ISFDB:Help desk/archives/archive 01

From ISFDB

Jump to: navigation, search

This is an archive page for the Help Desk. Please do not edit the contents. To start a new discussion, please click here.
This archive includes discussions from March - May 2007.

Archive Quick Links
Archives of old discussions from the Help desk.


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


Expanded archive listing


Contents

How to change a story title in a publication’s contents

Note: This section has been refactored to the Help pages, see its end'

As this happens several times per day I’m going to write up a note and hopefully it’ll reduce the number of publication updates that get put on hold. A secondary project is to figure out a good place to stick this in the general help screens.

In this case someone wanted to update the publication Best Science Fiction Stories of Clifford D. Simak and to change the title of the story Neighbor into Neighbour. If you look at the Neighbor title record you will see that this story is included in ten different publications meaning any changes to the title record will impact how the story is displayed in all ten publications.

Sometimes you want all ten records to get impacted – usually this is for data fields that are “about” the story such as the date, storylen, synopsis, notes, etc. In that case, just make the changes and the odds are a moderator will just review/approve the thing without a second thought.

However, when it comes to data fields that are about the title itself and not the story then the changes should normally be more limited in scope. Ideally, an editor could go to a publication record and just change the spelling of Neighbor to Neighbour to match what stated in the publication and ISFDB would then relink the publication to the title record for Neighbour. Unfortunately, that’s not how ISFDB works at the moment meaning we are faced with a more tortuous path.

Here are the steps to change the title of a story that’s listed in the contents of a publication.

  1. Bring-up the edit-publication page
    1. Click on [Add Title] that’s down at the bottom of the Contents section.
    2. Enter the information for the new title, in this case it would be page 226, Neighbour, 1954-00-00, SHORTFICTION, novelette, Clifford D. Simak.
    3. Click [Submit Data]
    4. Suggestion - If you are updating many titles it's been found to be helpful to set the page number for the old titles that you want to remove to something like, NA, (blank), Remove, Delete, etc. Wait for moderator approval of the publication edit and then proceed to the next step. Note that this is optional and if you are comfortable that you can keep track of the old/new titles then you can just move on to the next step without waiting for moderator approval of the publication edit.
  2. Go back to the original publication record and click “Remove Titles From This Pub” in the navbar.
    1. Look for the old version of the title, in this case “Neighbor, SHORTFICTION , Clifford D. Simak” and click the checkbox for that line.
    2. Click [Submit Data]
  3. Wait for moderator approval of your publication update and remove-title. Once approved you will need to merge the new title record for Neighbour with the existing record. ISFDB has several mechanisms available for this
    1. From the author’s bibliographic display page click “Dup Candidates” in the navbar, locate your story, Neighbour, in this case, set both check boxes, and merge them. If the two titles you want to merge are not spelled exactly the same then try the “fuzzy” mode of “Dup Candidates” (there’s a link at the top of the page for this) and see if that’ll group your titles.
    2. From the author’s bibliographic display page click “Titles” in the navbar, locate your story, Neighbour, in this case, set the for the shortfiction records, and merge them. The advantage of the “titles” method over “Dup Candidates” is that it can be used to merge title records that are dissimilar enough that the “dup” mechanism did not group them. A disadvantage is that “titles” only displays 100 per page and if the merge candidates are on two separate pages then you need to use one of the following method to merge these titles.
    3. From Advanced Search search for Neighbour by Simak, set the appropriate check boxes, and merge them. An advantage of this method is that you can use the and/or mechanism along with partial word scans to construct a search result that has a short list that includes titles you need to merge.
    4. Finally, sometimes the titles are dissimilar enough that you can’t get them grouped onto the same page using the methods outlined above. In that case, edit one of the titles so that its name matches the other (add a note to the moderator in the in title’s notes section), wait for approval, and merge the records using one of the methods listed above.

Before the “Dup Candidates” mechanism was added I used to add the new title to the contents with "(merge)" appended to the story title so that I would later know that the title record was a merge candidate. As the default “Dup Candidates” method uses exact matches by default and is so easy/handy I've switched to entering the new title as it's stated in the publication and “Dup Candidates” will find it. Marc Kupper (talk) 12:57, 6 Mar 2007 (CST)

Nice article Marc! A few comments:
1) It really DOES need to be made clearer how much you're affecting when you make an edit to contents. Particularly if you use "Clone" rather than "Edit" - the former sounds so much safer but it isn't. Hopefully this could just be a small change to add a warning to the edit screen? If so, is there a feature request in? BLongley 15:37, 6 Mar 2007 (CST)
We'll need to see what Al says about this. I have thought about asking for a feature request to make the title, title-type, and author fields entirely uneditible (exactly like clone-publication) but you would still be able to change the storylen and date. That would force people to go to the title record itself to make changes and maybe they will say "whoa" and realize their edit will affect all of the publications listed. It's entirely possible though that someone will fixate on "the title needs to be Neighbour" and just edit away.
A better course may be to allow the contents to be edited pop to add a warning. But, as you see, I wrote a lot of stuff above and you have many questions meaning it's not going to be a simple warning. Marc Kupper (talk) 03:12, 7 Mar 2007 (CST)
2) Sometimes you DO want to update All records - probably when there's only a few (probably one) other that you can also verify. There's no way to add a temporary note saying "Yes, I do mean to change them all, I've checked them all", but maybe there could be a convention to add this to the normal "notes" field on the understanding a Mod will delete it later? (I suspect this is far easier than adding a throwaway "Note to Mod" field that can be read and lost automatically, but I'd quite like that feature in the long-term.) BLongley 15:37, 6 Mar 2007 (CST)
Yes, just this morning I took advantage of the fact that I could go to pub-edit and nail both the publication and title record at the same time. In that case it was a bad data-import that had assigned a story to the wrong author. A note to the moderator may help but some editors already write notes and I often have trouble understanding them. The problem is I'm not in the mind of the editor and so often don't know just exactly what they are doing.
I would guess though the times when you need to edit all of the titles in a publication Contents is pretty rare. Maybe it's a collection where the author's name got misspelled during data entry. I believe though it would be better from a data security view if you simply could not make destructive changes while editing the publication contents. If you are already clicking through to the title to check that it's ok to change it then you might as well make the changes at the title level. Marc Kupper (talk) 03:12, 7 Mar 2007 (CST)
3) I haven't seen the Mod Screen, but does it give any warning about how many records will be affected? And how many of them have been verified by anyone bar the submitter? (For instance, at the moment you can do a title Search for "Robet" and be pretty sure what you find is a typo of "Robot", but as it's verified you might want to check if it's a KNOWN typo first.) If there's no warning, I suspect it makes the Mod work harder. Or maybe not, I'm not sure how confident Mods are about current data versus people-entered changes: as verification levels increase I'd expect people correcting mistakes of unverified data to be trusted more, but we're still pretty low on verification percentage. BLongley 15:37, 6 Mar 2007 (CST)
That's a good idea and I was thinking along the same lines though had not thought about that it could show a count of # of pubs and # of verified pubs. At present the moderator screen only shows the before/after data values but does not offer any hints that a particular submission may be more dangerous than it seems. At the moment there is one moderator-related warning message and that's about people editing publications that are marked as verified. My current "wish list" for the moderator screens are that they at least link to the records being affected so that I can click and take a look at them. Marc Kupper (talk) 03:12, 7 Mar 2007 (CST)
4) In the final article, I'd suggest we don't tell people to do the "remove titles" immediately IF there's a lot of them. It's far easier to mark the old titles as "NA" (Ok, that's my convention, feel free to suggest others) and wait for the first approval, they show up far more easily then. I know I've messed up some clones, particularly when I have to split a large US collection into smaller UK volumes, and it's been far more easy to fix when I allowed the mod to check the first step. As I add page numbers it also makes the next stage easier to check. Unfortunately, most of my publications won't sort their contents pages into alphabetical order - damn paper, it's useless at that sort of thing! ;-) BLongley 15:37, 6 Mar 2007 (CST)
That seems like a good idea. FYI it's a wiki and everyone is free to edit the article itself. :-) Marc Kupper (talk) 03:12, 7 Mar 2007 (CST)
Will do, if the final position of the article is clear! Is it anywhere but here yet? BLongley 16:58, 7 Mar 2007 (CST)
Actually, as I've received NO comments about whether my last update of a help page helped or not, I'm not sure we're reviewing this sort of thing enough. Most of us here don't NEED to read help pages, so won't encounter them naturally. But did anyone check whether I added "If you need to read this help, you're too stupid to work here" or something useful instead? BLongley 16:58, 7 Mar 2007 (CST)
Sorry - I don't patrol the recent-edits page very often meaning I'll miss random edits and edits to the middle of pages such as this one. I have not decided on a final place though I was thinking a standalone page off of Help:ScreenList and maybe linking to it from places like Help:Screen:EditPub. I agree that the general help pages are getting long. Marc Kupper (talk) 04:00, 8 Mar 2007 (CST)
In this case I added a suggestion as note #4 in part 1. My own convention has been to blank out the page number of titles targeted for removal because it cause the titles to jump around on the remove-title screen. Another idea is to set the page number to something like "Delete" or "Remove" and the edit-pub thing would automatically remove the title without us needing to make a separate trip to the remove-titles page. Marc Kupper (talk) 03:12, 7 Mar 2007 (CST)
The shorter the better, I feel. Some Mods are used to my "NA" convention now, and do the deletions for me. We do need a quick-entry option that can not be misunderstood, and preferably one that can be automated later. BLongley 16:58, 7 Mar 2007 (CST)
5) Do please add guidelines on "Make variant" for when it hasn't already been done. There ISN'T always one there already - I'm adding British Spelling variants as I find them, for instance, but I only have two or three thousand books so will not make much of an impact when you consider how many words used in titles WILL vary. (E.g. "Color" gets over four times as many matches as "Colour" does - but does anyone really believe 80% of US titles do NOT get published in Great Britain?) BLongley 15:37, 6 Mar 2007 (CST)
That seems like a good idea but I'm not sure how to word it - perhaps as a separate paragraph at the end? The trouble is that most of the time when I've needed to change a publication's contents it's not because of title changes that would result in a variant but rather an anthology was republished with a slightly different story list. I'm thinking what you are talking about is a more general note that would apply to novels as well as collections/anthologies.
Actually, I'm thinking of my own problems (I do a lot of short-fiction work), but the general problem might be that some Editing Help pages are getting HUGE. If all I want to recall is which small words do NOT need to be capitalized when "regularizing" titles, I don't want three screens of Help, I want the three lines I need to look at. BLongley 16:58, 7 Mar 2007 (CST)
Something I’d like to encourage as a regular practice is that if there are statements on the copyright page about “previously published as …” that these get recorded verbatim in the notes along with a note explaining where the statement was found. The idea is to create citable references on which titles should be variant titles and which are expanded/revised from other titles. Doing this might trigger editors into checking to see if ISFDB has the variant titles set up and if not then to add them. Marc Kupper (talk) 03:12, 7 Mar 2007 (CST)
Oh Yes - "Previously Published as" info is one of the better ways to improve our data. Even if it does mean I've bought the same story 8 times... :-/ BLongley 16:58, 7 Mar 2007 (CST)
6) Should we have some example titles of the problems rather than ones that (hopefully) will get sorted out? BLongley 15:37, 6 Mar 2007 (CST)
Yes, we can make an example of you <wink>. Seriously, I’d encourage all editors to set up a sandbox. For example, I have one at Marc_Testing where I play with random ISFDB operations and it’s ok if I screw things up. One hint is to use an unusual word, such as xtest, in the titles so that later you can do a title and/or publication search for xtest to hunt down strays that may have disappeared from your sandbox when you changed the author name. Marc Kupper (talk) 03:12, 7 Mar 2007 (CST)
I meant a PERMANENT (or whatever passes for it here) place for examples that can be quoted in help. Something people don't intend to mess about with any further. BLongley 16:58, 7 Mar 2007 (CST)
Sometimes I've done the examples right in the wiki. The problem is that much of the complicated stuff is "verbs" meaning we need to show before and after snapshots and that gets tricky when explaining something like changing a story title in the contents. Do you have some thoughts on what the permanent display should have? I have thought about coming up with a list of "model authors" which would be existing author bibliographies that demonstrate ISFDB's "best practices" or maybe it should be a "Sample Author" where the story titles and notes are descriptive. I personally learn a lot by experimenting and am always trying things "just to see what happens" which is why I brought up the sandbox. Part of the problem though is we all learn something new every day. Someone will come along and interpret the instructions in a way we had never considered resulting in another corner or edge case to discuss and then update the help files. Marc Kupper (talk) 04:00, 8 Mar 2007 (CST)
Anyway, thanks again Marc. On preview I decided one last suggestion: add the four tildes to each section that can be replied to, and it makes it easier for people to respond to various parts without messing up former comments. Or am i the only one that waffles this much in replies? ;-) BLongley 15:37, 6 Mar 2007 (CST)
Yep – that works well. I still need to figure out how to make my signature shorter. Marc Kupper (talk) 03:12, 7 Mar 2007 (CST)
I have copied Marc's guide to Help:How to change a story in a collection and linked it from the Help:How to page. Please take a look when you have a chance and polish it at your leisure. The polish is on the house! Ahasuerus 18:06, 15 Apr 2007 (CDT)

