Could any of the recent changes/patches have somehow frozen the Top Verifiers page? It hasn't changed in at least 4-5 days. I check my numbers a couple of times a month to make sure my insurance is keeping abreast of my additions [which always happens in great leaps at this time of year]. Thanks! --~ Bill, Bluesman 05:16, 15 January 2014 (UTC)
- That list is supposed to be refreshed every night at 1am server time. I see that the script that updates the list was last run 22.5 hours ago, but let me check to make sure that it didn't error out... Ahasuerus 05:19, 15 January 2014 (UTC)
- After re-running the script manually, I see that your count of primary verifications is now 11,815 vs. 11,800 yesterday and the count of secondary verifications is 80,419 vs. 80,396. Similarly, my count, which I incremented 30 minutes ago by verifying a book that I had lying around, has gone up by one. In other words, everything seems to be working OK. Ahasuerus 05:57, 15 January 2014 (UTC)
- Thanks! --~ Bill, Bluesman 17:41, 15 January 2014 (UTC)
Request for another clean-up script
Is it possible to write a script which finds publication records which give the ISBN-10 for post-2007 books? (Similar to the one that finds ISBN-13 in pre-2005 records.) I know ISBN-13 was supposed to be fully implemented by January 2007, but am certain some small press and self-published books continued using ISBN-10 well into the next year or so. So perhaps give January 2008 as the start for the search? At that point all major publishers were fully compliant. I would use the advance search to find such records, but the ISBN field of the publication search form only accepts full numbers. Thanks for considering it. Mhhutchins 04:25, 16 January 2014 (UTC)
- Sure, shouldn't be too hard to do. Let me add it to the list and I'll try to squeeze it in in between the next two patches. Ahasuerus 04:29, 16 January 2014 (UTC)
And since you've got nothing else to do...
- Good point. I'll add it to the list, thanks... Ahasuerus 05:00, 16 January 2014 (UTC)
Author multi-web-page editing
I think there may be a problem with the new multi-web-page handling, at least for authors. Dirk submitted a fill-out for Raymond Williams that included two web pages. The moderator screen showed it as http://www.sf-encyclopedia.com/entry/williams_raymond+http://en.wikipedia.org/wiki/Raymond_Williams. I accepted it, but when I went to check the author's summary page, there was only Webpages: SFE displayed, and my browser showed that link was for that full compound-URL text, as shown in the moderator screen. I figured maybe Dirk had made a mistake, so I went and edited it. Sure enough, the compound text showed up as the sole Web Page 1 value. So I fixed it up, removing the second part and putting it into Web Page 2. The moderator screen for that submission highlighted Webpages as a diff, but the text looked identical, and when I accepted it, the author summary (and any subsequent edit attempt) shows the expected two separate links. I suppose it's possible he entered the pair as one value, with a plus sign, but I figured I'd mention it. --MartyD 16:31, 26 January 2014 (UTC)
- Checking the body of the submission, I see that it contains a plus sign:
<Webpages> <Webpage>http://www.sf-encyclopedia.com/entry/williams_raymond+http://en.wikipedia.org/wiki/Raymond_Williams</Webpage> </Webpages>
- I guess Dirk was so used to using semicolons that he switched to plus signs once semicolons were disallowed :) Ahasuerus 17:29, 26 January 2014 (UTC)
The Last Space Ship
Are you certain that the "Designed by" credit in this book refers to the cover art? I've personally never seen an interior credit (usually on the copyright page) for design meant anything more than the design of the book unless it explicitly mentions jacket art. The cover art credit is usually on the dustjacket itself, or in the case of softcover books, occasionally on the copyright page, but more often on the back cover. But it's usually explicitly stated as "Cover art" and not "Design". I had a discussion with another editor who had credited Solomon with cover art based on the same interior credit, and has since reverted the changes. Thanks. Mhhutchins 17:38, 27 January 2014 (UTC)
- That's a good point -- typically I don't enter "designers"' names in the Cover Artist field, but I must have slipped in this case. I'll go ahead and remove it. Thanks! Ahasuerus 18:21, 27 January 2014 (UTC)
Latest Fixer submissions
I like that you're now giving the ASIN in the Note field of publication records of ebooks (like this one), but wonder about the reason for entering Amazon's "Print Length" in the ISFDB record's "Page Count" field. We know that MOBI formatted ebook files (like Kindle) allow the user to set the page count. So perhaps providing the Amazon's data about the print length to the ISFDB record's Note field would be better way to record it (if we have to record it at all). Just a suggestion, or maybe one that needs further discussion. Thanks. Mhhutchins 20:58, 28 January 2014 (UTC)
- Well, I am still experimenting with Amazon's data, so things are not settled yet. There are different fields like "ASIN", "EISBN" and "EAN", which Fixer can access, and I am not 100% that I got all the relationships right, at least not yet. There are other complexities involved, e.g. Fixer's internal ASIN-ISBN crosswalks which need to be maintained to ensure that the same record is not submitted twice.
- As far as "Print Length" goes, that's another can of worms, e.g. sometimes Amazon's records say things like "page count from ISBN NNNNNNNNN", where NNNNNNNNN is the ISBN of some dead tree edition of the same book. On balance you are probably right that it's safer to record Amazon-provided page counts in the Note field. I'll see if I can make the change later today. Thanks. Ahasuerus 21:28, 28 January 2014 (UTC)
- Thanks for the adjustments. BTW, are you intentionally omitting the Amazon list price for the ebook editions? They're not included in the latest Fixer submissions and I'm not sure whether I should add the price to the record. Mhhutchins 19:22, 29 January 2014 (UTC)
- List prices are not available for ebooks, at least not via the interface that Fixer has access to :( Apparently Amazon's intent was to make so-called "offer" prices unavailable, but in the process they also zapped list prices. Their official announcement says:
- "Unfortunately at this time Offer Prices for Kindle Items are not supported by the Product Advertising API. List Prices are however, provided by the API."
- but no such luck. And, since other users have complained about this issue, I assume that it's a problem with Amazon's API rather than with Fixer's logic. Fixer should be able to get list prices in other ways, but it will take some time.
- Also, in many cases ebook prices appear to be inconsistent, e.g. consider our publication record for the ebook edition of The Long War. The value of the price field is "£6.99" because that's what Amazon UK shows, but if you pull up the same ISBN on W. H. Smith's site, the stated price is £9.99. Ahasuerus 19:54, 29 January 2014 (UTC)
- Another reason for bibliographers to hate ebooks! (I need to start a list.) Should I leave the field blank or record the current price? At least the record will give the Amazon price and indicate the date at that price. Mhhutchins 20:47, 29 January 2014 (UTC)
- I think it would be best to record the current Amazon price in Notes and indicate where it comes from. That's what I plan to do with Fixer once I enhance his logic to grab list prices. Ahasuerus 21:02, 29 January 2014 (UTC)
- BTW, here is something else that I realized earlier today after processing almost 500 ebook ISBNs. Normally, Fixer gets two records for an ISBN, one from Amazon.com and another from Amazon UK. They are clearly different records and can have different publishers, titles, authors, etc, which make the manual reconciliation process time-consuming. With recent ebooks, however, the records returned by Amazon.com and Amazon UK are always exactly the same, which suggests that Amazon has a separate "Kindle" database shared by Amazon.com and Amazon UK. It may also help explain why there is no list price -- since prices are country-specific, they wouldn't be available in a shared record and would have to be pulled from some other location at page generation time. Oh well, the adventure continues! :-) Ahasuerus 21:02, 29 January 2014 (UTC)
- With the internet as the chief source for ebooks (I hear you can go into some retail outlets and buy them, but even then it's through the internet), I was pretty sure that international barriers would make very little difference. Regardless ebooks are still being released that are restricted to a specific territory. Michael Bishop re-sold almost his entire back catalog to Orion Books in the UK, but those rights didn't cover distribution in the US. So I think they (Amazon) can still determine to where the book is being downloaded and base their prices on the currency of that location. Giving the prices for different territories in the record's Note field has always been allowed when stated in a printed publication, so I guess the same logic should apply to ebooks. I'll moderate the current priceless Fixer submissions without entering the price in the Price field. Thanks. Mhhutchins 22:40, 29 January 2014 (UTC)
- Another reason to hate ebooks: Scanning without doing any copyediting for format errors. (Just click on the "Look Inside" and behold the countless errors in just a few pages of the introduction.) Ridiculous to charge people for so little effort into "publishing" a book. Mhhutchins 22:48, 29 January 2014 (UTC)
- Ouch! That's pretty awful :-( However, I suspect that it has less to do with this book being an ebook than with the people who scanned and "edited" it. Every month Fixer finds thousands of new paperbacks churned out by CreateSpace/etc-empowered "publishers", many of them reprinting public domain books. And every month I send 90%+ of them to "Queue 2" or "Queue 3", which means that they won't get processed for a very very long time. The only reason why I don't do it automatically is that CreateSpace is also used by commercially published authors like Norman Spinrad and Lawrence Watt-Evans to bring their older books back into print or publish minor works that commercial publishers are not interested in. Unfortunately, even though the process is semi-automated, separating the wheat from the chaff is still very time-consuming... Ahasuerus 23:05, 29 January 2014 (UTC)
Planet Savers/Sword of Aldones
Does the number line in [this] edition actually have '1' in it? I just picked up what seems to be an identical copy but the number line is 2 4 6 8 0 9 7 5 3. Same catalog #/price/etc. Thanks for checking! [if it's buried, don't worry about it, same question to Don Erikson] --~ Bill, Bluesman 03:28, 30 January 2014 (UTC)
- Yup, it has a complete number line: "2 4 6 8 0 9 7 5 3 1". It also says "This Ace printing: April 1980". Ahasuerus 04:00, 30 January 2014 (UTC)
- Thanks! Mine is also missing the date statement. --~ Bill, Bluesman 04:45, 30 January 2014 (UTC)
Strange bug when varianting titles in numbered series
Look at this series. I varianted the last three to a canonical author (unknown in the last case) and look what happened to the numbering of the series. I corrected #7 but left the others for you to see. Mhhutchins 18:02, 30 January 2014 (UTC)
- Thanks, let me take a look... Ahasuerus 18:09, 30 January 2014 (UTC)
- OK, I think I got it fixed. I also corrected the 30 affected records, so we should be in better shape now. BTW, "None" is a special value in Python and has to be handled in a special way when filing data into the database. When a programmer forgets about this "special case", "None" can pop up in unexpected places, e.g. as a "storylen" value for omnibuses (which I am sure you have seen.) Ahasuerus 18:42, 30 January 2014 (UTC)
- It also appears in other title types. An advanced search now for "None" as STORYLEN found 50 titles, including NOVEL, POEM, SERIAL and INTERIORART (the most common), but no OMNIBUS-typed records. (That's because I was aware of that bugh and occasionally do a search for them. I had not realized that it can happen in other title types.) Mhhutchins 19:41, 30 January 2014 (UTC)
- Curiously, I ran the same search on the development server a few minutes after I had posted and found 42 records. I guess it means that 8 additional records have been "infected" since the development server was refreshed on Saturday morning. Hm... At one point I thought that I could recreate this bug by flipping publication type and title type back and forth in Edit Title/Edit Pub, but I'll have to try it again. Thanks. Ahasuerus 20:02, 30 January 2014 (UTC)
- Uh, I was estimating. There's no way to tell the exact number without counting them. As it turns out it was a pretty good estimate.
