, is display just the works DMA wrote under that form of his name, and yet still have the ability to have a separate author whose canonical name is "David Alexander". So I wonder if we can represent this distinction in the database.

Exactly! Hayford Peirce 14:11, 29 Dec 2006 (CST)

Al, I was looking through the Database Schema and trying to figure out how an author's name is represented on a title of a shortstory, but couldn't. Where is it kept, exactly? Mike Christie (talk) 13:31, 29 Dec 2006 (CST)

Personal tools