Suspicious circular pseudonyms

There are more cases, but I noticed it most recently and strongly at Gardner_Dozois. Several of his stories are shown "[as by Gardner Dozois ]" (the space before the closing bracket, or actually CRLF in the source code, sic - apparently procedure DisplayAuthor, or perhaps Python's print itself, includes one, but that's the smallest problem we have). Of course this doesn't make any sense and would be best deleted. However, it occurs me now: shouldn't this be, and wasnt originally, "as by Gardner R. Dozois", who currently has no as-by publication, and isn't the circularity a remnant of some mistaken merge of the two (or perhaps switching the canonical name from one to the other, or some other effect of database migration?) http://contento.best.vwh.net/s86.htm#A1224 seems to take the opposite approach with R. as the standard form (he doesn't give exact form of the name in individual stories, so you have to click on the publications; I found just http://contento.best.vwh.net/t128.htm#A2725 without it and perhaps one or two more).

So, what now? --JVjr 08:28, 12 Mar 2007 (CST)

Right, this is a known problem with Variant Titles pointing to identical Titles. They are a pain to fix manually, but you can do it by removing them from their Publications, then merging them with their clones and then putting them back in their Publications. I just fixed one of the Dozois stories that way and it seemed to work fine. We should probably write a database cleanup script, though. Ahasuerus 11:53, 12 Mar 2007 (CDT)
I seem to recall approving an author merge of Gardner Dozois and Gardner R. Dozois (or maybe something similar). I had looked and the secondary name (Gardner R. Dozois) had two titles each with one publication each of which had cover images that clearly showed Gardner Dozois. This might be worth testing as it's possible if variant titles had existed that they are now circular after the merge. I know that author merges do next to nothing terms of SQL operations meaning it may not be checking for and fixing pseudonyms.
It seems the easier fix is to remove the VT relationship by setting the parent # to 0 or am I getting confused on what's meant by a circular pseudonym? For example, are these circular pseudonyms or circular variant titles? Marc Kupper (talk) 19:49, 13 Mar 2007 (CDT)

By "circular pseudonym" I meaned that a title is listed as pseudoymous, but the "as by" name is the same as canonical one: in Dozois's case, see frex Conditioned Reflex, records 89586 and 192086.

IIRC my previous dealing with those (see User talk:JVjr#Merging novel with shortfiction, removal of such pseudonym is easier than the process Ahasuerus describes: you just merge the two title entries, ignore the "WARNING: records X and Y both appear in the publication ..." and set the title_parent to none: the story remains where it was and just as it was, only loses the annoying child.

However, my point was whether such cases should be merged on sight rather than investigated, for obviously nobody would create such duplicates manually, so I think they must have originated in some non-standard way from a pseudonym that was originally different. I'm pretty sure that Dozois did publish a few things as Gardner R.; after all, couldn't somebody try to verify some of the titles concerned? --JVjr 04:54, 14 Mar 2007 (CDT)

Given that it seems like an author merge that should not have happened the best course was to change the author names for the variant records to Gardner R. Dozois. I verified the titles against Locus/Contendo and some of them should have been Gardner R. Dozois but others said Gardner Dozois. Only one publication has been verified and so I dropped a note on the editor's page to have the name rechecked. Marc Kupper (talk) 02:40, 15 Mar 2007 (CDT)

Star Wars mess

See http://www.isfdb.org/cgi-bin/pe.cgi?207 : There is an empty sub-series New Jedi Order, and TWO different subseries of the same name "Star Wars: The New Jedi Order" http://www.isfdb.org/cgi-bin/pe.cgi?7672 and http://www.isfdb.org/cgi-bin/pe.cgi?7672 which thus partially contain the same titles. (And that may not be all; I lost my way in it.) I'm afraid it will take an intervention directly in the DB; there isn't any WYSIWYG/web tool for deleting series, right? --JVjr 08:28, 12 Mar 2007 (CST)

There are a few different issues here. The first one is related to having two series with the same name, which confuses the software. The fix is to rename one of the Series, which is what I did in this case a few minutes ago and everything looks better now. The second issue has to do with apostrophes in Series names, which can cause the software to create duplicate Series records since it doesn't realize that the newly entered Series name is the same as an existing one. The third issue is that, as Jan points above, empty Series records are not deleted automatically the way Title-less Author records are. The reason is that some nested Series records are just placeholders for other Series records, so automatic deletion of seemingly empty series would cause problems. Al is still working on this part of the puzzle. Ahasuerus 11:44, 12 Mar 2007 (CDT)

Sleepless Nights in the Procrustean Bed

I'm trying to make this publication link to this title. At the moment it is linked to one of the publication's content titles. Is there an easier way to do this than to delete and re-enter the contents information? Thanks. Mhhutchins 15:45, 15 Mar 2007 (CDT)

A publication normally has one or more titles title records associated with it. When it displays the publication ISFDB goes down the list of titles and looks for one where the publication type (NONFICTION in this case) matches the title type. The first one it finds becomes the "parent title" which is displayed as the "title reference". Normally ISFDB also does not show this title in the contents list though it's visible in edit-title mode. In this case there is no "title reference" listed meaning I know there are no NONFICTION titles.
The fix is easy in that from pub-edit you add a new title, it can be named anything but let's use Sleepless Nights in the Procrustean Bed, of type NONFICTION. Once approved you will have the title reference. You then merge this with the existing Sleepless Nights in the Procrustean Bed title record. (note - I went ahead with this to make sure the script was correct and mid-way through realized it's also painful to delete the parent-title so that you could do it yourself. You delete the parent title by changing its title type so that it does not match the publication. At that point it becomes visible (and the title reference goes away) meaning you can then do remove-title and then delete-title on it.) Marc Kupper (talk) 16:27, 15 Mar 2007 (CDT)
Thanks, Marc! Sounds simple when you explain it. Much appreciation. Mhhutchins 17:44, 15 Mar 2007 (CDT)

Reviews for a misspelled author name

It turns out that if all an author has is book reviews that it’s rather hard to detect this in ISFDB. I had an Alastair_Reynolds book and did a search for Reynolds. ISFDB has a record for this name but also has one for Alistair_Reynolds which turned out to be empty. An ISFDB search turned up zero awards, titles, and publications but Google found the reason there’s an author record is four book reviews

I have no idea why four independent reviews of the same book were also the only things that misspelled the author’s name . You’d think the publisher’s preview copy had it wrong but two of the reviews are in the year 2000 and the other two in 2002.

I tried to get these reviews to show up on Alistair_Reynolds’s page without success, first by adding a dummy novel and then by renaming the novel to Revelation Space. Does anyone know of a trick to do this?

A secondary issue is what's the general procedure if a review misspells an author's name? I can't use the variant title mechanism to map these over to Alastair Reynolds. I suspect one option is to change correct the author name in ISFDB and to put a note in the notes field of the publications that contained the reviews. I’d add a note to the review’s title records too but those can’t be edited directly. Marc Kupper (talk) 16:12, 15 Mar 2007 (CDT)

Marc, I checked my copy of the March 2000 issue of Locus, where two reviews appeared. Both indicate that Alastair Reynolds is the author's name. Must have been an input error. Mhhutchins 20:16, 15 Mar 2007 (CDT)
Thank you - I updated ISFDB plus added publication notes. Marc Kupper (talk) 17:34, 17 Mar 2007 (CDT)

Stray publication: how to herd it with the title?

See RBLSTR2001: it correctly lists the author Shariann_Lewitt but lacks a Title Reference to her Rebel Sutra with 2 listed publications, hc RBLSTR2000 and pb RBLSTRMNRW2001 (the stray one appears to be a duplicate of the pb, but there are some differences so it can't be deleted outright). How does one join it to the title (or is it not possible at all)?

From RBLSTR2001 I first did an edit-pub as I know that will normally show all of the title records that may be hidden during display-pub. There were none and in the contents section added a title Rebel Sutra type NOVEL by Shariann Lewitt. After approval RBLSTR2001 now has a title reference (pup-type of NOVEL matches the title type of at least one record). I then went to the author's bibliography, clicked "Dup Candidates" and merged Rebel Sutra. Marc Kupper (talk) 19:42, 19 Mar 2007 (CDT)

BTW, I found it purely by chance as a duplicate in Tristan_Elwell's cover art - not having a title, it doesn't appear in searches, and it seems that Stray pubs are shown in author entries only when there are no "real" titles, and the stray-finding script has been taken offline. Shouldn't there be a concerted effort to hunt them down? --JVjr 10:11, 19 Mar 2007 (CDT)

Yes, I'm thinking there should be an effort to chase these down starting with publications that don't seem to have a title record at all just as this one had. Marc Kupper (talk) 19:42, 19 Mar 2007 (CDT)
Oh yes, we have had a stub for this project for a number of months, but we haven't seen much activity in the data cleanup area recently. Many (if not most) of us possess at least rudimentary programming skills and scrubbing a basic SQL database is not exactly rocket science, but it's a question of priorities. I wrote a couple of scripts last year and posted them here (plus a few as per r.a.sf.w requests), but figured it was more important to try to keep up with the flood of new publications that we get every fall. Then beta/moderation started and I had little to no free time left :( I figure once our core editor corps is done with entering their collections (many are up to "D"! :) we will have more time to spend on data cleanup projects. Ahasuerus 00:19, 20 Mar 2007 (CDT)
OI! Just because I entered my Dickson and Dicks over the weekend while the Moderators were sleeping doesn't mean I'm doing them in order! I did my Zelazny ages ago! ;-) (Not sure why, that meant scrabbling around on the floor in front of downstairs bookcase number eight, most undignified... I'm sure there was a good reason though.) Anyway, my edit count suggests I'm over 2000 edits now (remember I'm BLongley and BillLongley) so there should be only 500 or so to go before I finish the "true, reading copy" SF books and can go tackle the upstairs bookcases (including the collectibles, hardbacks, film novelizations, TV tie-ins, biographies, artist portfolios, etc): or AVOID having to clear out the old computer parts/furniture/comics etc to reach them and stick to a decent project instead. So if people have suggestions on a "Project" feel free to make them. BLongley 17:10, 20 Mar 2007 (CDT)

Thanks for the explanation, I'll try to do this in the future (alas, having to wait for the PubEdit to be approved makes the two-step process more difficult and lengthy; still, better an obvious duplicate title that anybody else can merge than an invisible stray pub). I see that the page even includes ISFDB:Data Consistency#Titles that point at themselves that I discussed above recently...