- I believe this happens when an editor accidentally enters a INTERIORART (or POEM, ESSAY, or SERIAL) content record as SHORTFICTION, and then updates the title records to correct the type. Because the system looks for a "length" in the SHORTFICTION record, it assign "None" to the newly-typed record. Of course, I could be wrong. Mhhutchins 20:15, 30 January 2014 (UTC)
- Well, maybe not. I tested several different permutations but couldn't get it show up. Back to the drawing board. Mhhutchins 20:19, 30 January 2014 (UTC)
When you get a chance, please read this message I left on another moderator's page concerning a submission-acceptance quirk that's happened at least twice in the past few weeks. Thanks. Mhhutchins 17:25, 31 January 2014 (UTC)
- That's right, it can happen when the server is particularly slow and the process times out during the approval process for whatever reason -- see Bug 495.
- One way to alleviate the problem would be to modify the approval logic so that it would change the submission status from "New" to "Approved" first and file the data into the database second. That way if the script stops half way through the approval process, the submission will be considered "Approved" and couldn't be processed a second time. It won't address the issue of partially approved submissions, but at least we won't have to deal with duplicates anymore. Ahasuerus 22:04, 31 January 2014 (UTC)
- That's good to know. How hard would the modification of the approval logic be? Mhhutchins 22:31, 31 January 2014 (UTC)
- Not hard, but I will need to review, change and test 27 scripts, so somewhat time-consuming. Ahasuerus 00:54, 1 February 2014 (UTC)
I accepted a few submissions that added artist image links using OpenLibrary.org. I thought they are an ok site to link to, but maybe that's just for covers? I found ISFDB:Moderator_noticeboard/Archive_09#Cover_images_from_OpenLibrary.org, but it's only discussing covers. More importantly, I realize we don't give hosting credits on displayed author images (e.g., here). --MartyD 20:01, 1 February 2014 (UTC)
- As I recall, author images have more stringent copyright standards than cover scans. I think the act of taking someone's picture creates a separate work protected by copyright while the act of scanning a cover doesn't (unless you do some kind of non-trivial rearrangement of pixels.) I am no expert, though, so perhaps my understanding is flawed or incomplete.
- As far as hosting credits go, that's a good point. Let me create a FR... Ahasuerus 03:58, 2 February 2014 (UTC)
- I'd like to see just how many of the photos on our author pages are hosted locally (most of which are probably in violation of copyright) or are improperly linked to non-permitting websites (which would be in violation of our policy about obtaining permission). I would bet money there are very few that avoid falling into either category. Mhhutchins 04:41, 2 February 2014 (UTC)
- Your wish is my command! :-)
- Total number of author URLs: 354
- Number of ISFDB-hosted URLs: 130
- Number of author images hosted by third parties: 224
- I'd have to write a new script to determine how many links are to sites that we don't have permission to link to. Ahasuerus 06:15, 2 February 2014 (UTC)
- Your wish is my command! :-)
Hi Ahasuerus. Just went to approve an update and when I clicked the "Current record link" I was given an error message. I did an update with the same series but a bogus parent and the link is also broken. I put them both on hold so you could see what's up.Kraang 04:12, 2 February 2014 (UTC)
- Did one other same problem.Kraang 04:14, 2 February 2014 (UTC)
- I did a test myself and found the same behavior. It should be linking back to "pe.cgi?#####". But instead, it links back to the non-existent "series.cgi?#####". Mhhutchins 04:33, 2 February 2014 (UTC)
- Good catch -- the bug was introduced about 10 days ago when the post-submission page for Series updates was changed to show the same table that the moderator approval page displays. Fixed now, thanks! Ahasuerus 04:47, 2 February 2014 (UTC)
- A minute later, it's displayed perfectly. Strange. Thanks anyway. Mhhutchins 18:59, 4 February 2014 (UTC)
- My bad -- I forgot to update the database definitions before installing the latest patch, so pretty much everything was broken for a couple of minutes. Sorry about that! Ahasuerus 19:01, 4 February 2014 (UTC)
HMH Books for Young Readers
Re this publisher: I've yet to find a pub that gives this as the publisher, using either the Amazon "Look Inside" or the OCLC record which gives publisher credit directly from the title page. This "imprint" is probably just used for marketing purposes (or just for Amazon), and not a true imprint or publisher name. Every book I've checked gives the publisher as "Houghton Mifflin Harcourt". I'll merge these publishers once you've seen this message. Thanks. Mhhutchins 20:52, 5 February 2014 (UTC)
- Done and done! Merge away! :) Ahasuerus 21:19, 5 February 2014 (UTC)
"You have new message" funkiness
I just saw the new new message indication for the first time, while I was looking at a title summary page. The entire screen (from the warning message down) was bolded and perhaps over-sized. I think perhaps a missing close font or close span or some such? This in IE 9. --MartyD 03:54, 11 February 2014 (UTC)
- Thanks, let me take a look... Ahasuerus 03:56, 11 February 2014 (UTC)
- Yup, it was a simple HTML error, which more modern browsers like IE11, Firefox and Google Chrome compensate for. And I should really start using the W3C validation service now that our headers have been cleaned up to the point where it is actually useful. Thanks for finding the critter! Ahasuerus 04:21, 11 February 2014 (UTC)
Tabbing on entry forms
While entering the latest issue of Locus this morning I noticed that there is some wonkiness when tabbing to get to the next open field or the next button after entering a second or third author. Hitting the tab brings me back to the top of the page. I'd never noticed this before. Mhhutchins 16:29, 13 February 2014 (UTC)
- Yes, it's a known issue with the "Add" buttons. I plan to get to it later this month once I finish processing Fixer's catch for March. 130 ISBNs to go :-) Ahasuerus 21:54, 13 February 2014 (UTC)
Link to submission XML in edit history?
Would it be possible to change the edit histories to link to the XML of the submission, in the same format that the submitted XML is displayed in? I'd often like to know what another editor has done to a title I've been working on. Or is there another way to see this information? Mike Christie (talk) 14:38, 15 February 2014 (UTC)
- That would be FR 43, submitted by one Mike Christie back in 2006 :-)
- Unfortunately, it's not as simple as it may sound because submissions do not capture the "before" and "after" states, so it requires a separate table to support the functionality. Al began working on it ca. 2007, but then his availability dropped to 0. There is a way to view author updates via "Recent Edits", e.g. this record, and there is an FR to make this history available on Bibliography pages, but the code is rather buggy -- see Bug 301. It's on my list of things to fix in the next 3-6 months and then I hope to add similar functionality to titles, pubs, etc. Ahasuerus 18:26, 15 February 2014 (UTC)
- Yes, I remember submitting that! I was actually thinking that just displaying the submissions XML, which would presumably be fairly easy, would be quite useful by itself. For example, if I create a pub and someone updates it, I can look at my XML and theirs and figure out what the change was. Not as handy as the original request, but (I was hoping) easy to implement. What do you think? Mike Christie (talk) 18:31, 15 February 2014 (UTC)
- I am afraid it's not so simple. Things like the submission type and record number(s) are currently hidden in the XML. There is no way of getting to them from individual pub/title/etc records short of parsing 1GB+ worth of submissions, which is not feasible in real time. We would have to add additional fields to the submission table and update dozens of scripts to populate them at approval time. In addition, the submission table is not a part of the publicly accessible backups while the "history" table is. Ahasuerus 20:05, 15 February 2014 (UTC)
- It's certainly an area that could use more TLC, but, unfortunately, there are only so many hours in the day :( Ahasuerus 20:13, 15 February 2014 (UTC)
- Yes, that's easier to do and has already been implemented for New Pubs when you view them via the moderator-only "Recent Integrations" list. At this time some of the required code is only accessible in the moderator area, but it will be changing soon: I am in the process of replacing the ugly XML dumps which we display post-submission with HTML tables, which requires moving some functions around. Once that has been done, it will be easy to make the bodies of recent submissions available to all editors. Ahasuerus 22:39, 15 February 2014 (UTC)
Help:How to verify data#Replacing a Primary Verifier
- Right, overriding primary verifications is no longer possible. I am not sure we can salvage anything in that section. Ahasuerus 18:14, 15 February 2014 (UTC)
Larry Niven's Rainbow Mars
Please see ISFDB:Community Portal#Larry Niven.27s Rainbow Mars. Thanks. -- JLaTondre (talk) 21:54, 15 February 2014 (UTC)
I had a thought about a feature for dealing with duplicate titles that might be more friendly for inexperienced editors and that also could help anyone dealing with newly entered anthologies. What about a check-for-duplicates function on a Title? And that could have all of the same modes and might even let you have aggressive mode when the author has > 1000, since now one side of the comparison is constrained to a single title. --MartyD 12:58, 23 February 2014 (UTC)
- Yes, that would be useful. Along similar lines, how about a Duplicate Finder for publications? It seems like it could be particularly useful when dealing with collections and anthologies. Ahasuerus 17:23, 23 February 2014 (UTC)
- Oh yes, even better! --MartyD 00:44, 24 February 2014 (UTC)
Unexpected behavior when varianting review records
Re this record: when I varianted the review record to the canonical author, the link to the reviewed title remained with the variant record. The parent record isn't linked to the reviewed title. Was that always the case, and I'm just now discovering it? Thanks. Mhhutchins 02:36, 1 March 2014 (UTC)
- Hm, I am not sure. Let's see... Checking the current version of the code and the history of this script, I don't see anything that would even attempt to link the newly created REVIEW title to the reviewed title. Sounds like a bug to me! Ahasuerus 03:21, 1 March 2014 (UTC)
- I'm thinking this was done on purpose to prevent a second review from being listed on the reviewed work's title record. Perhaps we could make the link appear on both the parent and variant of the review record, but only displayed once on the reviewed work's title record. (Similar to how series data is moved to the new parent record when a variant is made. The series data is displayed on both parent and variant, but the title only appears once on series list.) Thanks for checking into this. Mhhutchins 03:49, 1 March 2014 (UTC)
- It sounds reasonable, but let me experiment on the development server first... Ahasuerus 04:59, 1 March 2014 (UTC)
Fiona Patton's County
Plus sign as web page separator
One side effect I imagine I have noticed with the change from using ';' to '+' as the internal separator for the web page links is that the mod screens now get very, very wide (i.e., I'm frequently having to use the horizontal scrollbar within my full-screen browser window) because the text wrapping does not consider '+' a break. I assume it does consider ';' a break and used to wrap better, but I freely admit I may be imagining things.
What do you think about tweaking the mod screen display to replace '+' with br? --MartyD 12:06, 9 March 2014 (UTC)
- I noticed the same thing back in January:
- Display web pages and emails on separate lines in viewers.py. Currently they are separated with "+" signs, which make the differences hard to determine when there are many web pages.