I forgot to mention one thing: Lewitt has many titles listed only under her alternate names, not appearing in the entry for her canonical name. Making them do so means (currently) creating new titles under the canonical name, and making them variants, which is rather annoying drudgery to me; however if anybody else is willing, there is your opportunity. --JVjr 07:12, 20 Mar 2007 (CDT)


One more strange case: BKTG11776 has no Title Reference. However, it is linked from the title 16174. Might the problem with generating the link be that Michael_Moorcock has two different (at least in the DB; Locus1 says that as it happens just this volume has the same contents in both Eternal Champion series, modulo a small difference in the introduction) omnibuses called Hawkmoon? When I try editing the publication, it shows as content Hawkmoon - 1995-00-00 - omnibus - Michael Moorcock. So does the SFBC BKTG11777 - and, I see, it also lacks the Title Reference link. The other two publications of the title list individual novels they contain, and have the Title Reference. So...? --JVjr 09:33, 20 Mar 2007 (CDT)

That was just a Publication/Title type (Novel/Omnibus) mismatch, all fixed now :) Ahasuerus 11:15, 20 Mar 2007 (CDT)
Shame on me, such an elementary thing and I don't notice as if I were some kind of newbie :-) I must be going blind from all that staring at the screen. (But there was no "Type Mismatch" warning, which does appear elsewhere - ah hah, is it because title.py triggers them only for Anthology and Collection?) --JVjr 14:20, 20 Mar 2007 (CDT)
That's right, the mismatch identification algorithm tries to be conservative, perhaps a bit too conservative. Ahasuerus 22:44, 20 Mar 2007 (CDT)
JVjr, in that case you are right in that a type OMNIBUS title would never be expected to contain links to publications other than type OMNIBUS and thus the type-mismatch stuff should be flagging mismatches. A while ago I put in a feature request to open up editing of the code with the idea being that Al would be the gatekeeper on which feature requests can be farmed out. Something I had been wanting to do for a long time but did not have the right resources for was to set up a development box to run ISFDB. I believe I have that now but one, presumably minor, issue is that I haven't the slightest clue on if there is a particular flavor of Unix/Linux I should set up to run MySql and isfdb. Does anyone have pointers on this? Marc Kupper (talk) 02:07, 22 Mar 2007 (CDT)
Since the ISFDB data lives in a generic MySQL database and the scripts are all in Python, the application should run on any OS that supports these two (largerly platform-independent) technologies, which includes most Unixes and Windows. I have the MySQL part running under Windows, but haven't had a chance to configure the Python application yet. <insert the obligatory Prince Humperdinck quote here>
This "computer technology" thingie is changing entirely too fast. And what was wrong with the PDP-11 Assembly anyway?.. Ahasuerus 10:53, 23 Mar 2007 (CDT)

Make pseudonym hierarchy

I just approved a couple of MakePseudonyms on Catherine Shaffer that were of the form: A ps of B, B ps of C, D ps of C. I did the first without realizing the second was in the queue; the second looks right but I now see no way to remove the first one. Take a look at Catherine_Shaffer and you can see the situation. I assume it would be possible to create a bogus extra author record and merge them, nuking the pseudonym pointer; is there an easier way? Mike Christie (talk) 19:03, 27 Mar 2007 (CDT)

As Catherine_Haluska_Shaffer only had one title it was easy to delete the pseudonym connections related to that name by changing the author name in that title to Catherine Haluska Shafferx and then changing it back. The pseudonym connection for Catherine_Shaffer looks correct and so it looks like that fixed everything. I went ahead and made Catherine_Haluska_Shaffer a pseudonym of Catherine_H._Shaffer which seems to be the canonical name.
BTW - Do you think on Catherine_H._Shaffer's page it should say
Uses These Pen Names or Pseudonyms: Catherine Shaffer , Catherine Haluska Shaffer
Currently it says
Used These Alternate Names: Catherine Shaffer , Catherine Haluska Shaffer
with both the past tense and the term "Alternate Names" confusing me. Marc Kupper (talk) 23:54, 27 Mar 2007 (CDT)

Messed up publication entry?

There is a entry for Introduction by Asimov at http://www.isfdb.org/cgi-bin/title.cgi?487381 The Entry for the publication it's in does not list it. Dana Carson 23:27, 1 Apr 2007 (CDT)

The title type was set to EDITOR (which is reserved for magazine editors and should not be used for anything else since it triggers special logic) instead of ESSAY. Once I corrected the type, everything went back to normal. Note that you can't change or even remove EDITOR records from Publications. You have to pull them up directly in the affected Author's biblio and change their types there before you can do anything else. It's messy and there is an outstanding feature request to beef up the software in this area. Ahasuerus 02:17, 2 Apr 2007 (CDT)

Matthew Gregory Lewis

I've just added an addition of The Monk, his only novel, and noticed just a stray publication under his author name. He is in the ISFDB as Matthew Gregory Lewis. I've never seen a copy of this book published under that name (not to say it hasn't been) and as far as I'm concerned his canonical name should be Matthew Lewis. So I don't want to just make it a variant of the current name which appears, incorrectly in my opinion, against most editions of The Monk. Though I know if I just changed the canonical name we'd end up with two Matthew Lewis'.

So what's the best way to fix it? Leave it as it is until a few more titles have been verified. Make Matthew Lewis a variant, or... any suggestions?

--Unapersson 17:06, 3 Apr 2007 (CDT)

OCLC lists 92 editions of The Monk, a few of them duplicates. Most were published as by Matthew Lewis, some as by "M. G. Lewis" and a few as by "Matthew G. Lewis". There is at least one edition (2004, ISBN 1551112272) that used "Matthew Gregory Lewis", though -- you can see a scan of its cover on Amazon.com, among other places.
I suppose the easiest way to handle this situation would be to make "Matthew Lewis" the canonical name and then turn "Matthew Gregory Lewis", "M. G. Lewis" and "Matthew G. Lewis" into pseudonyms. Ahasuerus 17:25, 3 Apr 2007 (CDT)
Thanks that's useful. I'll look at sorting it out when I'm less tired :-) --Unapersson 17:33, 3 Apr 2007 (CDT)

When to add a new publication to one that already exists

I know that this is covered somewhere in the docs, but I can't find it. I have a pub that is exactly like another in the db, except for the publication date, the price, and the Publisher's Catalogue ID. I know I read something about when to add a new pub, and when it's ok to just verify, but I can't find it or remember exactlly what it says. Could someone please point me in the right direction? Thank you. CoachPaul 12:47, 5 Apr 2007 (CDT)

I suspect Help:Getting Started:Verify a Novel is the place you first read it. And all three differences would be enough to create a new publication - although if it's that similar, Clone from the other publication already present. BLongley 14:26, 5 Apr 2007 (CDT)

Editing an existing later edition publication

I want to add story information and page numbers to an already existing later edition publication. The story names in the anthology already are listed under the first edition publication for the book. If I clone the first edition, make the appropriate changes and editions, save it as the later edition, then merge the two later editions into just one entry, will it mess up something else in the db? Is there another way to do this to the existing entry for the later publication without entering all of the data by hand? Thank you. CoachPaul 14:22, 5 Apr 2007 (CDT)

You can't merge publications in this way, only titles. If the earliest publication has all the info, the easiest thing to do would be to create the clone as you suggest, adding all the data from the later publication, in effect creating a duplicate. Then delete the less complete publication. --Unapersson 14:28, 5 Apr 2007 (CDT)
As Unapersson explained, the ability to merge Publication records was removed a long time ago since it was causing problems, so now you can only merge Title records. At one point I requested the ability to "Clone a publication and merge it with an existing one", which is pretty much what CoachPaul wanted to do, but Al said it would be tricky due to the number of potential permutations, so it remains unimplemented. Ahasuerus 15:38, 5 Apr 2007 (CDT)
"Clone contents of one pub to another without any" was my variation: it doesn't help when everyone waits around for the first person to add contents though. (E.g. I chickened out here, and sure enough someone else did eventually add them. When the contents EXIST, but the page numbers don't, it puts people off too: e.g. who'll tackle this pub? There are plenty of reasons for more tools to help with situations like that, but we're short on programmers (not overall: I do that for a living, but am not an ISFDB programmer) but we mostly learn the fastest workarounds. "Cloning" is often the second step (after finding the best candidate TO clone.) BLongley 16:32, 5 Apr 2007 (CDT)
I'll always add contents, if I've got them. Though sometimes it stresses even my patience. --Unapersson 17:41, 5 Apr 2007 (CDT)
Eek! I think I want a full course of antibiotics after just LOOKING at that! ;-) I got tired after doing this one (making variants along the way too!) and will wait YEARS if necessary before admitting that I own the Drabble books. ;-) (For those too lazy to click: 100 Stories of exactly 100 words, price 100 Shillings. (I know I was a person to suggest that it should also be published on the 100th day of the year, and the second was: that may be my only claim to fame in the SF writing world apart from a review of "Return to the Forbidden Planet" in a British magazine I can't even recall the title of, or that they didn't put my correct name on if it's online: can I get British Filk awards considered for inclusion here?  :-/ BLongley 18:12, 5 Apr 2007 (CDT)

I'll tackle any pub, in English, that anyone wants to send me as a gift! CoachPaul 21:20, 5 Apr 2007 (CDT)

If you'll pay the postage, I have about four yards of shelf-space I want to clear out: prior girlfriends acquired a lot of books I have no intention of reading. I've already verified 5 feet of Fantasy books I never want to see again, you could do the rest! :-/ BLongley 21:46, 5 Apr 2007 (CDT)
I've thought twice about adding a new publication, when I know contents have to be added. Sometimes you just have to grit your teeth and start typing. As when I added RVRFRMSTRR2005. But then, I really had no choice, since I was responsible for the contents to begin with :) Mhhutchins 21:50, 5 Apr 2007 (CDT)
Ahem. Take a look at these 4 books, folks. I have a couple and, if anything, the Locus Index understates the number of reviews in each volume... Ahasuerus 01:45, 6 Apr 2007 (CDT)
Once you've entered their contents, Ahasuerus, then you will have won the contest. No one will ever be able to moan again about submissions with contents of a paltry 20 or 30 titles. :-/ Mhhutchins 17:33, 6 Apr 2007 (CDT)
Yes, ONCE you've entered them! ;-) I notice you didn't point to the ones we have here which look mostly empty... In the meantime, a Gold Star to anyone that completes a publication with 100 entries or more. BLongley 16:04, 7 Apr 2007 (CDT)
Going back to CoachPaul's original question - I would just enter your later publication and add a note that it seems to be a later publication. If someone comes along with a different edition (earlier or later, it does not matter), then hopefully they will see the work you have done on your publication, clone that, and verify the results. It may mean they will end up deleting a content-less publication record for their book as there's no way at present to clone or copy just the contents from one publication to another. The other night I had a rather messy set of editing in that a large hardcover collection was reprinted as a pair of paperbacks with the stories shuffled out by year and not the order they appeared in the hardcover. There were ISFDB pub-records for all three editions but the two pb records did not have contents. I ended up cloning the hc publication twice, unmerging the clones as the two pb editions had different titles, doing careful remove-titles from both of them, and deleting the original content-less pb publication records.
Something I have wished for but still need to think of a clean mechanism on how to do it is a way to export a publication to a file and to import or merge into a publication from a file. That would allow us to copy the contents around and/or to edit them using systems that are much more convenient than ISFDB's editor screens. Marc Kupper (talk) 14:13, 8 Apr 2007 (CDT)
I just did the first Drabble book, and it would have been MUCH easier to take content from here and massage it rather than copy and paste a 100 times, fixing on the way.... BLongley 17:25, 30 Apr 2007 (CDT)

Edward L. Ferman Problem

Under Series title "Best from F&SF", you can tell just from looking at the dates that "25 The Best from Fantasy and Science Fiction: 25 (1974)", is clearly out of place. When you open it's link, the pub is "The Best from Fantasy and Science Fiction, 25th Anniversary, (1974 , Edward L. Ferman, Doubleday, hc, anth)", which is already under the Series "F & SF Anniversary Anthologies", and has two editions listed there. THBSRSY441974, the one that is incorrectly placed, has the most information listed of any of the editions that have been entered into the db. Can this title be removed from the incorrect series listion, and placed in it's proper place? I can't figure out how to do it, and don't want to screw something up. CoachPaul 20:20, 6 Apr 2007 (CDT)

Yes, it was possible to remove this Publication record from the incorrect Title record (using "Unmerge Titles" on the left), which would have created a new Title record for this Publication. You could then merge the newly created Title with the Title in the Anniversary series. However, since there were no other Publications under the "bad" Title record, there was an easier way. Simply merging the two Titles (using the "Titles" option on Ferman's biblio page) resulted in the "good" Publication record with all its Contents data getting moved under the merged Title record. Once I did that, all that remained to do was a little bit of polishing (adding the ISBN from OCLC etc) and things are in synch now. The 1977 paperback contentless paperback reprint looks suspect, but I left it alone for now. Ahasuerus 00:02, 7 Apr 2007 (CDT)
Thank you. I got this cloned to a pb, I'm not sure if the suspect one is good or not, but my copy is wierd. It lists a bunch of dates in the beginning of the book, but none of them are specificaly attributed to any stories, nor is there a date of publishment for the book as a whole. I left the dates as they already were figuring that they were in the hb, and used the last date listed as the date for the pub. Should I change the date on the pub to "unknown"? CoachPaul 08:35, 7 Apr 2007 (CDT)
Unfortunately, this is not uncommon with reprint anthologies. OCLC lists this paperback reprint as "1975?", so I have changed the date to 0000-00-00 (displays as "unknown") and added a Note about it. I also created a clone for the British hardcover reprint based on what OCLC has listed. The page count is the same as the first edition's, so I assume that there were no textual differences, but we won't know for sure until somebody verifies it. Ahasuerus 17:53, 7 Apr 2007 (CDT)

Variant Title Problem

I've come across this problem multiple times and don't know how to fix it. Any help would be appreciated. On an Author's page, under the title Short Fiction, I will find a title "AAA", on the next line, or somewhere near by, will be a title "xyz" and below that "Variant Title: AAA". The listing under the "Variant Title" will be empty, and the listing under the title will be full of pubs. I can't seem to find a way to merge the titles, nor to move the pubs from the title to the variant title. There are at least three examples of this under Ray Bradbury, one of which is "The Long Years". If someone could tell me how to fix this, I could do it whenever I see it pop up. Thanks. CoachPaul 10:33, 9 Apr 2007 (CDT)

I figured it out. Sorry to have taken up the space. CoachPaul 14:01, 9 Apr 2007 (CDT)
No problem, but you may want to tell other editors HOW you solved it! ;-) BLongley 15:02, 9 Apr 2007 (CDT)
OK, on the Author's Bibliography page I clicked on Titles, when the new page opened up,I then found the "AAA" title, and then the "AAA" Variant Title, checked the two boxes next to them, and then pushed the MERGE button on the bottom of the page. When you get to the next screen, make sure that you select the option that keeps the "parent information". CoachPaul 15:41, 9 Apr 2007 (CDT)
Thanks! The mod(s) that approved them might want to say what they checked as well - I started looking and they were all done before I understood why you were doing it... :-/ (I'm a newbie mod, I tend to be slower than the rest.) BLongley 18:14, 9 Apr 2007 (CDT)

Adventures in Tomorrow

The hc version printed in the 50's has Kendell F. Crossen listed as the author in the db, and uses the same name as the author of "Restricted Clientele", one of the short stories in the book. My pb edition, published in the 70's lists Kendell Foster Crossen as author of both. How do I handle this situation. There are currently two entries for "Restricted Clientele" one under each name. If I add the book under the Foster name, then there will be two entries for both of those also. Is there an easy way to merge the various entries of Mr. Crossen's, including the misspellings of "Kendell" as Kendall" into one author bibliography? CoachPaul 16:33, 10 Apr 2007 (CDT)

We use "Make This Author a Pseudonym" to connect the different spellings of an author's name - even if they're obviously wrong, if it appeared in print we want to record how it actually appeared. There appears to be a bit of a mess at the moment... :-/
 Author 	Legalname 	Birthplace 	Birthdate 	Deathdate
 Kendall F. Crossen 	- 	- 	- 	-
 Kendall Foste Crossen 	- 	- 	- 	-
 Kendall Foster Crossen 	- 	- 	- 	-
 Kendell F. Crossen 	- 	- 	- 	-
 Kendell Foster Crossen 	Crossen, Kendell Foster 	USA 	1910-00-00 	1981-00-00
Then we can use "Make This Title a Variant Title or Pseudonymous Work" to connect the publications done under the different names.
If somebody's verified a particular spelling, then err on the side of caution and check with the verifier. BLongley 17:07, 10 Apr 2007 (CDT)
I get all of this, but on some records I've seen for instance, "Look Homeward, Spaceman (1956) [as by Calvin Knox ] ". This was from the Robert Silverberg bibliography page here in the db. How do we get the Crossen stories to do this all on one page? The Kendell F. Crossen bibliography already says "Used As Alternate Name By: Kendell Foster Crossen", but none of the entries connect like they do with Silverberg/Knox. CoachPaul 17:34, 10 Apr 2007 (CDT)
I have done something wrong. I Cloned the pub, changed the name from "Kendell F. Crossen", to "Kendell foster Crossen", but it still appears under Kendell F even though the new edition of the pub has Kendell Foster as the author. CoachPaul 17:42, 10 Apr 2007 (CDT)
There are a few things to remember here. The first one is that you have to make every single title which was published pseudonymously into a Variant Title using the "Make This Title a Variant Title or Pseudonymous Work" option in the navigation bar on the Title page for the affected pseudonymous Title. This is what makes comments like "[as by Calvin Knox ]" appear in bibliographies.
Next, you will want to link the main (i.e. canonical name) Author record and his/her pseudonyms using the "Make This Author a Pseudonym" option in the navigation bar on the "Summary Bibliography" page of the pseudonym. What this does is make messages like "Used As Alternate Name By: Kendell Foster Crossen" appear at the top of the pseudonym record's page. Also, it will display a big Pseudonym. See: [canonical name] message at the top of the page, but only if there are no non-pseudonymous Titles associated with this Author record in the database.
You may wonder why it takes so much work to create what seems to be a pretty simple link "XYZ is a pseudonym used by ABC". But consider authors like V. C. Andrews, who was a real person and wrote a number of popular novels in the 1970s and 1980s. The books proved so popular that after her death her publisher hired a ghostwriter to continue producing books under her name. At this point the minority of the books published as by "V. C. Andrews" were actually written by her and the majority have been penned by the ghostwriter, Andrew Neiderman, since her death some 20 years ago. Worse, there are "house names" like Alexander Blade that are/were shared by a dozen+ writers and you need to do additional research to determine who was behind each story.
Granted, the examples listed above may represent a minority of all pseudonym cases in the database, but we had to account for them in some way, which led to the current approach. Ideally, we would have another option in the navigation bar, "Make all Titles by this Author into pseudonymous Titles by another Author", but we are not there yet.
Finally, to address the "Restricted Clientele" issue, there are currently 4 Titles in the database for this story, 2 as by "Kendell Foster Crossen" and 2 as by "Kendell F. Crossen". One of the latter ones is not a Variant Title, which is why it appears on Kendell F. Crossen's biblio page. We just need to merge the two "Kendell Foster Crossen" titles and then the two "Kendell F. Crossen" titles (keep the parent title ID) and we should be all set.
I will be the first to admit that the current setup is rather time consuming and error-prone, but once you do a few dozen times, it gets to be a second nature :) Ahasuerus 19:48, 10 Apr 2007 (CDT)

How do I now get "Adventures in Tomorrow, (Feb 1968 , Kendell Foster Crossen, Belmont, B75-215, $0.75, 236pp, pb, anth)" from Kendell F. Crossen to Kendell Foster Crossen? CoachPaul 21:15, 10 Apr 2007 (CDT)

The Title record for this book currently points to two publications: the 1951 original hardcover published as by "Kendell F. Crossen" (according to Contento) and the 1968 paperback reprint by Belmont, which was published as by "Kendell Foster Crossen". The way to separate the two is to go to the Title page and click on Unmerge Titles on the left. You will then be able to select a checkbox for the Belmont edition, which will convert it to a separate Title. Then it should be easy to convert the newly created Title record to a Variant Title of the 1951 book. Ahasuerus 21:36, 10 Apr 2007 (CDT)

Duplicate Short Fiction Title Entries

I run into this alot, where there are two Title entries for the same Short Fiction piece on an Author's Bibliography page. The only difference between the two is that one will have the month of publication in the date, and the other will have the month listed as 00. In other words, one will have more detailed information then the other. Can I go ahead and merge these occurances when I come across them, or should they be researched on a one by one basis? This is what comes from working on almost nothing but Collections and Anthologies. CoachPaul 15:18, 12 Apr 2007 (CDT)

Yes, you can merge them. When I approve, I check first to make sure that the month comes from a known first magazine publication, or known publication date of a book. Sometimes it's one we don't have here, in which case I might go try and find a reference I'm comfortable adding: e.g. we don't commonly have Playboy or Collier's magazine entries set up, but they're normally easily findable. I know this does make "auto-merge" a bit more difficult as the references to original magazine dates in many books are unreliable, or just gives a copyright date rather than publication date: but if so the differences are worth recording in the notes. There's a case for a feature request where adding "manual merge" content goes on a "To-Do list" - but I suspect we'd argue whose "To-Do" list it goes on! ;-) BLongley 20:16, 12 Apr 2007 (CDT)

Edit and/or Verify?

In the following situation which is the proper course of action? I have the following pub. Title: Man Plus Authors: Frederik Pohl Year: 1977-00-00 My information in the pub is the same as in the db, except that my pub lists the year of this edition as September 1977. Should I edit the pub to reflect the September '77 date? If I do this can I then "Verify" the pub? As a second choice, should I leave the date alone and just verify by Primary? As a third choice, should I "Clone" the pub and then add the September part of the date to the clone? CoachPaul 21:51, 13 Apr 2007 (CDT)

If nobody else has verified that pub, adding to the accuracy with a month is fine (In My Humble Opinion). If they have, ask the first Verifier to check. BLongley 22:10, 13 Apr 2007 (CDT)

Also, if working on Collections/Anthologies, and I add page numbers, can I then "Verify" the pub by Primary? CoachPaul 21:51, 13 Apr 2007 (CDT)

If it matches everything your pub says, sure, verify it. Verification doesn't finalise things - it does put you up as a person that will be niggled with requests to double-check a spelling or an entry for Interiorart or such, no harm done. You'll be considered an expert on that title! ;-) BLongley 22:10, 13 Apr 2007 (CDT)

A bit of help with this merge?

56844 and 325911 need to be merged, but they're variant titles on each other, and I'm unsure of how to proceed. grendel|khan 23:08, 15 Apr 2007 (CDT)

If my understanding is correct, the underlying problem is that there are four related Title records here. The original 1958 story was published as "Segregation by Brian W. Aldiss". It has since been reprinted as "Segregation by Brian Aldiss", as "The Game of God by Brian Aldiss" and as "The Game of God by Brian W. Aldiss". We had a circular variant Title relationship set up, which is always a bad thing, and we had two other Title records that were not linked to them at all.
Here is how I fixed it. I pulled up all linked Titles and broke the links by using "0" as the parent Title in "Make This Title a Variant Title or Pseudonymous Work". I then somewhat arbitrarily picked the earliest Title, "Segregation by Brian W. Aldiss", as the primary Title and then made the other 3 into Variant Titles of this one. Take a look at the result, I think it now does what we want it to do. Ahasuerus 23:23, 15 Apr 2007 (CDT)
Excellent! grendel|khan 11:36, 16 Apr 2007 (CDT)

Adding titles to a publication duplicates them

I added an omnibus (536605), which referenced several previously extant titles. However, it created new copies of those title records, and I had to merge them--this, despite that I'd copied and pasted the title, date and type of the contents directly from the title records for the individual novels. I've had to do this for (I think) every collection or omnibus I've entered. Am I doing something wrong?