- I hope to get to it in the foreseeable future, but between Fixer's submissions and software improvements I am a bit burned out right now and may need to take a few days off. Ahasuerus 15:36, 9 March 2014 (UTC)
- Done. Ahasuerus 17:21, 10 March 2014 (UTC)
- Just noticed it today on a submission with 5 or 6 web pages. Nice! Thanks! --MartyD 02:12, 16 March 2014 (UTC)
- Sure thing :) Ahasuerus 03:15, 16 March 2014 (UTC)
I don't know if you still Watch the Award Talk page (it seems it was just you, me, and Bill), but I added some notes there about things I've just done to the awards. (In addition to adding the 2011 World Fantasy Awards.) Chavey 20:13, 10 March 2014 (UTC)
- Yes, I keep an eye on it since I am working on awards right now. There are a couple of things in the pipeline that should hopefully improve things, but there is a lot more that needs to be done. Ahasuerus 02:53, 11 March 2014 (UTC)
"Untitled Awards" documentation
I've been filling out the Norma K. Hemming 2010 Awards (we had the winner, but not the extensive list of nominations). Four of the nominations that year have to be left as "untitled" awards -- three because they were online publications and one because it was a graphic novel. I found myself wanting to be able to justify why we didn't have those titles in the database, e.g. to tell a visitor that "No, it's not because we don't have something entered that we should, it's because these are outside our scope of coverage." I'm not sure of the best way to do that, and was hoping you might give a little thought to how this could be handled. For now, the the two authors ([here, here, and here), I added notes to the "Bibliographic Comments" on their Summary page. But that wouldn't work if they weren't already in the database. I considered a feature request to add a field for such data, but since it would be necessary so infrequently, that seems inappropriate. If you have a better idea, let me know. Chavey 23:42, 15 March 2014 (UTC)
- The first thing that comes to mind is that we need to enhance the "Award Bibliography" page to display the "award_title" field for non-title-based awards. They are already displayed on year-specific Award pages, but they don't appear on Award Bibliography pages. For example, "Proserpine When It Sizzles" appears on the 2010 Norma K. Hemming Award page, but not on Tansy Rayner Roberts' "Award Bibliography" page. That should be easy to change. Ahasuerus 00:26, 16 March 2014 (UTC)
- Done. How does it look? Ahasuerus 00:53, 16 March 2014 (UTC)
- That looks good. (And easier to do than the two award FR's I submitted today :-) ) Thanks. And thoughts on how to document "why this award isn't linked to anything?" Chavey 01:27, 16 March 2014 (UTC)
- I guess the easiest thing to do would be to display a note at the top of each Award Bibliography page explaining that title-based awards are linked to title records if and only if the title record is in the database. And that some titles may not be eligible for inclusion. Ahasuerus 01:35, 16 March 2014 (UTC)
- That could work. I wonder if it would be worth it to try to distinguish when such an "untitled award" was listed because we didn't have the data (or time) to add the actual title vs. when it's "untitled" because it's outside our scope of eligibility? Chavey 04:00, 16 March 2014 (UTC)
- Great, thanks! :) Ahasuerus 23:53, 20 March 2014 (UTC)
thanks for approving the Hungarian Hesse editions and attaching them to the correct original German titles. I would have a comment for Hesse's author page - is seems that all of his novels assigned to the novel category not considering if it fits to the sf/fantasy/horror genre or not. As I remember mainstream novels should be classified as nongenre or something like that... I read only the Peter Camenzind and the Klingsor... both of them mainstream novels - not sf/fantasy, and I am not sure if the others can be classified as fantasy or not. maybe the author's page would require some cleanup in this sense - of course I would not delete any entry because Hesse is really an important author, I agree.
GaborLajos 09:35, 21 March 2014 (UTC)
- That's a very good point. Although Hesse wrote a fair amount of fantasy, e.g. see Pictor's Metamorphoses and Other Fantasies, much of his work is not SF. Let me post on the Community Portal and see if we have any Hesse experts who could help sort it out. Thanks! Ahasuerus 14:10, 21 March 2014 (UTC)
Publications with Suspect Images
Two requests with respect to the "Publications with Suspect Images" page:
- Could you add a bit more vertical padding between the rows? I've missed the correct "Click to Resolve" link a couple of times.
- Upon clicking "Click to Resolve", can you dispense with the "Suspect URL resolved" page and simply return to the "Publications with Suspect Images" page? There is a lot of clicking associated with fixing these and my mouse finger would really like the break. :-)
- Yes, it is certainly doable. The second change will require adding a server-side redirect, which is currently handled as a special case and needs to be generalized. Ahasuerus 00:41, 24 March 2014 (UTC)
- Change #1 has been implemented. Change #2 will be done tomorrow -- the "server-side redirect" enhancement is already in place. Ahasuerus 04:41, 24 March 2014 (UTC)
- Done! Ahasuerus 16:07, 24 March 2014 (UTC)
- Thank you. It's much appreciated. Mhhutchins 14:08, 24 March 2014 (UTC)
A particular [record] appears on [this] bibliographic page for Moorcock's The Twilight Man twice. Both records have the same number. I tried to clone the record and then delete the two but ended up with two anyway. Then I created a new record, imported the contents and deleted the [now] two clones and ended up with another two identical records. Glad I don't have any hair left to pull out. Do you have a magic wand or do we have a gremlin just playing games .... ? --~ Bill, Bluesman 15:20, 24 March 2014 (UTC)
- Curious. Let me take a look... Ahasuerus 16:47, 24 March 2014 (UTC)
- I suppose we can change the software so that the Title page would only display each Publication record once, but we'll need to think about it -- there may be other unusual cases where it may not be desirable. Ahasuerus 16:53, 24 March 2014 (UTC)
- Thanks! --~ Bill, Bluesman 17:15, 24 March 2014 (UTC)
- The problem is exacerbated by the fact that SF Reprise 1 is actually a re-binding of two issues of New Worlds (#144 and #145)each retaining their own pagination, thus the serial is presented as two separate titles. Would that make this publication an omnibus rather than an anthology? It's a sticky situation. This seems to be a rare kind of publication. I don't believe changing it to an omnibus would correct the problem. I'm not sure if we should change the software to suppress the display of a title appearing twice in the same publication record. In the past, it was a good indication of an extra editor record. That doesn't happen now because of software improvements, but if it did we have a clean-up script that finds them. (I check that script every time I run the others, but it's been years since a record with that problem popped up.) Mhhutchins 18:21, 24 March 2014 (UTC)
Fixer and Amazon UK
Tonight I've been doing a bunch of the Fixer May 2014 submissions of ebooks from Amazon UK, and I've noticed a few things about the submissions and prices, just FWIW:
- Most of the submissions have no price.
- Amazon UK does provide prices, although it's not clear whether they are list or not. I think they might be, but it's hard to tell (the publisher sites don't seem to be giving prices, at least not on the ones I checked).
- Sometimes Fixer is getting a price from B&N, but those are US prices. I'm not sure they're appropriate, especially when you go to the data source -- Amazon UK -- and find a price in pounds instead of dollars.
I decided to go back and put the price from Amazon in the ones that had no price, and I changed one of the ones that had a $3.99 price from B&N to the £2.99 from Amazon (and adjusted the note to record the B&N price). I'm not sure what's best, and I'm not asking you to do anything, I'm just passing along the observations. --MartyD 01:49, 25 March 2014 (UTC)
- Good observations, thanks!
- Unfortunately, I have been unable to find a viable source of UK prices. Amazon UK is as unreliable as Amazon.com when it comes to prices, e.g. they want £10.74 for Unwrapped Sky by Rjurik Davidson, but the list price as given by Pickabook is £10.99. Pickabook is not a bad source, but the site is very slow, which limits its usefulness.
- The use of B&N prices for UK books is a bug and I plan to fix it shortly. Ahasuerus 19:04, 25 March 2014 (UTC)
- P.S. Having said that, there are cases when Amazon's prices are valid. For example, they own a variety of imprints like 47North, AmazonEncore, AmazonCrossing and Montlake Romance. Pretty much everything published by these imprints is only available through Amazon, so the price that Amazon shows is the list price. Ahasuerus 19:13, 25 March 2014 (UTC)
Mills & Boon [/ Harlequin]
If we believe the Mills & Boon page, we should be entering all recent Mills & Boon things as published by Mills & Boon / Harlequin. If that's true, two things: Would you convince Fixer to record its new findings that way? --MartyD 02:21, 25 March 2014 (UTC)
- Good point, done. Ahasuerus 19:27, 25 March 2014 (UTC)
And could you zap anything with date > 1971 and publisher ID 38559 to have publisher ID 1779 instead? I edited some, but the 2013 and 2104 lists are large and would take a lot of edits. Thanks. --MartyD 02:21, 25 March 2014 (UTC)
- In this case we are looking at 30-40 edits and I am afraid that my "ad hoc script" threshold is much higher than that. I may be paranoid, but I have seen a number of cases where an ad hoc script ran amok and messed things up so badly that the only way to recover was to go back to the backups. Imagine typing "update pubs set publisher_id = 1779" and entering a semicolon before you realize that you meant to add "where publisher_id=38559". I have actually done something very similar a few times, but only on the development server... Ahasuerus 19:27, 25 March 2014 (UTC)
- 2013 & 2014 have all been changed to "Mills & Boon / Harlequin"Kraang 01:45, 26 March 2014 (UTC)
- Thanks! :-) Ahasuerus 01:47, 26 March 2014 (UTC)
Hi, would you agree that Last Orders forms part of the 'Masters of Science Fiction' Series? And, as a bonus: Would you agree that the covert art is the same as on Hegira?--Dirk P Broer 18:50, 3 April 2014 (UTC)
- Yes and yes :) Ahasuerus 19:47, 3 April 2014 (UTC)
The Glory Game
We seem to have several records for this title in hardcover, one of which does state First Edition,  and two others for later printings. All have gutter codes noted. Perhaps one matches [this] record?? Cheers! --~ Bill, Bluesman 04:01, 8 April 2014 (UTC)
- Ah yes, I remember this issue. The problem here is that the way my copy was glued together, the gutter code is all but illegible. You can sort of figure out that the first character is an "O", but the second character is hopeless. And since we have two pub records with two different gutter codes starting with an "O", there is no telling which one my copy matches. I thought I had updated the record with this information a couple of years ago, but either I forgot or my changes got subsequently overwritten.
- Have you heard from the other verifier, Bob, re: his copy? Perhaps he will have better luck with the gutter code, in which case we can delete the problem record once he moves his verification to the right pub. Ahasuerus 04:26, 8 April 2014 (UTC)
Linking to/from SFE
I've had a scripting enquiry from Dave Langford, and I've told him I'll pass it on to you:
- I've been thinking that we should add more links from SFE entries to ISFDB. Initially we linked only from artist entries, since ISFDB usually has some artwork examples and we hadn't then invented the Picture Gallery feature. But reciprocal links to all ISFDB entries that link to us would surely be a Good Thing.
- Not knowing my way about the API at ISFDB, I have no idea whether this is possible, but might there be a way to generate a table of URLs for ISFDB pages that include SFE3 links, along with said links? I could write a scraping script to do something of the sort, but that feels like bad manners. And doing it the "legitimate" way by e.g. hand-searching for all SFE3 authors in turn and hand-copying the URLs would take a long time.