Nope, you are not doing anything wrong, it's just the way the software works at the moment :-\ Ideally, each submission would be checked against the database to see if any of the records that are being added are the same as the records that we already have on file. The software would then give you the option of merging the new recod(s) with the currently existing one(s). Or, better yet, use Ajax to let you select existing records during the data entry entry phase. But we don't have anything like that in place, so for now everything has to be merged manually :( Ahasuerus 12:39, 16 Apr 2007 (CDT)

This is the SQL that was the result of adding the omnibus record. I see that it does no searching for the entered title data. Are we supposed to have to manually merge the titles? grendel|khan 12:24, 16 Apr 2007 (CDT)

  • insert into pubs(pub_title) values('xxx');
  • update pubs set pub_title='Classic Star Wars: The Lando Calrissian Adventures' where pub_id=181921
  • update pubs set pub_tag='CLSSCSTRWR1995' where pub_id=181921
  • update pubs set pub_year='1995-00-00' where pub_id=181921
  • update pubs set pub_pages='409' where pub_id=181921
  • update pubs set pub_ptype='tp' where pub_id=181921
  • update pubs set pub_ctype='OMNIBUS' where pub_id=181921
  • update pubs set pub_isbn='0345394437' where pub_id=181921
  • update pubs set pub_price='$10.00' where pub_id=181921
  • select publisher_id from publishers where publisher_name='Ballantine Del Rey';
  • update pubs set publisher_id='15' where pub_id=181921
  • insert into pub_authors(pub_id, author_id) values('181921', '427');
  • insert into titles(title_title) values('xxx');
  • update titles set title_title='Classic Star Wars: The Lando Calrissian Adventures' where title_id=536605
  • update titles set title_copyright='1995-00-00' where title_id=536605
  • update titles set title_ttype='OMNIBUS' where title_id=536605
  • insert into canonical_author(title_id, author_id, ca_status) values('536605', '427', '1');
  • insert into pub_content(pub_id, title_id) values(181921, 536605);
  • insert into titles(title_title, title_copyright, title_ttype) values('Lando Calrissian & The Mindharp of Sharu', '1983-07-00', 'NOVEL');
  • insert into canonical_author(title_id, author_id, ca_status) values('536609', '427', '1');
  • insert into pub_content(pub_id, title_id) values(181921, 536609);
  • insert into titles(title_title, title_copyright, title_ttype) values('Lando Calrissian & The Flamewind of Oseon', '1983-10-00', 'NOVEL');
  • insert into canonical_author(title_id, author_id, ca_status) values('536613', '427', '1');
  • insert into pub_content(pub_id, title_id) values(181921, 536613);
  • insert into titles(title_title, title_copyright, title_ttype) values('Lando Calrissian and the Starcave of Thonboka', '1983-12-00', 'NOVEL');
  • insert into canonical_author(title_id, author_id, ca_status) values('536617', '427', '1');
  • insert into pub_content(pub_id, title_id) values(181921, 536617);
  • update submissions set sub_state='I' where sub_id='533813'
  • update submissions set sub_reviewer='134' where sub_id='533813'
  • update submissions set sub_reviewed=NOW() where sub_id='533813'


You didn't have to copy all that; it appears that the relevant script is pa_new.py and you can see the complete algorithm there (if you understand it, which I don't at least at the first glance).
In short: yup, we are supposed. As stated above by Ahasuerus and most likely even in the appropriate help page, it's just the way the things are now, a (missing) feature rather than a random bug: adding any title to any publication creates a new title no matter if there already is one such, and they need to be merged manually ex post. (So I took the liberty of changing the section title above to be more descriptive and easy to find.)
IMO it is one of ISFDB's most annoying holes / urgent feature requests currently (and I really hope that something can be done in coming weeks with it before Al disappears again): it wastes manpower terribly, as well as the unicate title ID #s which are immediately thrown away, just incrementing the global counter.
Until this is fixed, I can recommend only a few bypasses:
  1. Wait with adding collections/anthologies/omnibuses and concentrate for now on the novels, or verifying those multi-title publications that are already in the system.
  2. Wherever possible, clone (and modify) an existing publication - that's AFAIK the only way where duplicates aren't created. (Yes, it's not an option with most omnibuses; but many coll/anths already have one such.) It also saves typing anyway.
  3. And if you really need/want to add new content, copying detailed dates is really a waste of time: if there are already correct title entries for the contents, they'll be preserved when merging the new duplicates into them.
--JVjr 04:02, 17 Apr 2007 (CDT)

The Weapon Shops of Isher by A. E. van Vogt

I've got a list of things to do to this entry on User talk:CoachPaul#The Weapon Shops of Isher by A. E. van Vogt. Can one of the mods help me with #3 there please? CoachPaul 13:38, 16 Apr 2007 (CDT)

Now I could use help with #5. CoachPaul 17:41, 16 Apr 2007 (CDT)

All reviews in a magazine issue have duplicate entries

All reviews in SPCSFAMAR1953 turn up twice in George_O._Smith. However both the magazine's display, Edit Publication and Remove Titles opening screen appear OK - I haven't tried submitting anything so far. Anyway, can somebody (able to analyse the DB dump, say) discover what's the root of the problem here, how likely it is that it will repeat and whether plain merge suffice or a more complicated fix is needed? --JVjr 08:06, 17 Apr 2007 (CDT)

Well here's what I've figured out so far:
  • The ENTIRE magazine is duplicated, not just the review entries.
  • There are 2 distinct (and complete) versions of the magazine in the database. See:

http://www.isfdb.org/cgi-bin/pl.cgi?181985 and http://www.isfdb.org/cgi-bin/pl.cgi?181981

  • There is only 1 submission for this magazine in the submission history. This means that the submission was approved twice.
  • There is code in the approval app to check that the submission is in the NEW state, but there is a theoretical race condition that can occur between the time the state is checked, and the time the submission is placed in the INTEGRATED state. If the MySQL check can cause the application to give up the processor, that would allow another moderator to come in and approve it simultaneously.
  • There were 2 moderators (Scott Latham and Mhhutchins) who were approving submissions at that time. It's possible that the first approval happened during one of those short system freezes we see on occasion, allowing another moderator to start an approval before the prior one finished.
At any rate, the bottom line is that there are two distinct versions of that issue, including stories and editor entries. I'll have to do some experiements to see if I can demonstrate any race conditions (which will become more prominant as the number of moderators and system load increases). Alvonruff 12:49, 17 Apr 2007 (CDT)
Al, I'm certain I did not approve this submission. Is there any other plausible explanation? (Scott Latham 15:31, 17 Apr 2007 (CDT))
I've personally seen occasions where I appear to be the only moderator around, press the approval button, and get an error that the submission is not in the new state. That implies that it is somehow possible for a single click to generate two responses - perhaps an error with the squid proxy server. That's all speculation at this point, but it is good to verify that only one moderator was involved. Alvonruff 18:18, 17 Apr 2007 (CDT)
Well that explains that. Wondered what was going on when I was verifying my recent data entry. So duplicate keys are allowed in tag field? Should one magazine record be deleted?--Swfritter 12:00, 18 Apr 2007 (CDT)
The duplicate tag fields are another probable indicator of a race condition. I think the best solution for the duplicate magazine would be to select one of them, do a title remove and select all the titles, and then delete the shell publication record. Alvonruff 13:05, 18 Apr 2007 (CDT)
Will do. --Swfritter 14:31, 18 Apr 2007 (CDT)
Title removal approved, publication deleted. Ahasuerus 21:45, 18 Apr 2007 (CDT)

Magazines Pub Type

What kind of Pub Type do I put for an 8 1/2 X 11 size magazine? CoachPaul 12:53, 17 Apr 2007 (CDT)

That's "Letter" size paper to me. Slightly different from the "A4" used elsewhere - Brin1 has recently been using "A4" for some British magazines, I'd like to keep the two distinguished. If we go for the physical dimensions, please keep them accurate and don't round them, e.g. A4 is 8.3 x 11.7 inches and could easily be abbreviated to 8" x 11" as well. And anyone that's photocopied one size onto the other knows they are NOT compatible. :-( BLongley 13:30, 17 Apr 2007 (CDT)
Still, if you're going to be adding 200 of them I think YOU get to choose! ;-) BLongley 13:30, 17 Apr 2007 (CDT)

Adding Cover Art

Is it possible to add the cover art from the following web site to my Dragon Magazine project? http://nightmare.org/dnd/gallery/main.php?g2_itemId=10349 If so, could someone please tell me how to do it? I came up with a blank on the Help Pages as far as addint art was concerned. CoachPaul 07:25, 18 Apr 2007 (CDT)

The help is here, but as you can see there's only a few sites we have permission to link to. Looking at the site you propose made me think 1) it's too slow, 2) it's not given us permission to link to it (although you could ask them) and 3) we're never going to get a stable image. (That "main.php?" bit and the result suggests that we aren't accessing a stable file that's posted on a webserver for all, we're causing it to go drag it out of a database.) It does look as though they don't mind giving it away for free as I just downloaded it via their shopping-cart for nothing, but I would imagine somewhere there are terms and conditions about the uses of it.
In cases like this, I just add the URL to the notes so that someone can manually go find it, e.g. as with this magazine. (If you just go search that site in general you have to pull up EVERY picture for the year to find what week it relates to, so a direct link could save a lot of time.) BLongley 14:37, 18 Apr 2007 (CDT)

All Judgment Fled

I want to move James White's serial "All Judgment Fled" for example to its main title that includes the book publications and I don't see an obvious way.Kraang 21:27, 19 Apr 2007 (CDT)

If there is a novel called All Judgment Fled for James_White then the serial records will get collected with that novel. I don't see one. If the novel version was published under another title then make the new All Judgment Fled title record the canonical title and the novel a variant. You will then see the serials and the variant novel underneath All Judgment Fled. Marc Kupper (talk) 03:14, 20 Apr 2007 (CDT)
We have it on file, but the title was misspelled :) I have fixed the title and everything looks good now. Moving the discussion to the main ISFDB:Help desk page. Ahasuerus 04:04, 20 Apr 2007 (CDT)

Audio books

Is there a consensus on how to handle audio books, books-on-tape (-cd, -mp3, etc.)? --WimLewis 22:44, 20 Apr 2007 (CDT)

I just enter "audio" or "CD" as appropriate. Now that the field length has been increased to 32 characters, it should be able to handle almost anything we can throw at it. Well, within reason :) Ahasuerus 01:47, 21 Apr 2007 (CDT)
Well, we've thrown 98 things at it so far... :-/ "audio" leads with 181 pubs, with "audio cd" and "audio MP3" well behind at 9 and 7. We have 4 "audioboo" (no "k") and only 1 "Audiobook". There's a solitary "audio ca" too - "cassette"? I may have to remember that, somewhere I have a set of "SFX magazines" that were published that way - although I'm pretty sure there was no SF content. There were "SFX Magazines" on CD-ROM though that did... BLongley 17:06, 22 Apr 2007 (CDT)
I think the answer is "no, we don't have a consensus" - so just do what you think best, and if it looks good we'll follow your example. BLongley 17:06, 22 Apr 2007 (CDT)

Correcting/regularizing an author name

I just saw a submission with with the cover by "t. Winter-Damon." I was going to add a comment to the editor but then took a look at t._Winter-Damon and realized that he regularly uses "t." The question is - Is there an easy way to fix this in ISFDB? At first I tried an edit-author to change the canonical name and was rewarded with the apparently silent creation of a pseudonym.

  • t. Winter-Damon - Summary Bibliography
  • Used As Alternate Name By: T. Winter-Damon
  • Used These Alternate Names: T. Winter-Damon

I ended up setting up a dummy author, T. Winter-Damonx, author merges "t. Winter-Damon" into this and then to set up a second dummy author, T. Winter-Damon, and author merged back to the correct name. Is there a better way to handle this? Is there a way to get rid of the author pseudonym entry? Marc Kupper (talk) 16:50, 20 Apr 2007 (CDT)

Hm, that is peculiar! I have never seen a Canonical name change result in pseudonym creation.
Based on prior experience, my guess would be that there were already two Author records for Winter-Damon on file, one under "t. Winter-Damon" and one under "T. Winter-Damon" and they had some kind of pseudonym relationship already set up. Once you changed the canonical name in one of the records, it resulted in a "two records with the same canonical name" situation, which is a big non-no. The reason is that the software uses the canonical name as the main key to access Author records and a duplicate key makes one of the records inaccessible. When I see a duplicate canonical name, I change it (for the accesible record) to "Firstname Lastname (test)", which lets me access either records and analyze what needs to be done. We really need to beef up the software to prevent editors from accidentally creating duplicate canonical names, though :( Ahasuerus 17:49, 20 Apr 2007 (CDT)
I believe it's possible the pseudonym was there before I started but the puzzle is that I did author searches a number of times while doing this and it always found one record. Marc Kupper (talk) 02:39, 21 Apr 2007 (CDT)

Making a Series part of another Series

Paul Kidd has three books on his biblio page that are in concern here.

White Plume Mountain (Greyhawk) (1999)
Descent into the Depths of the Earth (Greyhawk) (2000)
Queen of the Demonweb Pits (Greyhawk Classics) (2001)

They are all part of the Greyhawk Series, although they have not yet been linked there in the db. I will remove the "(Greyhawk)" parts from the titles and link them to the Series in the db, however these three books are a "Sub-series" of three books that should be read in they order that they were written. I wish to be able to give this sub-series the title of "Justicar" as he is the main character of these books, and number them 1-3. Most of this I know how to do, however, how do I make one series, a sub-series of another? CoachPaul 11:59, 23 Apr 2007 (CDT)

Put them into the 'Justicar' series first. When you've created the 'Justicar' Series, bring details of that up (normal search of type 'Series', enter 'Justicar' as what to search for) and then use the 'Series Data' link under 'Editing Tools'. There you can set the parent series of 'Justicar' to be 'Greyhawk'. BLongley 13:13, 23 Apr 2007 (CDT)

Double Entries Again - Imaginative Tales May 1957

Just added. I will clean it up later in case somebody wants to make an attempt to use the entries to find out why this happens periodically.

Pandora's Legions

The various parts of Pandora's Legions were previously published as Part I, Pandora's Planet ss, Parts II, IV, VI, VIII, rest of the Pandora's Planet novel, Part III, Pandora's Envoy, Part V, The Toughest Opponent, Part VII, Trap. So in one sense the collection is a novel called Pandora's Planet plus a short called Sweet Reason, in another it's a novel and several shorts.

So does my way of entering it make sense to the mods? Wanted to be able to show the connections to the short stories that make up the parts of the novel. Dana Carson 03:53, 24 Apr 2007 (CDT)

I have approved the submission and massaged the phrasing in the Note field a bit, but we may want to go over it again and clarify what sections first appeared on what form. Policywise, that's pretty much what we usually do with fixups: record them "as is" and then add an explanatory note about the work's origins. Ahasuerus 15:24, 24 Apr 2007 (CDT)


Creating a title record for magazines

I see that the latter years (the Wizards of the Coast issues) of Amazing Stories is in a mess, especially the title records, making it hard to do searches. Before I start cleaning it up, I want to group the issues by years. How do I go about creating a title record so that I can merge various issues into groups? Mhhutchins 12:43, 25 Apr 2007 (CDT)

Here are some notes on the EDITOR title type that I put together a while back and Mike archived:
  1. Every magazine Publication should have the "Editor" field filled out, but this data element is specific to the Publication and doesn't affect Titles;
  2. Every magazine Publication should have a Title (type "EDITOR") associated with it and its Title_title should be the same as the title of the magazine issue and the author/editor is the magazine editor'
  3. This EDITOR Title is created automatically for newly entered magazine issues, but it is not displayed in the Contents area and not pulled up in Edit Pub;
  4. Some previously entered magazine Publications do not have this EDITOR Title associated with them, so they will be showing up as Stray Publications in their Editor's biblio until an EDITOR Title is added to them;
  5. We typically do a Title Merge on all EDITOR Titles for the same magazine/year so that someone like John W. Campbell or Stanley Schmidt doesn't have 300 "Editor" records in his biblio. This results in each Editor record in the Author/Editor's biblio fanning out into up to 12-13 Publication records when you pull the Title up;
  6. The way to fix pre-existing magazine Publications without this EDITOR Title (Open Display Bug 20036) is to go through affected Pubs one at a time, manually add an Editor Title to them, and them merge them when needed.
Which reminds me that I need to create a Help page for this stuff... Ahasuerus 13:14, 25 Apr 2007 (CDT)
Thanks for the info. I created an EDITOR Title for each of the 1987 TSR issues of Amazing and merged them into this title record. Looks OK?
Looks good! I have added these EDITOR Titles to the Amazing Stories series. Not only does it force them to be displayed together (which is important when you have folks like Hugo Gernsback who edited 3+ magazines at the same time), but it also allows our users see complete editor bibliographies for a given magazine. Or at least it will once we have EDITOR records added throughout :) Ahasuerus 15:17, 25 Apr 2007 (CDT)
Funny thing is, when I first searched for the editor, Patrick Lucien Price, all of the issues of Amazing that he edited were listed on his page, but as "stray publications". After I merged the year 1987, that was the only thing that appeared on his page. All of the remaining issues disappeared. I had to do an advance search in order to find them. What exactly is a "stray publication"?
"Stray Publications" are Publication records that do not have associated Titles. They are only displayed on the Summary Bibliography page when there are no legitimate Titles for the author on file. The most frequent cause of their appearance is a pseudonymous Author record that has had all of its Titles set up as Variant Titles while the Author record itself hasn't been "Marked as a Pseudonym" yet. In this case, however, the problem was the lack of EDITOR records in the Magazine issues in question, which you have no addressed. Ahasuerus 15:17, 25 Apr 2007 (CDT)
When I did further research into the Wizards of the Coast/Kim Mohan edited issues I discovered that all of the records that I checked into must have been created by a bot. None have more than the basic ISBN info and are shown as novels instead of magazines. Looks like it would be better to start from scratch with those issues. Mhhutchins 14:07, 25 Apr 2007 (CDT)
This is the exact same problem we have for the latter Dragon Magazines, also by WotC. I'm thinking that when I get to them it will just be easier to delete what is already there and start anew. CoachPaul 16:01, 25 Apr 2007 (CDT)
I've ran across another problem. When I tried to make "Omar Gohagen" an alternate name of Elinor Mavor for issues in 1979-1980 of Amazing, I must have done something wrong. Now both are listed on the Amazing Stories series list. Any idea how to correct this? I tried unmerging the issues from the editor title but didn't notice any change. Mhhutchins 17:21, 25 Apr 2007 (CDT)
Nope, you didn't do anything wrong. It's our old friend Open Series Bugs 30013:
The reason why Series display is so rudimentary at the moment is that Al has been making a lot of changes to the main Summary Bibliography page and other pages' logic has fallen behind. Now that the core logic is reasonably stable, it needs to be propagated from the Summary page to all the other pages, including the Series page. Ahasuerus 18:26, 25 Apr 2007 (CDT)

Moving Awards From One Title to a Similar Title

How do I move the Locus 1999 Anthology Award from "The Fantasy Hall of Fame" (1983) ed Silverberg & Greenberg, to "The Fantasy Hall of Fame" (1998) ed Silverberg? Also there are two reviews that appear under both titles, but judging by the dates, they may be for the latter book, however I'm not 100% sure. CoachPaul 19:09, 25 Apr 2007 (CDT)

I don't think you need to move the reviews, they already appear under the correct title, they just also appear under the incorrect one as well. :-/ And award data doesn't appear editable at all. I suspect the matching logic could be fooled by retitling the two entries, e.g. adding "(1983)" and "(1998)" suffixes, but if you can't do that to the award title as well you'd lose both connections. BLongley 12:52, 27 Apr 2007 (CDT)
Reviews are currently displayed based on whether the author string and the title string match, so there is no way to distinguish between two identically titled anthologies edited by the same person. Al has plans to overhaul review (and Serial) support so that they would point to internal database IDs, which are unique, but I don't think it's at the top of his list of priorities at the moment. Ahasuerus 14:11, 27 Apr 2007 (CDT)
Well, Awards has an "Author String" (with both Authors/Editors in) but it must be matching against a derived Titles "Author String" which isn't quite right. Similar with Reviews I guess. And the unimportance of Author ordering to us doesn't help... :-/ Still, it's not worth fixing the lexical match any more for so few problems, I'm happy to wait for the final fix - although I'm sure we've got some sneaky data problems that will ensure further problems, however brilliant the programmer. BLongley 14:38, 27 Apr 2007 (CDT)
Actually, Award records are no longer lexically matched (thank Cthulhu!) as of late 2006. Only Reviews and Serials still suffer from this problem. Anyway, have to run, have a plane to catch, should be back online by midnight East Coast Time. Ahasuerus 14:50, 27 Apr 2007 (CDT)
OK, Awards CAN be fixed then, but only by the programmer? BLongley 15:19, 28 Apr 2007 (CDT)
Right, somebody with direct access to the database (which only Al has at the moment) can change the pointers. There may be a way to do it via the application as well, but I haven't played with this area lately, so I am not sure. Al added a rudimentary Award editor around Christmas time, then took it off line to beef it up, at which point I figured that it wasn't worth spending much time on until Al had something more permanent in place. Ahasuerus 21:57, 28 Apr 2007 (CDT)
It seems silly but - Where are the lists of awards? Using Google I can find individual pages such as the 2003 Ditmars [1] (the last year in ISFDB) but I can't find a directory of the awards. Marc Kupper (talk) 10:46, 1 May 2007 (CDT)

(unindent)I don't think the Award list made it from ISFDB-1 to ISFDB-2. You can find awards for individual authors by selecting the "Awards" option in the navbar and then you can choose "Previous Year" and "Next Year" to navigate within the award, but I am unaware of any other user-accessible functionality at this point. Hopefully Al will address the issue as part of his "award editor" project mentioned above. Ahasuerus 12:07, 1 May 2007 (CDT)

Be careful what you wish for Marc! There'll be a lot of work when the Award editing code comes online, I think. Looking at a few examples, we don't have links to titles that are almost always in the database, just authors... and there'll be the temptation of correcting the IMDB links for TV shows, etc. (This example is particularly bad for IMDB links.) BLongley 13:32, 1 May 2007 (CDT)
As far as accessing the awards go - would it work for now to just add a wiki page or would that make it too obvious that the awards have not always been updated in the past couple of years? Nebula goes to 2005 [2], 2006 Hugo [3] looks like an early revision, etc. The awards pages are available and while there may be bad links here and there the data should mostly be ok. The whole reason this came up is because Locus has two typos in an author's name - I had spotted one and was wondering why the ISFDB awards page did not show the award despite my trying both versions of her name. I hunted for the ISFDB awards page itself for that year before giving up and spotting the second typo in the name. Marc Kupper (talk) 01:22, 3 May 2007 (CDT)
So long as I don't have to create it, I think a Wiki page would do for now. You could add the missing years in from other sources too, and then we could use those to populate the "official" pages when done? BLongley 15:37, 3 May 2007 (CDT)

Camouflage by Henry Kuttner

On the biblio page for this story are listed three serials with the name "Camouflage" (2004). When you click on any of these three links, they take you to a story named "Camouflage" written by Joe Haldeman. I can't figure out how to remove the serials. Somehow I feel like I'm asking about Awards again...Am I opening another can of worms? CoachPaul 16:45, 27 Apr 2007 (CDT)

Sort of - you're just reminding us there's another bug to fix. :-/ Serials do end up with the strangest associations, it seems to be a mere title match. Just report the weird ones you see, and it might move up the "to-do" list. BLongley 15:24, 28 Apr 2007 (CDT)
G. Harry Stine serial "Manna" linked to John Christopher. John Christopher has a short story of the same name and the serial appears underneath that.--Swfritter 19:19, 28 Apr 2007 (CDT)
Yup, the three of you have run into Open Display Bugs 2046 -- "Only Serials titles (not Authors) are matched against Work titles in Title biblio pages, therefore Serials written by other Authors can be linked to unrelated Titles" -- I am afraid. Ahasuerus 19:26, 28 Apr 2007 (CDT)