I think this is what Dave was discussing with Bill Longley last November, a discussion which obviously went no further. His e-mail is drl (at) ansible (dot) co (dot) uk … no need to reply via me. Good luck with this! PeteYoung 11:53, 14 April 2014 (UTC)
- Thanks, I'll take a look :) Ahasuerus 15:40, 14 April 2014 (UTC)
I have two submissions sitting in the queue (left intentionally for your perusal). The first shows the submitter as "None", and the second is correctly attributed to me. Is this a bug in the "LinkAward" function? I noticed it earlier this morning, but it may have been there all along. Mhhutchins 19:48, 21 April 2014 (UTC)
- Most curious. Let me take a look... Ahasuerus 20:09, 21 April 2014 (UTC)
- Confirmed as a Link Award-specific bug, now to figure out what causes it. Ahasuerus 20:34, 21 April 2014 (UTC)
- Hulk smash bug! :) Ahasuerus 00:31, 22 April 2014 (UTC)
Author Award Listing Error
- Thanks, looking... Ahasuerus 01:20, 24 April 2014 (UTC)
- Fixed! Ahasuerus 01:35, 24 April 2014 (UTC)
- Darrah is probably as close to an awards guru as it gets. However, before we ask him (or start asking for volunteers), let me finish the next round of award changes which will implement a new Web page to add Award Types... Ahasuerus 23:09, 27 April 2014 (UTC)
- I did not have those on my radar -- thanks for mentioning them. Right now, the focus is on bringing existing awards up to date, which will probably keep me and others busy at least until Ahaseurus has implemented more convenient ways to add new awards. I've been keeping a list of the awards we should eventually add at the Awards Talk page, and I've added this suggestion to the list. If you look at that list, though, you might be surprised to see how many awards we need to add to get to where we want to be. But we certainly hope that we get there, and the Parsecs are certainly an award that we should have included. Chavey 03:25, 28 April 2014 (UTC)
- Thanks for looking into this! :) Ahasuerus 14:06, 28 April 2014 (UTC)
Greatheart Silver: Collection vs. Novel
If there's a consensus among the verifiers, I'd like to change Greatheart Silver from a Collection to a Novel, per [suggestion]. Would you please respond affirmatively or negatively there? I am leaving this note for all verifiers. Thanks. --MartyD 16:07, 14 May 2014 (UTC)
SFE3 Link Script
Dirk P Broer has been adding SFE3 links to author pages. As I've approved his edits, I've been checking the script's page to click the appropriate resolved link. I think I've found an issue with the "Authors with Missing SFE3 URLs" script. He's found matching links which don't show on the script page. Recent examples include Cherie Priest, Richard Owen, and Neil Oram. In each of these cases, there is a direct match with our page name and the SFE3 page name, but yet the author was not listed on the script page. The "Previous versions of this entry" dates on the SFE3 pages indicate these aren't new pages.
- Dirk has been given access to the script to find missing links and to remove them from the list once they're linked ("resolved"). Mhhutchins 00:40, 17 May 2014 (UTC)
- Thanks for chiming in, Michael! :-) Ahasuerus 01:33, 17 May 2014 (UTC)
My wife mentioned that she went to the ISFDB to try to find books on time travel and didn't have much luck. She didn't know about tags (which aren't that helpful anyway). --MartyD 14:41, 26 May 2014 (UTC)
- Once you showed her "Tag Search", did she not find "Titles marked with tag time travel" helpful? Ahasuerus 18:04, 26 May 2014 (UTC)
She did find advanced search and tried to search Notes, which of course didn't work out for her. In trying to help her, I realized we have no way to search synopses.... --MartyD 14:41, 26 May 2014 (UTC)
- Yes, but I don't know if this would be helpful in cases such as this one, because I do suspect that you would have an even bigger bunch of possible phrases to search for. What we would need is a greater percentage of tagged titles, and maybe some kind of parent tag: perhaps 'time' as parent for 'time travel', 'time manipulation' and maybe even 'alternate history'. Or did you have something else in mind for the synopses? Stonecreek 16:57, 26 May 2014 (UTC)
- The tag search was for "time-travel", unfortunately. --MartyD 22:20, 26 May 2014 (UTC)
Open Road Media
Fixer is adding a bunch of Open Road Media publications. I've looked at a number of these via Amazon Look Inside and the publisher is listed on the title page as Open Road Integrated Media. I believe the publisher should be renamed and Fixer modified to replace Open Road Media with Open Road Integrated Media. So far, there is only verified pub by this publisher, but I'm not sure the verifier is active. Thanks. -- JLaTondre (talk) 13:08, 11 June 2014 (UTC)
- Publisher name changed, verifier notified, Fixer adjusted. Thanks! Ahasuerus 19:13, 11 June 2014 (UTC)
- A question about the publication series which some of these records are being entered: where did this info come from? It's not apparent in the Amazon Look Inside. Mhhutchins 14:19, 12 June 2014 (UTC)
- These records are listed by Amazon as published by "Open Road Media Sci-Fi & Fantasy", "Open Road Media Teen & Tween", etc. When Fixer finds these records, he automatically converts the value of the "publisher" field to "Open Road Integrated Media" and moves the Amazon-provided value to the "publication series".
- I am trying to recall whether this logic was based on:
- what Look Inside showed for some ISBNs, or
- an assumption that Amazon's data comes from the publisher and thus reflects the publisher's categorization
- but I am drawing a blank. (Memory is the second thing to go :-()
- If Fixer's assumption is invalid, then I can creates a script to delete these values. Ahasuerus 18:23, 12 June 2014 (UTC)
- I'm not certain that it's entirely invalid, because I've seen such cases on Amazon's listings. But I don't believe one can assume that this is always valid. Thanks. Mhhutchins 19:15, 12 June 2014 (UTC)
- OK, I have changed Fixer's logic so that he wouldn't be populating the Publication Series field any more -- see the latest batch. How does that sound? Ahasuerus 19:38, 12 June 2014 (UTC)
- Sounds fine. Thanks. Mhhutchins 22:12, 12 June 2014 (UTC)
(unindent) FYI, this one came through with publisher "Open Road Media Romance", but the Look Inside shows the same Open Road Integrated Media as do all of the others. --MartyD 10:29, 13 June 2014 (UTC)
- Thanks, I have tweaked the pattern match. Ahasuerus 14:00, 13 June 2014 (UTC)
There are two Ace editions of the above title that are virtually identical. this and this. The differences are in the notes for the first link. I don't know which copy you have .... Cheers! --~ Bill, Bluesman 01:32, 15 June 2014 (UTC)
- Let's see... Mine is 1977 and correctly verified. Nice detective work! :-) Ahasuerus 03:31, 15 June 2014 (UTC)
Script finding publications with suspect images
Since clearing the several hundred items found by this script several months ago, not one additional publication as come up. Does it have be rewritten to find current suspect pubs or have the Amazon links become considerably more stable? Mhhutchins 19:33, 18 June 2014 (UTC)
- This list has to be regenerated every so often since, unlike other cleanup lists, it can't be maintained dynamically or even nightly. Rebuilding it is a time-consuming process because we have 125,000 links to Amazon and we can only check so many images per minute due to Amazon-imposed constraints. The current plan is to rebuild the list twice a year -- once I am done entering Fixer's data for August, I will start the process again. Ahasuerus 19:46, 18 June 2014 (UTC)
- No rush. I was just concerned that we'd gotten so good at linking images that my job was at jeopardy. :) Mhhutchins 20:37, 18 June 2014 (UTC)
- I am almost positive you have nothing to worry about :-) Ahasuerus 21:06, 18 June 2014 (UTC)
Is it possible to write a script that could find "stray publications", i.e. titles which have been varianted, but the author has yet to be made into a pseudonym? Thanks for considering it. Mhhutchins 20:25, 18 June 2014 (UTC)
- Well, anything is possible, but it will likely be a resource hog and may affect user-experienced response time. Which reminds me that I was going to move some of the cleanup scripts to the nightly job so that they wouldn't affect our users. Let me start by adding a simple cleanup script, which I have been meaning to implement for some time, to the nightly job and see how it goes.
- In the meantime, we can try to define "stray publication" titles for the purposes of the proposed cleanup script. Presumably it should find variant titles none of whose authors are pseudonyms, right? I guess there may be some oddball ghosted titles, but let me try this logic and see where it takes us... Ahasuerus 20:57, 18 June 2014 (UTC)
- Thanks. But only if you get a chance in the middle of all your other chores. And if no one has mentioned it lately (or at all), I really appreciate all that you've done in the development of the database. Having nil experience at it, I can only imagine the time and effort it must take in the writing, testing, etc. of the software changes. Mhhutchins 21:33, 18 June 2014 (UTC)
- Appreciate the kind words! As they say, we all do our part :-) Ahasuerus 21:55, 18 June 2014 (UTC)
Re-visiting M.Z.B.'s "Fall of Atlantis"
Last February we began a conversation about Marion Zimmer Bradley's Web of Light / Web of Darkness / The Fall of Atlantis. I finally pulled all three books to compare and, I believe, have (nearly) finished that conversation. Please take a look here. Chavey 06:08, 19 June 2014 (UTC)
Time's Last Gift
Hello. After adding a note and uploading a new cover scan to your verified, I wondered whether the art credit, which is at the present time “Darrell K. Sweet”, shouldn't be changed to “Darrell Sweet”, as this is the way it appears on the copyright page. Thanks for your opinion about this. Linguist 09:03, 20 June 2014 (UTC).
- My copy is currently in a hard-to-reach place, so I will happily go along with whatever the other verifiers decide to do :) Ahasuerus 14:35, 20 June 2014 (UTC)
- OK, thanks for responding. Linguist 15:25, 20 June 2014 (UTC).
- Linguïst is right. I corrected the artist (and varianted the cover to “Darrell K. Sweet”). --Willem H. 19:56, 20 June 2014 (UTC)
- Thanks! Ahasuerus 20:12, 20 June 2014 (UTC)
- Ditto. And as far as the tréma (aka diaeresis) on Linguist is concerned, I'm afraid it is superfluous. Thanks all the same ;-) ! Linguist 21:12, 20 June 2014 (UTC).
Deleting Award Categories
- You have to have at least 10 contributions of a given type to appear on the list :-) Ahasuerus 14:10, 20 June 2014 (UTC)
- Thanks for the explanation. Mhhutchins 15:25, 20 June 2014 (UTC)
Merging award categories
I've noticed that several awards have arbitrarily titled a category as "Best X" one year and the next year the same category would just be "X". (See the Analog Awards.) Is it possible to merge two categories under one name? Mhhutchins 03:33, 21 June 2014 (UTC)
- Yes, it's possible to code a "Merge Award Categories" option, but I am not sure how often it will be used once the initial cleanup has been completed. Perhaps wait a month or two and re-evaluate? Ahasuerus 06:35, 21 June 2014 (UTC)
- In the meantime, I'm doing manual edits to bring each of them together, especially in the major awards. The Nebulas were all over the place when it came to the titles of category. I'm not sure what they are "officially", but I don't think they should vary from year to year. There are some specialized and obviously unique categories which I've left alone, e.g. in the Analog awards, I'm not touching "Best Cover Artist" as it was labeled one year, while the others are given as "Best Cover". Thanks. Mhhutchins 16:59, 21 June 2014 (UTC)
- Excellent! That's exactly what I hoped would happen once we had the tools in place :-)
- As far as minor variations in category names go, I think once we determine that they are real and not just data entry artifacts, we can document them in Award Category Notes, which I plan to add shortly. Ahasuerus 22:17, 21 June 2014 (UTC)
House of Shadows
- Thanks for the heads-up! Unfortunately, the way the ISFDB database and application are configured, "é" and "e" are used interchangeable for searching and some other purposes. For example, if you were to enter a new book by "André Norton", it would be filed under "Andre Norton" because the software would associate the new titles with the existing author record. We have similar problems with "ï"/"i" and a few other character pairs. It's on our list of things to fix. Ahasuerus 22:30, 21 June 2014 (UTC)
- One more thing I've learnt. Thanks for the explanation. --AliHarlow 06:09, 22 June 2014 (UTC)
Coverart title / Cover Artist modification bug
I just ran into an interesting problem. We had pub X with cover artist A. For the resulting coverart title C:XA, two variants were made, C:XA-1 and C:XA-2. An editor then went and modified pub X to have cover artist A'. When that was accepted, the system deleted C:XA, leaving C:XA-1 and C:XA-2 referring to a non-existing parent title record.
If you're working with a back-up prior to today's, the pub ID is 60033, the original parent title ID was 149861, and the two variant title IDs are 1087005 and 1359712. --MartyD 16:22, 22 June 2014 (UTC)
- Thanks, I will take a look... Ahasuerus 19:20, 22 June 2014 (UTC)
Fixer and Mills & Boon
You might want to look through the Fixer submissions I have on hold. Independent of my question about Mills & Boon, there are quite a few "Mills & Boon" variations represented among them. I have not yet gone to check each Look Inside to see if they're all credited the same way, but on the few I looked at, I see no sign of the secondary series/imprint-ish information that's getting included in the publisher names. --MartyD 12:02, 24 June 2014 (UTC)
- Let's see...
- "Mills & Boon M&B" is clearly "Mills & Boon" and can be easily fixed within Fixer's logic.
- "Mills & Boon Nocturne Bites" appears to be the Mills & Boon version of Harlequin Nocturne Bites, up to and including the original Harlequin cover. The only difference is the ISBN, which is not available in the US, presumably for legal/business reasons. I guess Fixer should change the publisher to "Mills & Boon / Harlequin" and the publication series to "Harlequin Nocturne Bites".
- The "Mills & Boon Nocturne" and "Mills & Boon Silhouette" pubs have "Silhouette Bombshell" on the covers, so they are presumably a part of the Silhouette Bombshell publication series from Silhouette, a Harlequin Mills & Boon Ltd. imprint. Our record states that Silhouette was discontinued in 2010, but it looks like it has been revived.
- Let me see if I can get a larger sample before I make any further changes to Fixer's logic. (And thanks for reporting these issue!) Ahasuerus 15:08, 24 June 2014 (UTC)
- After reviewing a bunch of M&B ISBNs, I think the best Fixer can do is submit them as "Mills & Boon / Harlequin" (subject to whatever changes we may agree upon based on the current Moderator Noticeboard discussion) and leave the Publication Series field blank. Better safe than sorry... Ahasuerus 18:14, 25 June 2014 (UTC)
Digital audio player
This record was submitted by Fixer with a pub format of ebook. The linked photo (and price!) prompted me to go the the Amazon page and learn that it should have been entered as digital audio player. Is Fixer able to distinguish this format from other digital formats? Thanks. Mhhutchins 16:45, 25 June 2014 (UTC)
- Hm, let me take a look... OK, I see what's going on. Amazon's API reports the format of some e-books as "Digital". As a result, Fixer was confusing "Preloaded Digital Audio Player", which should be submitted as "digital audio player", with "Digital", which should be submitted as "ebook". All fixed now, thanks! Ahasuerus 18:51, 25 June 2014 (UTC)
- Thanks for looking and for fixing it. Mhhutchins 19:49, 25 June 2014 (UTC)
Make Variant Form Validation: Possible Feature Request?
Could you have a look at this short discussion and weigh in on whether my suggestion seems plausible and/or necessary? If so I can submit a feature request. Thanks, Albinoflea 05:53, 27 June 2014 (UTC)
Minor Interface Tweak Suggestion
In Publication Series tables and Publication Search Results tables, pubs without pages numbers are shown with '?' in that field. Since we don't list pages for ebooks, the question marks could be dropped when it's an ebook format. Obviously, minor so understand if you never get to it. -- JLaTondre (talk) 14:19, 29 June 2014 (UTC)
- Back in the day, when our data was often incomplete, Al decided to display question marks for blank values so that editors would know that a record needed to be updated. Eventually our data improved so much that blank values were more often than not legitimate and I changed the Publication display logic to eliminate question marks. However, I missed the Pub Series page and the Publication Search Results page, which have now been fixed. Thanks for reporting it! Ahasuerus 20:15, 29 June 2014 (UTC)
- Originally it was "Verified", but that contributed to the previously discussed spacing problem, so the last three letters got trucidated :) Ahasuerus 20:15, 29 June 2014 (UTC)
Heroes and Horrors
- Thanks! :) Ahasuerus 14:25, 30 June 2014 (UTC)
I added three feature request tickets (627, 628, 629) and one bug ticket (518) to the Sourceforge lists. Hopefully they are pretty self explanatory, but figured, since they went in as "Anonymous", that I'd let you know I was the submitter in case there were questions. Thanks. -- JLaTondre (talk) 17:19, 8 July 2014 (UTC)
- Thanks, I was wondering about that :) Ahasuerus 17:28, 8 July 2014 (UTC)
The unrecognizable character ï
INTERIORART records with a "StoryLen" of "sf"
An advanced search found more than 400 records typed as INTERIORART which have a story length. I was prompted to run this search when an editor made a submission to change a work of interior art which was mistakenly entered as SHORTFICTION. I noticed that "shortfiction" remained in the story length field of the work's title record. Is there a way to clean this systematically, or will I have do it manually. Thanks. Mhhutchins 22:12, 16 July 2014 (UTC)
- Hm... Perhaps we should take a step back and consider the big picture. At the moment, the "title type" field and the "story length" field are independent. You can enter any value in the former and it won't affect the latter. In a way, it makes sense because the "story length" field is currently used for things like "jvn", "/1,2,3" and "nvz" as well as for "sf", "ss", "nt" and "nv". Eventually, we'd like to move "jvn" and other non-length values to separate fields, but for now it will have to do.
- That said, there is nothing preventing us from changing the software to blank out the value of the "story length" field if the title type is not SHORTFICTION and the "story length" value is "sf", "ss", "nt" or "nv". The programming will be a little tricky because the logic behind New/Edit/Add/Clone Pub is convoluted, but it's certainly doable. What do you think?
- And just as a point of reference, here is how many non-SHORTFICTION records with "sf/ss/nt/nv" story length values we currently have:
- sf - 1,267
- ss - 208
- nt - 40
- nv - 136
- Ahasuerus 02:45, 17 July 2014 (UTC)
- Whoah! That's a lot more than I thought, but then I was only checking one non-SHORTFICTION type. If it can done programmatically without doing any harm to the records, then I so it's better than manually editing 1500+ records! Thanks. Mhhutchins 15:39, 17 July 2014 (UTC)
- After sleeping on it, it occurs to me that the underlying question is whether it's safe to assume that all non-SHORTFICTION titles with "sf/ss/nt/nv" storylen values should have the latter zapped. Specifically, if a NOVEL title says "nv" in the storylen field, can we safely assume that the "NOVEL" part is accurate and the "nv" part is not? Or would it be safer to do a manual review, at least of the 176 titles with "nv/nt" (and possibly "ss") values? Ahasuerus 19:29, 17 July 2014 (UTC)
- I have seen NOVEL-typed titles with a novella length, most of which are serialized "complete novels" and Ace Double-sized works. I'm not sure that it's necessary to keep that even if the work is less-than-novel length, but I will concede that some editors might want to retain that data. Mhhutchins 21:51, 17 July 2014 (UTC)
- I would argue that this kind of information belongs in the Note field. Ahasuerus 23:09, 17 July 2014 (UTC)
Clean-up script request: publess uncredited works
Is it possible to write a script which can find title records credited to "uncredited" and are publess. I've ran across some randomly and deleted them. They seem to arise when an editor removes them from records after discovering entry errors and doesn't follow up by deleting them from the db. These publess title records are easier found for authors with relatively few records, but would take days on end to find for this "author". We have a similar script which looks for publess reviews and interviews. Can the criteria be adapted for this new script. Thanks for considering it. Mhhutchins 22:20, 16 July 2014 (UTC)
- It is certainly possible, but it occurs to me that there are other types of pub-less titles that we may want to be able to find. Perhaps a more comprehensive script is in order. Let me sleep on it and hopefully I will have some ideas tomorrow morning. Ahasuerus 03:15, 17 July 2014 (UTC)
- There are many publess records which should remain in the database, especially any fiction types and ESSAY records. I've often entered publess titles into the database because the publication that contained them were non-genre. I don't know if other editors are familiar with the "trick" to do that. I'm afraid that a clean-up script to find all of them will result in some good ones being deleted offhand, because without their being contained in a pub, there is no back-up warning. I'd rather have several separate and more narrowly defined scripts than a single comprehensive one. Of course, that's your decision. Mhhutchins 15:44, 17 July 2014 (UTC)
- It's certainly true that there are different types of "publess titles" and many of them shouldn't be deleted. In some cases, we didn't have enough information to create pubs for them at the time when the data was entered, but perhaps things have changed and we have more information available to us now. For example, I see that someone (perhaps you?) has entered publess VTs for foreign language translations of Michael Bishop's works like this Japanese translation or this Spanish one. In some cases, WorldCat has enough data to create publication records for them, e.g. see this WorldCat record. A cleanup script would make it easy (well, easier) to find these types of problems and fill in many lacunae. (As an aside, I generally tend to create publication records for non-genre pubs and add a note that only known SF works are listed, but that's a Rules and Standards issue.)
- Before we continue this discussion, let me run a couple of scripts to see what kinds of numbers we are looking at... OK, here is a breakdown of the 3,706 publess titles that we had as of the last backup:
+--------------+----------+ | title type | count | +--------------+----------+ | ANTHOLOGY | 37 | | COLLECTION | 35 | | COVERART | 60 | | INTERIORART | 230 | | EDITOR | 168 | | ESSAY | 513 | | NOVEL | 804 | | NONFICTION | 244 | | NONGENRE | 93 | | OMNIBUS | 15 | | POEM | 156 | | SERIAL | 35 | | SHORTFICTION | 1306 | | CHAPTERBOOK | 10 | +--------------+----------+
- Perhaps we could go after the low-hanging fruit like COVERART and INTERIOART records first and see where it takes us? Ahasuerus 20:05, 17 July 2014 (UTC)
- Sounds like a good plan. I can't see reason to keep any of those types of records, deleting them without any thought. The records for container types (ANTH., COLL., NOVEL, etc.) will require the editor to make some critical decision of either deleting it or creating a pub for it. The records of contained types (ESSAY, POEM, & SHORTFICTION) will almost all have to be left alone, so I don't think we should create such a script and add it to the clean-up list. (I fear an unsuspecting moderator will delete them willy-nilly.) With one exception: those contained types which are credited to "uncredited", which was the original impetus behind the request. Mhhutchins 22:04, 17 July 2014 (UTC)
- On the one hand I share your concern about "unsuspecting moderators" accidentally deleting publess titles. On the other hand, we have publess "contained" titles that could really use a careful review. For example, consider William Gibson's Agrippa - A Book of The Dead. Our Title note reads "Published on diskette only. The original being readable only once as it has a built-in self-destruct program. Very rare." We also point to a Wikipedia article which provides additional information about this unusual poem's history. The article claims that there were at least three paper-based edition -- "small" ($450), "deluxe" ($1,500) and "collectors'" ($7,500). Once we confirm Wikipedia's information, we will presumably want to create a CHAPTERBOOK title record and at least 4 CHAPTERBOOK publication records.
- In the grand scheme of things, I think we are better off creating publication records for all of our pub-less titles. 3,000 is only 1% of the 303,183 pubs that we currently have on file, so it's not like it's going to make the database unmanageable. Ahasuerus 23:02, 17 July 2014 (UTC)
Fixer and publishers
You could get Fixer to change:
- Skylark to Bantam Skylark and Dedalus Limited to Dedalus.