Entering unicode for foreign author names

ISFDB has links to Adam_Wiśniewski-Snerg but his bibliography can't be displayed as this UrlEncode()s to %c5%9b. Currently the ś is unicode 015B I suspect what I'm looking for is a character somewhere in the unicode 2000 range that will map correctly but so far I have not found it. Anyone have ideas? Marc Kupper (talk) 03:39, 29 Apr 2007 (CDT)

015B is correct according to Microsoft's keyboard page. My guess is that you have run into one of many Unicode bugs and/or data corruptions in the ISFDB. To quote Al, "Have I mentioned that I hate Unicode?" Ahasuerus 11:02, 29 Apr 2007 (CDT)
I did some playing around with this. It looks like the the browser converts links into wikipedia:UTF-8 and as it's similar to ASCII everything works well until we hit characters that are represented by multi-byte sequences. The question then becomes - how is the data getting encoded/stored in the database? It must not be UTF-8 otherwise Adam_Wiśniewski-Snerg would just work meaning Al needs to decode the UTF-8 into Unicode and then recode that into whatever format he uses in the database. FWIW - the wiki seems happy with Bio:Adam_Wiśniewski-Snerg though it's not clear if they are decoding the UTF-8 or the file name would have characters 0xC5 and then 0x9B in the middle.
The irritating aspect is that it seems everyone is expected to reinvent the wheel code/debug their own encoder/decoder. For now I changed the guy's name in ISFDB to Adam_Wisniewski-Snerg. Marc Kupper (talk) 15:09, 29 Apr 2007 (CDT)
The biggest concern that I have with using Unicode characters is that they make searching more difficult. How many users do we have who know how to search for "Wiśniewski" or Звездный Клондайк? Ideally, searching for either "Wiśniewski" or "Wisniewski" would pull up the same record(s), but I don't know how hard it would be to implement in MySQL. And the only way to do non-Latin alphabets in a Latin-friendly way is to use one of many (emphasis on many) transliteration schemes, so we introduce an element of ambiguity, not to mention even more variant titles. See Al's favorite quote above :( Ahasuerus 23:08, 29 Apr 2007 (CDT)
The search function already maps some normal characters into their accented version. For example, searching for Jose Farmer finds Philip_José_Farmer. I agree this may be more of a challenge for the multi-byte unicode characters and a quick test finds that it does not work for Wiśniewski but it's not clear if that's because it's not in the database correctly or if Al needs to do something extra to allow for searches of ś. Marc Kupper (talk) 01:03, 3 May 2007 (CDT)

Twilight by John W. Campbell Jr.

How do I fix this?

Twilight (1934)
Variant Title: Twilight (1934) [as by Don A. Stuart ]
Variant Title: Twilight (1934) [as by John W. Campbell ]
Variant Title: Twilight (1934) [as by John W. Campbell ]

I can find no way to merge the two identical variant titles. CoachPaul 17:15, 1 May 2007 (CDT)

You can pull up the John W. Campbell (no Jr.) page and select "Titles" in the navbar on the left. You will then see a list of all Titles published as by "John W. Campbell" and you can then merge them as needed. Note that the Summary Bibliography page for "John W. Campbell" will not list "Twilight" or any other titles that are marked as "Variant Titles", but the "Titles" list will. Not very intuitive, but that's how it works for now. You can always merge Title records via the "Advanced Search" route as well, although for common words like "Twilight" it may require limiting the search in some way. Ahasuerus 22:21, 1 May 2007 (CDT)

Adding a tag

I finally stumbled across a book that had some categories listed beneath it and I followed that link for a while until I came to the Fritz list of all the Tags, with the numbers of works to which each had been attached. I saw one for Dinosaurs and went there too. Now I want to add a tag to my own book about dinosaurs, The Thirteenth Majestral, AKA Dinosaur Park but can't figure out how to do it. I'm sure it's excruciatingly simple, once it's been explained. Many thanks! Hayford Peirce 15:01, 2 May 2007 (CDT)

I would start by pulling up the book, then clicking on "Add Tags", then scrolling down to the field "Tag1" at the bottom of the screen. I would then enter the word "dinosaurs" in the field and click on Submit Data. There is no moderator approval involved, so the tag will be added to the book right away. 17:09, 2 May 2007 (CDT)
P.S. I have created Help:Screen:AddTags help page for tags. Please feel free to improve/rewrite/uplift it :) Ahasuerus 17:34, 2 May 2007 (CDT)
Thanks for the help -- I feel like an idiot. But I was looking on the *left* of the screen, where there are all the other editing functions, rather than the top, where the Add Tag link is, of course, very visible -- once your lying eyes, as Chico Marx might say, look in that direction. Hayford Peirce 18:46, 2 May 2007 (CDT)
Would it help if there were a link to add a tag (and probably for votes as well) to the left navbar? Alvonruff 12:11, 4 May 2007 (CDT)

Another tagging question, hopefully not as dumb as the first one

Okie, I've added a bunch of tags to "The Thirteenth Majestral", including a "Jack Vance pastiche" tag, which I then also added to 4 other works cited in the Wikipedia article on Jack as emulating him. When you look at "Majestral", one sees that a variant title is "Dinosaur Park". When I clicked on that, however, I see that none of the tags added to "Majestral" followed along to "Park". They could, of course, be added individually, but that would be tedious, particularly if there were 17 different editions of a given work listed, say. Is there a specific policy on tagging variant names and different editions of the same work? In makes sense, in one way, to just have the tag applying to the original work. On the other hand, suppose someone picks up the later edition, renamed, say "Dinosaur Park" and does a search in the database. He'll find that book but NOT the tags associated with it -- unless he consciously makes the effort to move one level up to the original title. I don't feel strongly about this one way or the other but I *am* curious.... Hayford Peirce 19:06, 2 May 2007 (CDT)

This is something that I have been thinking about for the last few weeks. "Parental" ratings and tags are currently not displayed for Variant Titles, which should be fairly easy to change programmatically, but is that what we really want to happen?
Some Variant Titles are just title changes and all tags/ratings can be applied to the original title or to the variant title(s) pretty much interchangeably. However, some variant titles are actually revisions and may have separate ratings. To use an extreme example, Keith Laumer's Once There Was a Giant as published in 1968 was one of his better efforts, a solid 8 or 9, nominated for the 1968 Nebula, etc. On the other hand, the 1984 rewrite of the same name was very poor, probably a 3 (less if you consider how badly he had ruined a perfectly good story for the younger generation), so it clearly needs a different rating.
I suppose the same considerations could conceivably apply to tags as well. Suppose a Variant Title was an expansion/revision that introduced new topics or topes, although I would expect it to be a less common occurrence. Ahasuerus 20:15, 2 May 2007 (CDT)
I've occasionally run into a similar issue in the opposite direction. I'm often working at the publication level and when I click on the title reference I'm taken to the variant title record. If I'm not paying attention I end up adding series names/numbers, title notes, tags, etc. to this record which is usually invisible (it's only visible if a title has more than one variant). I'm also wondering how often someone's on a bibliography, clicks a title, and then clicks "add new publication" not realizing they may need to first track down the variant title record.
I personally would vote for votes, tags, series notes, storylen, and the title notes to apply to both the parent and all variants. If a story is rewritten for the YA market then technically it's not a "variant title" though it's convenient to use the VT mechanism to link the story to the original edition. I'm not sure if it's practical in terms of coding but it seems there is a need for some sort of linking thing that looks/acts like variant titles but would be used for revised titles. The link may need to be fuzzy as sometimes a story is based on or adapted from another story. Marc Kupper (talk) 01:37, 3 May 2007 (CDT)

Tomorrow 1 by Robert Hoskins

The db, the front cover, and Contento all call it "Tomorrow 1". The book spine, and the title page call it "Tomorrow I", with the Roman Numeral "I" instead of the "1". Does the front cover trump the title page, or visa-versa? Do we leave it with the "1" because it is more visible, or change it to the "I" because it is probably more correct? CoachPaul 11:51, 3 May 2007 (CDT)

As per the Help:Screen:NewPub:
Books. For a book, use the title page to get the title. This is typically the page with the copyright information on the back. Don't use the title on the cover, spine, or page running heads.
In other words, the title page trumps the title on the cover. It is generally advisable to record any variations in the Notes field. Ahasuerus 12:39, 3 May 2007 (CDT)
I might have a stab at proposing some changes to that help this weekend. For instance, what if there IS no title page? What's the second-best option? I know it's uncommon, but I've seen at least one recently with no title-page, and it didn't appear to be a damaged book. When I did some Duplicate ISBN checks recently, a lot of the problems were down to Series information being included, whereas Subtitle information is all we should be adding, IMO. Actually, I may leave that till later, I'm beginning to form a strong opinion against the use of '0000-00-00' for publications where you can definitely put an earliest possible date and printing number on instead, and at least separate the unknowns a bit more clearly. Although that ties in with a rant about how, if we recorded publishers more consistently, we could narrow down the dates even more... ah well, I'll see if I'm still this grumpy when I have time. :-/ BLongley 15:19, 3 May 2007 (CDT)
The thought of making the help longer bothers me. The hydra has many heads already.
Well, we are trying to model a (not so small) part of the real world, which is inherently fairly complex, so it makes our conventions unavoidably complex as well. On top of that, other parts of the Real World (tm), e.g. collectors, hava certain pre-existing rules and conventions (e.g. what is and is not a "first edition"), which we are also trying to follow, hence even more complexity.
It's not that OCLC bibliographers wanted to make their rules as complex as they currently are, it just kind of happened over time as more and more special cases were identified and accounted for. Of course, OCLC folks (and librarians in general) have much more time to spend on these kinds of rules than a motley crew of volunteers...
I agree though that additional help pages would assist in making the decision process more consistent in terms of how to deal with missing/conflicting data. For example, one I run across frequently are publications that have an old style publisher code on the cover, an SBN on the spine, and a pseudo-ISBN (It's titled "ISBN" but has the 3-digit price code appended like an SBN) on the copyright page. I usually enter the ISBN (without the price code thing) and lots of publication notes. Bill, I'm completely with you on the '0000-00-00'. Marc Kupper (talk) 19:30, 4 May 2007 (CDT)
If I understand the issue with 0000-00-00 correctly, it's a presentation problem, not a data one, right? We will still capture the same date information regardless of which convention we agree to follow, it's just a question of whether we decide to (a) use 0000-00-00 in the Date field and document what we know in the Notes field or (b) whether we use the less-than-100%-certain date in the date field while still documenting our sources/details in the Notes field, correct? If so, then I agree that we could relax the Date field standard as long as we tell our users exactly where the date information came from, e.g. Tuck, OCLC, Locus, etc. Ahasuerus 20:31, 4 May 2007 (CDT)
Many publishers seem to only state the first printing date and a printing number meaning 0000-00-00 becomes all too common. As ISFDB does not have First Printing and Printing # fields I've experimented with a hack of putting the First Printing in the YYYY-MM field and setting the day of the month to be the printing number. That gets the records sorted in the correct order for that publisher. Sometimes I'll have a probable printing date and will at least document the source and what the probably date is though I usually don't set the publication's Year field to this. One source of probable printing dates BTW is ISFDB itself as the publication records are often already dated with a something that seems probable but is not stated in the publication. In that case I note that the ISFDB record has YYYY-MM-DD but and while that seems reasonable given the cover price or other sources the date is not stated anywhere in the publication. Marc Kupper (talk) 03:35, 6 May 2007 (CDT)
Interesting idea with printing number instead of day... it even seems to work when the month is unknown. Doesn't work with my 37th printing of Harry Potter 1, though, it silently gets put back to 0000-00-00... :-/ Still, I think it helps if we allow people to enter the only date they can see if that's all the physical pub has. For a new title, this would probably get the title date correct at least rather than losing all that information: and entering the Copyright or First printing date if that's all there is, then modifying the pub date to '0000-00-00' seems more work than necessary for everybody, especially if the intention is to modify it back later at some point. I suspect we'll never be able to do that though, and I wouldn't want to be the person finding out if my 37th printing had already been entered if all I can see is three dozen '0000-00-00' versions in no particular order.... BLongley 08:05, 6 May 2007 (CDT)
Just thought of another example where I wouldn't want to sort out '0000-00-00' dates: The Wolves of Willoughby Chase keeps getting reissued with different covers under slightly different Red Fox imprints, and goes through several reprints of each one. Each Edition gets a new first printing date but the reprints only have that and the printing number. The price can go up or down depending on whether it's a luxury new edition or a cheap one. Yes, they change the ISBN but I'd rather have each edition grouped together than pick them out of a long list of '0000-00-00's. Not that it's a huge problem yet, few new people seem to bother with such thorny issues.BLongley 08:17, 6 May 2007 (CDT)
It does seem that adding printing number as a field in the database might be a good idea. Dana Carson 16:38, 6 May 2007 (CDT)
I forgot to note but to help find that 37th printing I have experimented with putting it in the publisher name field as in "DAW Books (37th printing)." Fortunately, it's a rare book that goes beyond 30 printings. I have contemplated, but never done, a hack-hack of just going to the next month. 1997-06-00 is the first printing, 1997-06-02 to 1997-06-29 are the 2nd to 29th printing, and 1997-07-01 would be the 31rst printing. The publication notes would need to be very clear that the publication states "First Printing, June 1997 / 31 32 33 34 35 36 37". I have not done it yet but suspect the wiki bibliography page for the first printing should be used to document all of the printings along with the ISFDB date that was assigned to each plus extra notes that may give hints on when it was published. All of the publications would reference that wiki page in the pub-notes rather than adding a long detailed note to each publication. Marc Kupper (talk) 14:56, 7 May 2007 (CDT)
I'd love to know what the record is for printing number. This pub tired me out, as that one actually RECORDS all previous pubs. I dread to think what it's up to now, 33 years later... I'm not going to buy another copy to find out though. If we can just agree that an "only date shown" can be kept as is, with notes about the printing number and any other date-related info kept, I'm happy for now. Long-term: yes, printing number should be a database field, IMO. BLongley 15:16, 7 May 2007 (CDT)
That one was pretty cool and there were even 28 dates listed. Personally I'd deal with it by using a publisher name of "Dell (28th printing)." One puzzle is the printing history goes back to Jan-1970 but if you look at title 2269 you will see a verified Sep-1965 Dell #1149 followed by your verified Dell Dell #1149 / 0-440-11149-8 at Nov-1973. Marc Kupper (talk) 12:36, 11 May 2007 (CDT)
Re-re-re-checked and MORE notes added. That edition doesn't give any details of the prior Dell edition that had the same number: presumably as they wouldn't fit on the copyright page. ;-) It's an excellent example of why I DON'T add "stub" publications for every book I have with prior publication info in, and leave them in notes. (Yes, I know I asked everybody about whether it was a good idea to do so, but I was only thinking about one or two publications when I asked. I still do if it's a small number and gives us a better first pub date.) BLongley 14:57, 11 May 2007 (CDT)