--Chris J 00:33, 19 July 2014 (UTC)
- Done and done. Thanks! Ahasuerus 00:47, 19 July 2014 (UTC)
Pre-filled fields during the cloning process
I worked a submission in the queue in which the editor cloned the English language publication record to create a German language publication record. It never occurred to me that an editor can change the author and title field during a cloned function. I rarely use the clone function myself, preferring the "Add Pub to This Title" function, which pre-fills the title and author fields. If the clone function did the same, we wouldn't have a situation where an editor adds a different language publication to the wrong title record. Would that be a complicated change in the software? Thanks. Mhhutchins 04:43, 26 July 2014 (UTC)
- Hm, that's a good question. Let me take a look... Ahasuerus 05:06, 26 July 2014 (UTC)
Please see the merge I have pending for "Extract from Captain Stormfield's Visit to Heaven". When I try to approve it, it results in a SQL error message. Thanks. -- JLaTondre (talk) 13:13, 27 July 2014 (UTC)
- Thanks for the heads up! I will restore the latest backups on the development server and go from there. Ahasuerus 16:49, 27 July 2014 (UTC)
Fantasy Book, Vol 1, No. 2
I'd like to make a couple of small changes to Fantasy Book, Vol 1, No. 2. I have discovered the reason for the discrepancy between the inside cover credit and the signature on the cover. It appears that two editions of the magazine were done. I recently picked up both of them, thinking I was getting two different issues. I can't say whether I would have purchased both had I known. The issue with the Hunt cover is printed on pulp paper and has a 25¢ price, whereas, the one with the Crozetti cover is printed on book paper and is priced at 35¢. There is a note in Miller/Contento that makes this clear, and the subscription note in the magazine does mention both regular and deluxe editions. I'd like to adjust the existing note to explain this. I also intend to enter a record for the other edition giving the alternate cover and price.
I've also noticed that the illustration on page 11 which we have credited to Tanner appears to have a more complete signature. It looks to me like "Wm F. Tanner". I would assume that "Wm" is for William. Unfortunately, we don't have any other records for either "Wm" or "William". I would suggest that we enter the name as "Wm. F. Tanner" regularizing it with the addition of the period after Wm.
Finally, I'd like to add the letters by Wright, Cowie and Spencer, who we otherwise already have in the database.
- Sounds like a plan! :) Ahasuerus 19:14, 27 July 2014 (UTC)
Going through the latest set of Fixer submissions (2012 ebooks), I find a surprising number of them for non-genre books by Nora Roberts. Is there a way to set these on the back burner while we work on adding spec-fic books by spec-fic authors? Thanks. Mhhutchins 00:22, 1 August 2014 (UTC)
- Well, I could ask Fixer not to submit any ISBNs associated with NONGENRE titles, but that would also affect L. Ron Hubbard, John Barnes and other writers whose NONGENRE works are not as numerous of Roberts's. I could also try to get Fixer to move Roberts's NONGENRE titles to "queue 2", but that's a fair amount of work for just one author and, if done hastily, can make Fixer's internal database unusable (been there, done that.) I guess the easiest thing to do would be for me to try to remember not to submit Roberts's NONGENRE books and move them to "queue 2" instead. Which is also somewhat error-prone, but at least the consequences of an error will be trivial. Ahasuerus 02:06, 1 August 2014 (UTC)
- I mistakenly thought it would be a simple fix. Wrong! Carry on as before. Thanks. Mhhutchins 02:39, 1 August 2014 (UTC)
"Untitled" award records
Is there a way to change an award record currently in the database into an "untitled" award, i.e. one that is not linked to a title record. I can't think of a way to clean the remaining items on this clean-up script. They're all either for series which aren't linkable, or for art books in which the award is given to both the author and the artist. Since we don't credit artists in the title records of these publications, they are not linkable to the title record without losing the valid award data. Mhhutchins 01:20, 4 August 2014 (UTC)
- If an award record is currently linked to a title record and needs to be unlinked, using the "Link Award" page and entering "0" in the "Title #" field should do the trick. However, the awards that you linked above are already "untitled", so perhaps you had something else in mind? Is it the fact that they show up on the "Suspect Untitled Awards" list? Ahasuerus 01:37, 4 August 2014 (UTC)
- That's part of it, but I'm not sure why some "untitled" awards show up on this list while others do not. In other words, how do we clear items from this list? How do we tell the system that they are not "suspected" but actually are "untitled awards? Mhhutchins 02:08, 4 August 2014 (UTC)
- The logic that looks for "suspect" untitled awards is rather complicated. It checks a dozen different conditions, including the presence or absence of different words like "illustrator", "book", "collection", "graphic", "publisher", etc. I have modified it to ignore Locus awards for years prior to 2011, Aurealis awards for 2008, and Hugo awards for 1964 (see FR 655), so the report page should now say "No record found" -- until someone enters another "suspect" award, that is :-) Ahasuerus 02:37, 4 August 2014 (UTC)
- Wonderful. Thanks for the extra effort. Mhhutchins 04:20, 4 August 2014 (UTC)
- One step at a time :) Ahasuerus 05:18, 4 August 2014 (UTC)
Possible duplication of ebooks pubs
Re this title: This is the third one I ran across today, and the first times I noticed that Fixer has very likely duplicated records for the same publication. (You can see that the ASINs are identical.) It's probably because back in 2012, Fixer wasn't pulling ISBNs. I'm not really sure how this can be prevented, except to have moderators check the title records with each submission, and that seems like overkill just to find the rare duplicated record. What do you think? Mhhutchins 06:06, 4 August 2014 (UTC)
- Fixer was still ignoring e-books in 2012 (because I didn't know how to associate their ASINs with ISBNs at the time), so the ISBN-less pub record must have been created by a human editor. I am reasonably sure that we have dozens, perhaps even a few hundred, pub records with missing "e-ISBNs" because they were created using Amazon's data. Ideally, Fixer would create EditPub submissions to add the missing ISBNs (instead of AddPubs), but that may be asking too much of him.
- Hm... I guess I could modify the AddPub pre-approval page (for e-books only) to warn the reviewing moderator if there is another ebook associated with the "to-be-added-to" title that has no ISBN. And I could also create a cleanup script to find duplicate ASINs in e-books' notes. How does that sound? Ahasuerus 22:02, 4 August 2014 (UTC)
- Sounds like a reasonable solution, unless you feel it's more effort than it's worth. But if not, go for it! I think the clean-up script should be able to find most, but then I may be one of the few humans who are adding ASINs of ebooks to Note fields. I'll go ahead and delete the duplicate of the above title. Thanks. Mhhutchins 22:25, 4 August 2014 (UTC)
Rogue Angel series novel Celtic Fire
- Excellent, many thanks! (I could have sworn I checked Google Books -- perhaps they have added the details in the last 48 hours.) Ahasuerus 02:00, 6 August 2014 (UTC)
- Make that 24 hours! I checked it yesterday. Mhhutchins 02:35, 6 August 2014 (UTC)
Billy the Kid
- Yup, that's the cover all right. I remember it well :) Ahasuerus 04:37, 12 August 2014 (UTC)
I believe I have a copy of this pub that you verified. I am not entirely sure, though, the two things that make me uncertain are:
- This pub note: "At least the seventh Ace printing, possibly the eighth". Is this note referring to completely different earlier Ace books (e.g. this one)? Or is the note referring to other printings of the same Ace book that have the same ISBN - and it just happens that we don't have any pub records in the DB for any of the other printings? If the latter interpretation is correct, then do you know how I might distinguish different printings? Because I don't see where the evidence for the pub note's claim about the 7th or 8th printing comes from - after all there is no printing history in the book (numberline or otherwise), no price differences, and I believe Ace does not use gutter codes.
- Virtually the only difference between my copy of the book and the information in the pub record is that in my book the story "Huddling Place" starts on page 45, but the pub record says it starts on page 44.
Under normal circumstances I would have said the page difference is an error in the database, and that I have the same book. But given the (to me) ambiguous pub note I am unsure whether this page difference might distinguish two printings, and I might have a different book? Any illumination on the matter would be greatly appreciated, but if you're too busy I can also ask Bluesman, he's PV2. Thanks, Patrick -- Herzbube Talk 07:49, 14 August 2014 (UTC)
- Checking my copy, I see that "Huddling Place" starts on page 45, so "44" must have been a data entry error and we have the same printing.
- As far as the printing number goes, I am not sure whether that note was added by me or by Bluesman, but I am reasonably sure that it refers to older Ace printings of "City".
- Also, keep in mind that Ace is a special case. Well, sort of. At some point in the 1970s, they apparently decided to forget all about their older printings and start the numbering scheme from scratch. The result is that different printings of an Ace books may say something like the following on their respective copyright pages:
- First Ace printing - 1964
- Second Ace printing - 1966
- Third Ace printing - 1969
- Fourth Ace printing - 1972
- First Ace printing - 1976
- Second Ace printing - 1979
- This Ace printing - 1982
- Lovely, isn't it? :) Ahasuerus 23:19, 14 August 2014 (UTC)
Award Display Order
- The award list is sorted by year, which is admittedly not particularly useful when all/most of the awards were given in the same year. Let me add a sort by award level so that wins would appear first... Ahasuerus 21:04, 16 August 2014 (UTC)
- Bad HTML. (You didn't close the URL with a final ".) Try again, or I could fix it for you. This would have shown up on a clean-up script, but I've not ran that one in the last 24 hours or I would have already found it and fixed it. Mhhutchins 14:30, 18 August 2014 (UTC)
- There's another submission in the queue which can't be accepted or rejected because of the bad HTML. Once you've seen this message, I'll do a forced reject of the submission and you can make a new one. Thanks. Mhhutchins 14:41, 18 August 2014 (UTC)
- Thanks for looking into this! Ahasuerus 14:59, 18 August 2014 (UTC)
- Thanks, Michael. In my defence it wasn't my bad HTML, but I've fixed it and it all looks good. --AliHarlow 15:48, 18 August 2014 (UTC)
- Sorry, I assumed you wrote the whole Note. Thinking more about this, I realize that the clean-up script would not have found this HTML error, because it only looks for missing (either opening or closing) angle brackets (< and >). Ahasuerus, is there a way to update the script to find URLs in Note fields which are missing either an opening or closing quotation mark? I know it'll also work properly if neither are present, but not if only one is present. Thanks. Mhhutchins 16:26, 18 August 2014 (UTC)
- That's a good point, I'll create an FR to update the cleanup script to look for mismatched quotation marks. However, searching all Notes records in the database is extremely resource-consuming and making it do more things will make it even more so. We need to move this cleanup script to nightly processing so that it would only be run once a day. We also need to generalize it so that it would look for all kinds of bad Notes, not just publication-specific Notes. I'll see what I can do... Ahasuerus 16:51, 18 August 2014 (UTC)
- The principal problem is that a Summary page for an author without a language code has no way of displaying language information correctly. For example, take Philip K. Dick's Summary page. The Collection and Omnibus sections include entries like Ubik [German ] and Philip K. Dick Omnibus [Dutch ], which would look like ordinary English titles if the author record didn't have a language code. Ahasuerus 02:39, 21 August 2014 (UTC)
- I know how this affects display on summary pages, but remember there was a specific reason why we were adding languages to authors where the language had not been given. I'll try to look back over some of the talk pages and see what brought up the idea of creating this in the first place. Thanks. Mhhutchins 02:46, 21 August 2014 (UTC)
- Ah, yes, I remember it now... Ahasuerus 05:29, 21 August 2014 (UTC)
The Worlds of George O.
"Remove Titles from Publication" function
Is it possible to add an option to remove ALL "regular titles" when using the "Remove Titles from Publication" function? I'd make a feature request, but I haven't figure out how to negotiate that process. Thanks. Mhhutchins 20:52, 30 August 2014 (UTC)
- Let me make sure that I understand the desired functionality correctly. Suppose you are working on some magazine pub, e.g. Heyne Science Fiction Magazin, #2. If you click on Remove Titles From This Pub, you can then select which titles to remove. Are you suggesting that we add a "Select all Regular Titles" check-box, which, when checked, would cause individual check-boxes for regular titles to be checked? If so, then yes, it's possible to implement. Ahasuerus 21:06, 30 August 2014 (UTC)
- That's it exactly. Often I find that it's better to remove all titles and import them from another publication where the contents are paginated. It's very hard to add numbers in the edit page where the contents are in alphabetical order instead of how they're displayed on the record. Thanks. Mhhutchins 21:37, 30 August 2014 (UTC)
- Sounds like a good idea -- FR 665 has been created. Thanks!
- P.S. As far as the ability to create FRs goes, just go to this page, click on "Create Ticket" on the left hand side, enter the FR title in the "Title" field at the top of the page, then a description in the big box at the bottom of the page and click "Save". That's it! Ahasuerus 22:19, 30 August 2014 (UTC)
- Thanks. I'm going to copy these instructions on my user page. Mhhutchins 17:36, 8 September 2014 (UTC)
Thrilling SF Adventures #14
- Let's take a look... It's really three small, almost overlapping, pictures with long-ish captions (1-3 sentences) plus a short (7 sentences) uncredited essay at the bottom of the page. I suspect that the first title record was supposed to cover all of the pictures and the second record was supposed to be an ESSAY rather than SHORTFICTION.
- Also, the Note currently says:
- The illustrated essay "Riddles of Science: Mystery of the Sunspots" is signed by Joe Sewell and it is assumed that he is responsible for both the text (7 sentences) and the art.
- but all the other "Riddles of Science" essays are currently entered as by "uncredited".
- Based on the above, I plan to change the second INTERIOART record to ESSAY and change its author to "uncredited". Thanks for bringing this issue to my attention! Ahasuerus 17:42, 9 September 2014 (UTC)
A broader definition of OMNIBUS
A pub record appearing on the cleanup script that finds "Omnibuses without Content Titles" prompted this discussion. Please join in the discussion when you can get a chance. Thanks. Mhhutchins 04:53, 9 September 2014 (UTC)
For some reason when I uploaded an image for [this] edition, the link is showing properly in the field but no image is present in the record? It worked for another book a couple of minutes ago?? --~ Bill, Bluesman 18:18, 9 September 2014 (UTC)
- That's because you entered the URL in the "Artist" field rather than in the "Image URL" field -- an easy mistake to make, as I know from personal experience :-) Ahasuerus 18:21, 9 September 2014 (UTC)
- Ooooops. After 20,000 or so that really shouldn't happen ..... (((; --~ Bill, Bluesman 18:25, 9 September 2014 (UTC)
- Oh, I still do it from time to time, especially late at night :-) Ahasuerus 18:34, 9 September 2014 (UTC)
Re this title: a translation of Carroll's Alice into Nor-Aest Scots, which is not Scottish Gaelic, and may not even be a language but the Doric dialect of Northeast Scots. How should we handle it? Mhhutchins 04:54, 13 September 2014 (UTC)
- Scots Language Centre claims that it's one of four dialects comprising Scots. I think the easiest thing to do would be to add "Scots" to the list of supported languages and then make a note to the effect that this particular translation uses the Doric dialect of the Scots language. What do you think? Ahasuerus 05:25, 13 September 2014 (UTC)
- Great plan. Thanks. Mhhutchins 15:35, 13 September 2014 (UTC)
- Scots has been added and the title record has been corrected. Ahasuerus 22:05, 14 September 2014 (UTC)
- Great. Thanks. Mhhutchins 22:18, 14 September 2014 (UTC)
Similar Mode Suggestion
Please see ISFDB:Community Portal#Duplicate Finder changes. Bit of a late response so thought I'd point it out in case you no longer have that discussion on your radar. I just ran into this use case. ;-) -- JLaTondre (talk) 12:57, 14 September 2014 (UTC)
- Done! Ahasuerus 18:39, 14 September 2014 (UTC)
Novels first serialized as ebooks
A fairly recent publishing phenomenon is the serialization of novels as ebooks, for example here, here, and here. Because of the software's assumption that all serials first appear in print magazines, it gives the title records of the variants (the SERIAL records) as "Magazine Appearances" on the publication page and the author's summary page. Is there a way of recognizing this trend to accommodate both forms of serialization without having to do much damage to the older form? Perhaps just changing the name to "Serialized Publications" would cover both forms and be the easiest way to update the software. Thanks. Mhhutchins 15:13, 14 September 2014 (UTC)
- Hm... How about "Serializations"? Ahasuerus 15:37, 14 September 2014 (UTC)
- (Slapping forehead!) Even better. Mhhutchins 22:19, 14 September 2014 (UTC)
- All done. (What can I say? Insomnia...) Ahasuerus 06:21, 15 September 2014 (UTC)
Omnibuses without contents
I've worked this cleanup report down to the four last records. The first is for a Russian edition of Lord of the Rings, which needs someone with a knowledge of the Russian language. I've left notes on the talk pages of the editors who verified the second and third records.
(The latter to reconsider typing this record as a COLLECTION.)
Before I leave a message on the talk page of the verifier of the fourth record, I need some clarification. Since this is a binding of the ten issues of a magazine, shouldn't it be typed as MAGAZINE instead of OMNIBUS? If you consider it to be an OMNIBUS, we'd need to add title references (editor records) for each of the magazine issues, and I'm not sure if the software can handle a record containing more than one EDITOR record. (There's a cleanup report that finds such anomalies.) I think we should just consider it a MAGAZINE, and there'd be no issue with the cleanup report or the way it's displayed. What do you think? Mhhutchins 04:21, 20 September 2014 (UTC)
- Update: the third record was fixed by the verifier, leaving the one I called the "fourth" record above as #3 on the current list. Mhhutchins 15:32, 20 September 2014 (UTC)
- Excellent work! I have updated the Russian pub, so we are down to just two outstanding records.
- As far as The Arkham Collector, Vol. 1 goes, how about entering it as an ANTHOLOGY, which is how we usually handle magazine reprints? Also, looking at the Content section, it would appear to be incomplete. Of the 348 pages in the book only the first 178 are covered and issues 6-10 are not listed. Perhaps Ron started entering the data and then got sidetracked? Ahasuerus 16:38, 20 September 2014 (UTC)
- I don't think so. Look at the pages for the contents, and you'll see many are duplicate. That's a good indication that each issue kept its original pagination. Ron notes that there are no separate title pages. Looking at the verification date, the record was created this before we had the ability to order how the contents are displayed using the pipe method. I'll leave him a message pointing him to this discussion. Thanks. Mhhutchins 17:44, 20 September 2014 (UTC)
- Oh, I see. A poster boy for the utility of pipes :-) Ahasuerus 17:49, 20 September 2014 (UTC)
- Forgot to address the original problem: since most of the contents appear to be ESSAYs (with seven SHORTFICTION pieces and many scattered POEMs) wouldn't NONFICTION be more appropriate? The only way I could see this being an ANTHOLOGY is if the essays are about the poems and stories. And they don't appear to be. (There is one SHORTFICTION content which is very likely mistyped.) Mhhutchins 17:52, 20 September 2014 (UTC)
- Hold the press. The pagination in the individual issues was continuous. The contents of the last four issues haven't been added to the record. So your original theory is correct. Sorry. Mhhutchins 17:58, 20 September 2014 (UTC)
- (after edit conflict) Let me take a closer look... <puts his good glasses on> By my count, it's 48 poems and 7 short stories, not counting the mistyped "shortfiction" record. The short stories account for about 50 of the 178 pages, so the combined "fiction page count" is close to half the total number of pages. Wouldn't it make the book closer to an anthology? Ahasuerus 18:09, 20 September 2014 (UTC)
- If each of the poems had a separate page, maybe. But it also appears that many of the essays are less than a page reviews (which are entered as two content records). Even so, "close to half" means "less than the majority", right? (Unless you're a Republican senator!) I'll leave it to Ron to determine whether the fiction or the nonfiction are the dominant type. Thanks. Mhhutchins 18:54, 20 September 2014 (UTC)
- Well, according to Help the NONFICTION title type is for "books that are predominantly or completely non-fiction". "Predominant" is somewhat vague (is 51% "predominant"?), but this happens so rarely (in books) that a little vagueness probably doesn't hurt :) Ahasuerus 19:33, 20 September 2014 (UTC)
- My sense is that anthology would be more appropriate. However, let me go over these, adding the piped page numbers in each issue (and correct the errant SHORTFICTION). I'll probably drop everything in the collected volume and re-import to get the page numbers right. I can't imagine why I missed the later issues before, but I'll ensure they're all there now. I'll probably have a better idea over whether fiction or nonfiction is predominant after I've done all that. I'm also thinking that the title record for collection (currently the OMNIBUS title) should be taken out of the series. It's causing the EDITOR records to appear in Derleth's Fiction Series section on his bibliography page. --Ron ~ RtraceTalk 14:25, 21 September 2014 (UTC)
Request to add a new award
- Faster than Superman! Thanks. Mhhutchins 16:37, 23 September 2014 (UTC)
Bug in cleanup report
- Wow, that's impressive! Let me take a look... Ahasuerus 17:05, 23 September 2014 (UTC)
- Got it. The bug was introduced a few days ago when I moved this report to nightly processing. Let me fix it real quick... Ahasuerus 17:22, 23 September 2014 (UTC)
- Done! Ahasuerus 17:46, 23 September 2014 (UTC)
Re this Fixer submission: I noticed this was priced very high for a public domain work. When I check the price on both B&N and Amazon, it's listed as $5.99. Should I correct the record after accepting it? Mhhutchins 16:37, 24 September 2014 (UTC)
- Hm, that's curious. Could you please keep it on hold for a few hours? I'd like to research it when I get back. Ahasuerus 16:57, 24 September 2014 (UTC)
- OK, I think I see what's going on. Fixer found this ISBN on 2014-03-03 and captured B&N's "list price" on 2014-03-14. At the time it was apparently $10.97. That's what Fixer added to his database and that's what he used to create the submission. However, there is a flaw in the submission creation logic -- it uses the current date rather than the "date price was captured" when populating the Note field. If the "list price" changes between the time Fixer captures it and the time he creates a submission for the ISBN, then the data in the Note field will be invalid.
- We can't prevent publishers from changing their list prices retroactively (perhaps to correct a data entry error?), but at least I can fix Fixer's logic. Thanks for finding the bug! Ahasuerus 21:03, 24 September 2014 (UTC)
AddPub missing a title record to "add pub"
Re this Fixer submission. There's no such title in the database, thus a hard reject candidate. Could the title have been deleted between the time it was queued by Fixer and the time of this submission? Does Fixer check for current titles at the time of the submission and when it gathers the data from the source? Mhhutchins 17:33, 27 September 2014 (UTC)
- You are right, that's exactly what happened. Fixer uses a copy of the ISFDB database as of the last backup, which can be up to a week old. A week ago we had 9 NOVELs, 1 COLLECTION and 1 NONFICTION title by Hilary Mantel. As of today we are down to 4 NOVELs, presumably because the rest were non-genre and somebody has deleted them. When Fixer created AddPubs for some of the defunct title records, they were determined to be invalid by the approval logic.
- The next question is what Fixer can/should do about it. At submission time, Fixer uses our Web API to check for the existence of publication records with the about-to-be-submitted ISBN. The Web API doesn't let you check for the existence of title records, but that can be easily rectified, which will address the immediate technical issue.
- Then we have to decide what we want Fixer to do when he finds that the title record is gone. In this particular case the logical thing would have been to delete the ISBNs from Fixer's database and forget the whole thing. However, a title record may have been deleted (or, more likely, merged) because it was a duplicate, in which case we would want Fixer to create a NewPub submission which the approving moderator would then need to merge with the pre-existing title.