Christmas on Ganymede by Asimov (copyright dates)

I'm sure this is in the HELP files somewhere, but I never can seem to find what I'm looking for there. This short story in the db has a date of 1942 due to it's first publication in "Startling Stories, January 1942," issue published by Better Publications Inc. The book I have, "Christmas on Ganymede and Other Stories" ed Martin H. Greenberg, lists the copyright date for the short story as "1941, by Better Publications Inc." Should we change the date of the story to 1941 based on this evidence, or leave it as 1942? I remember reading somewhere about magazine dates and that most magazines are published several months before their cover dates, but I don't remember what the consensus, if any, was. CoachPaul 17:40, 4 May 2007 (CDT)

Yup, it can be confusing. As Help:Screen:NewNovel states:
Note also that the publication date does not always perfectly match the calendar date. For example, a January issue of a magazine is usually available in December of the previous year, and often earlier than that. Books with a January publication date may often be bought in the closing weeks of the prior year; they will show the later year's copyright date, even though that year has not yet started. In these cases, the convention is to use the official publication date rather than to try to identify when a book actually first became available. If there is a large discrepancy -- for example if a book was printed but unexpectedly delayed before release -- then this can be noted in the notes field.
The only time this becomes a real issue is when you are working with a reprint anthology which lists copyright dates only and when you can't check the original magazine. Caveat lector (and bibliographer)! Ahasuerus 18:28, 4 May 2007 (CDT)
The date in ISFDB would be 1942-01-00 as that's it's the stated first publication date for the magazine. I would add a title note with an explanation about the copyright date that's pretty much a copy/paste of what you already discovered so that if someone comes along wondering about 1941 vs. 1942 they can make an informed decision about what date they want to use. Marc Kupper (talk) 19:19, 4 May 2007 (CDT)

David Bischoff

The db, in the "Chrysalis 5" entry, lists the author of "All The Stage, A World", as David Bischoff. My copy of "Chrysalis 5" on the ToC lists the author as David F. Bischoff, and on the first page of the story, lists the title as Dave F. Bischoff. Do I use the name on the ToC, or the first page of the story? CoachPaul 14:48, 11 May 2007 (CDT)

More checking had David F. Bischoff on the copyright page, so I used that and put a note into the "Note" section of the record. CoachPaul 15:04, 11 May 2007 (CDT)
As per Help, first page of the story. "For short stories, essays and poems, take the title from the heading on the page where the work begins, rather than from the table of contents" BLongley 15:07, 11 May 2007 (CDT)
What about when this differs from the Copyright Page? And this isn't the Title but the Author. CoachPaul 15:10, 11 May 2007 (CDT)
Whoops, wrong quote: "Short stories and artwork. For short stories that appear in magazines and anthologies, the author's name should be taken from the story heading, rather than from the table of contents, if there is one." BLongley 10:59, 12 May 2007 (CDT)
Unfortunately, copyright notices describe who owns the rights to the story, not who wrote it, so we tend to shun them. With the recent proliferation of sharecropping, they have become particularly unreliable. Having said that, copyright information can be occasionally useful when authors use thinly disguised pseudonyms, but put their real name on the copyright page.
In Bischoff's case, he used "Dave" and "Dave F." during the first few years of his career, but then switched to "David F.". This story was presumably published during the transitional period. I would record the name used on the title page as per the Help text and then add a Note with detailed information to the Publication record. Ahasuerus 15:39, 11 May 2007 (CDT)
Thank you both for your help. CoachPaul 15:46, 11 May 2007 (CDT)

Is there some kind of problem with making Dave F a psuedonym? CoachPaul 09:46, 12 May 2007 (CDT)

Not really, I just left it for the moment until I could get back to find out what was decided here. It seems we've settled on "First page of story trumps ToC". (I wish the help was that succinct!) I'll approve the pseudonym, we've got plenty of notes now. BLongley 10:59, 12 May 2007 (CDT)

Changing a variant relationship

In Norman Spinrads Summary Bibliography the novel Your Name Shall be...Darkness(1964) is a variant of The Ersatz Ego(1970), the listing should be the other way around. How is this fixed? Kraang 19:16, 9 May 2007 (CDT)

It's a two step process at this time. First, you break the relationship by pulling up the current Variant Title, selecting "Make Variant Title" in the navigation bar and then entering "0" in the parent field. Once the submission is approved, the two titles become unlinked and you can submite a regular "Make Variant Title" submission. P.S. I have created Help:How to reverse a variant title relationship since this question has come up a few times and I don't think we have anything in the Help pages that addresses it. Ahasuerus 19:48, 9 May 2007 (CDT)

Also is there a simple way to leave messages on the main Help desk page and not the talk part of the Help desk?Kraang 19:16, 9 May 2007 (CDT)

When you have the main Help desk page displayed, click on "Edit" at the top of the page and then you can add a section at the bottom. Ahasuerus 19:48, 9 May 2007 (CDT)
I was looking for the plus button, ok i'll try that in the future.Kraang
That's one of the downsides to using the "ISFDB" namespace in Wiki page names: you don't get the "plus" button for some reason. I assume it's just another Eddorian conspiracy. Ahasuerus 20:16, 9 May 2007 (CDT)
Cool - I had seen the + button, wondered what it was long ago, and forgot all about it. I'll have to start using it. Marc Kupper (talk) 03:32, 10 May 2007 (CDT)
I can't find it in the MediaWiki help and don't want to waste time by searching in the source code, but I think/STR that the plus sign is intended ONLY for "talk" pages, so one can start a new thread easily. The main namespace is presumably mostly used for more structured collaboration, where the "add new section" button wouldn't be used that much.
However, you can force it by ading "" to the page - alas, only from the version 1.7 up, which ISFDB doesn't, and is not going to in foreseeable future, use. But there's a hack around it using direct URL, I'll add it. --JVjr 08:49, 11 May 2007 (CDT)
And instead of having to edit the whole page you can click the edit link for the last section and add your new section with the usual == title ==. This cam be needed on real long pages that not all browsers can fit in the edit box. Dana Carson 16:56, 11 May 2007 (CDT)

Best Way to Inform Moderators of Errors?

They usually catch them - as Blongley did when I changed a canonical name when I meant to update a legal name, but what is the best way to broadcast an "I screwed up and you need to know" error?

Post it here? :-) Or The Moderator Noticeboard maybe.
Mods aren't Gods, and can/do/will make mistakes too (the main advantage of being a Mod so far is that I can correct my past mistakes quietly without admitting them to anyone else!) but I'd suggest here first. But there is no "Attention, active Moderators!" option that I've found yet. If you spot who is doing more approvals than any other Mod at the time, then maybe their talk-page is the place to post it, but being a Moderator doesn't put you on call 24/7 (I hope!). It's not ideal - we should have a "withdraw submission" option for every editor in cases like that - but the Programmers are even shorter in supply than the Bureaucrats, and we do (not officially) have semi-moderators that stick to what they know and don't mess with what they don't. (I do experiment at times - I'd like to become a better, more useful, Mod, but I don't think anyone else should suffer from my experiments.)
I do see a drop in activity from the current Mods though, and some new ones wouldn't go amiss. So if there's a respected Editor that's been overlooked, talk to them. (Don't nominate them until you HAVE talked to them though.) BLongley 18:56, 12 May 2007 (CDT)
Right now there's no perfect way to get a moderator's attention. Usually when I'm working on the queue I'm not monitoring the wiki and checking Special:Recentchanges is lower on the queue. In fact, tonight I am catching up on several days of Recentchanges. I have thought about ISFDB improvements including allowing editors to put their own submissions on hold and/or to reject them. Being able to edit them is a lot more work. In the mean time, I've had people send me User:Talk messages and that works as I do pop over to the wiki from time to time. As for becoming a better moderator - that just takes practice. I've screwed up things many times and am really happy that Google caches stuff so that I can look at what the old (and correct) state of ISFDB was. Marc Kupper (talk) 00:25, 15 May 2007 (CDT)
... which happens to be the main reason why I keep the last 10+ ISFDB backups around :) Ahasuerus 00:32, 15 May 2007 (CDT)
Speaking of which, when do the rest of us get a new download available? I'm running out of "these books need entering!" options, and need a reminder of the "you edited this, why haven't you verified it?" pubs. Which I won't actually get from the backup, but a reminder that I've only verified 3 pubs from someone I own 20+ pubs from helps. BLongley 17:01, 16 May 2007 (CDT)
Al recently updated the Moderator Availability section of the Moderator Noticeboard with May 26 as the day when he may return to the realm of the living, so I would guess May 26-27 is the way to bet. It's not a terribly time consuming process, but Al deletes all passwords and e-mail addresses from the database before he can post the backup file, so it's not something that he can do with one keystroke. Ahasuerus 18:56, 16 May 2007 (CDT)
Personal tools