- So it looks like the safe thing to do is to create a NewPub submission if the reference title is no longer available. Does this sound reasonable? Ahasuerus 18:29, 27 September 2014 (UTC)
- Very reasonable. Just the fact that you can have Fixer make such logical decisions which many humans have trouble making is a wonder to me. :) Thanks. Mhhutchins 01:19, 29 September 2014 (UTC)
- Sounds good. I will create an FR and work on it tomorrow. Ahasuerus 02:54, 29 September 2014 (UTC)
Peculiar behavior on the "Titles Without Publications" cleanup report
See here. I corrected the type of a publess title from NONFICTION to ESSAY, thinking it would disappear from the list, since you've not added ESSAY type to the those that the script are looking for. The same thing happened last night, and I just "ignored" it. It's no big deal to do the same with the one today, but I wanted you to know, just in case this isn't the expected behavior. Thanks. Mhhutchins 14:59, 1 October 2014 (UTC)
- It looks like it may have been a timing issue. Rerunning the nightly report manually made the errant title go away. Ahasuerus 16:17, 1 October 2014 (UTC)
Under data elements to consider, there is an entry of Character names (requested on rec.arts.sf.written). The request seems to be from 2008. I've thought about this myself a bit and wondered if there had been any discussion? Thanks. Doug 19:48, 1 October 2014 (UTC)
- I don't think the issue has come up since 2008-2009 or at least I don't see any Feature Requests along similar lines. If you'd like to see it implemented, I would suggest posting a proposal on the Community Portal to see how much interest may exist. It's always possible to re-juggle development priorities if there is enough grass roots support :-) Ahasuerus 21:50, 1 October 2014 (UTC)
- I figure there's a lot of discussion first before we think about development. Using IMDB as a model may help start the discussion, but is a long way from being a spec. I'll give it some more thought first before starting something new. Thanks for the feedback. Doug 00:35, 2 October 2014 (UTC)
Shirley's New Taboos plus...
Might be a wild assumption, but I have the feeling that you might be the one who added the tag 'criticism' to this CHAPTERBOOK. Reading the book and finding that the leading novelette takes more than half of the feature space, what do you think of moving the tag to the title giving essay instead? I am right now two thirds through the shortfiction and have added already some tags (still have to decide if it sums up to a dystopia). Probably the tag was added because of the theme of the title essay and seems to fit it perfectly, but IMHO it seems a bit too much to characterize the whole book. What do you think? Please forgive me, if the assumption was a false one. Christian Stonecreek 13:52, 3 October 2014 (UTC)
- Guilty as charged! Chances are that I added it as part of a Fixer run. I have now removed the tag and will leave it in your capable hands :-) Ahasuerus 15:18, 3 October 2014 (UTC)
Pub dating question
After accepting the submission adding this pub record, I changed the date of the title record to reflect this pub record's earlier publication date. That's when I noticed it had cover art that differed from the first edition, unusual for a simultaneous (or near-simultaneous) ebook edition. So I check its publication date on Amazon and see that it gives January 5, 2010 and not the Fixer date of 2009-10-21. Have I uncovered an anomalous situation or is it possible that the publication dates for ebooks published close to five years ago aren't really that reliable, more often than we'd previously suspected? Mhhutchins 01:37, 13 October 2014 (UTC)
- If you search the Amazon page that you linked for the words "Publication Date", you'll find a line which reads "Publication Date: October 21, 2009". That's what Fixer gets back from the Amazon API as well. However, if you scroll down to the Product Details section, it says "Publisher: Tor Books; Reprint edition (January 5, 2010)". So I guess the questions are as follows:
- How often can these two dates get out of sync?
- When they do get out of sync, which one is correct?
- Does this discrepancy affect e-books only or are paper books affected as well?
- Ahasuerus 02:24, 13 October 2014 (UTC)
- Strange. I never knew that Amazon displayed the publication date in two places, so I could not have suspected that the dates could disagree. When I work from data on Amazon, I always get the date from the publisher field, since there is no date field in the "Product Details" section. If you go to the publisher's website, you find a listing of the ebook of this title that gives the publication date as January 2010. Mhhutchins 03:38, 13 October 2014 (UTC)
- I'll have to review what Fixer gets back from the Amazon API. Perhaps there is another field there that contains the same date as the one displayed in the "Publisher" field. Ahasuerus 03:54, 13 October 2014 (UTC)
- It turns out that the Amazon API does have a separate field for what it calls "Release Date". It's the date that is displayed next to the publisher's name on regular Amazon pages. Here is what Fixer gets back for "Hidden Empire":
- PublicationDate: 2009-10-21
- ReleaseDate: 2010-01-05
- The question then is whether Fixer should always use the "ReleaseDate" field instead of the "PublicationDate". Ahasuerus 23:14, 14 October 2014 (UTC)
- Does Amazon specify why there are two different dates and why they don't necessarily match? Mhhutchins 23:26, 14 October 2014 (UTC)
- Here is how they describe these fields:
- PublicationDate: Publication date.
- ReleaseDate: Date on which the item was latest released. Items that have been released multiple times have both release dates and original release dates. See ReleaseDate.
- The first description is not very helpful. The second is self-referential and there is no "Original Release Date" anywhere that I can see, but at least it suggests that "ReleaseDate" is the latest date. It seems like we should be using the "ReleaseDate" field, but first we may want to confirm that it hasn't been used as "the latest date that any edition of this book was released on". Ahasuerus 00:09, 15 October 2014 (UTC)
- Hopefully you'll be able to parse their logic, and the "ReleaseDate" is NOT just the latest release of any edition.
- Since the original post, I've checked the Amazon listings of several Fixer "AddPub" submissions which has a warning about an earlier publication date. Roughly one in three were more than a week or so. In cases where the publication date is more than two weeks, I've updated the date of the title record, since it's obvious that the publisher released the ebook edition months before the print edition (sometimes a whole year or more in the case of Samhain Publishing.) The remaining title dates I've left alone, which may cause some concern in the title record display for for titles published at the end of the year, e.g. a title record dated January 1, 2014 (based on the original print edition) and displayed as "2014", might have an ebook publication displaying "Dec 2013", because it was published on "2013-12-24". At this point, I would feel uncomfortable changing the publication date of the title to "2013-12-24" based solely on Fixer's interpretation of Amazon's date. Perhaps Fixer can compare the dates given on both B&N and Amazon, and flag any that have discrepancies of more than a couple weeks? Mhhutchins 01:20, 15 October 2014 (UTC)
(unindent) I am feeling a little better tonight, so I too have poked around a bit. Here is what I see. Consider the e-book edition of Robert Jordan's Knife of Dreams, ISBN 9781429960816:
- Amazon's Look Inside shows this ISBN and says "First E-book Edition: August 2010].
- Under "Book Description" Amazon says "Publication Date: April 21, 2010"
- Amazon's API returns 2010-04-21 in the "PublicationDate" field.
- Under "Product Details" it says "Publisher: Tor Books (August 24, 2010)"
- Amazon's API returns 2010-08-24 in the "ReleaseDate" field.
- Checking B&N, I find "Publication date: 8/24/2010".
So Look Inside, Amazon API's "ReleaseDate" and B&N all agree that the book was published on 2010-08-24. "Publication Date" is apparently an outlier. I am not sure where it originally came from -- perhaps something that the publisher submitted early in the process, but then things got delayed? In addition, spot-checking other records on Amazon, I note that only e-books appear to have this "Publication Date" value displayed.
Could you please double-check my findings? I am still under the weather and heavily medicated, so a second pair of eyes would be good :-)
As far as using B&N's dates goes, it's possible, but there would be some technical issues to overcome. If we can determine that Amazon's ReleaseDate field is accurate, it would be better to leverage it. Of course, I will need to instruct Fixer to go back and check Amazon's data for discrepancies between PublicationDate and ReleaseDate and create submissions to make appropriate corrections.
The good news, to the extent that there is any, is that recent e-books do not seem to have this problem, or at least spot-checking the e-books coming out this fall shows the same date displayed in both places. Ahasuerus 03:49, 15 October 2014 (UTC)
- Done. It will disappear from the list after the next nightly job run. Ahasuerus
Our discussion about Amazon's two ways of recording the dates of publication made me think of what could be a useful clean-up report. How about one that finds titles which have publications with dates earlier than the title date? I know this is supposed to be caught in the warnings for AddPub submissions, but I'm sure some moderators aren't as diligent about correcting the title date fields as others. Also, there are very likely many pubs which were added with earlier dates before the warning was added to the submission page. (I found one the other day that was years apart.) We could start with the records that are more than a year apart, just to get an idea of how many are in the database. And then we could adjust the script to closer dates as we fix them. Mhhutchins 04:49, 15 October 2014 (UTC)
- Sounds like a good idea. We may want to clean up the current Fixer/Amazon mess first, though, or else some well-meaning moderator may change a bunch of title dates based on yet-to-be-corrected publication dates.
- Also, a report that finds titles (book length titles at first) all of whose associated pubs have a later publication date may also be useful. Presumably it will be an indication that we don't have the first edition on file. Of course, we will need to account for variant titles, serializations and such. Ahasuerus 05:08, 15 October 2014 (UTC)
- Yes, I can see that being useful as well. Mhhutchins 05:11, 15 October 2014 (UTC)
- OK, 2 FRs have been created. Ahasuerus 05:43, 15 October 2014 (UTC)
Another Similar Mode Suggestion
- Done! Ahasuerus 02:01, 16 October 2014 (UTC)
- You are very welcome! Since the fundamentals are now in place, adding new replacement pairs is easy :) Ahasuerus 23:43, 17 October 2014 (UTC)
Voyage of Slaves
- Excellent, thanks! :) Ahasuerus 00:46, 18 October 2014 (UTC)
Pseudonym Submitted Warning on Make Variant
When making a record that uses a pseudonym into a variant of a new record under the canonical name, the "Proposed Make Variant Submission" moderator screen displays a "Pseudonym submitted" warning. This throws me off as I assume it means a pseudonym was submitted as the parent record. A pseudonym as the variant would be the correct thing so am I missing something about the meaning of this warning? Thanks. -- JLaTondre (talk) 14:35, 18 October 2014 (UTC)
- That's a good point. The software currently determines whether a warning is needed based on what's displayed in the Current and the Proposed columns, but it should only be checking the value in the Proposed column. I'll see what I can do... Ahasuerus 22:07, 18 October 2014 (UTC)
- Fixed. Thanks for reporting the problem! Ahasuerus 01:36, 19 October 2014 (UTC)
Variant type mismatches
If I recall correctly, at one time there was a wiki-based project to find varianted titles that don't match the type of their parent record. Is it possible to do this as a database clean-up report? (This record prompted the request.) Thanks. Mhhutchins 21:41, 20 October 2014 (UTC)
- That's a very good point. Here is what I see in the database after eliminating SERIAL-NOVEL, SERIAL-SHORTFICTION and INTERIORART-COVERART mismatches:
+--------------+----------+ | title type | count | +--------------+----------+ | ANTHOLOGY | 4 | | COLLECTION | 27 | | COVERART | 68 | | INTERIORART | 4 | | EDITOR | 1 | | ESSAY | 52 | | NOVEL | 142 | | NONFICTION | 1 | | NONGENRE | 5 | | OMNIBUS | 5 | | POEM | 7 | | REVIEW | 1 | | SERIAL | 62 | | SHORTFICTION | 109 | +--------------+----------+
- Some of them may be legitimate, e.g. SERIAL-COLLECTION (?) mismatches, but the majority are probably not. I will see if I can create a cleanup report and fine-tune its logic later tonight. Ahasuerus 22:01, 20 October 2014 (UTC)
- Well, I was waiting for the dust to settle to see if this report would require the ability to ignore records. As you said, the verdict is in and some records are indeed unfixable, so I will be adding the ability to ignore them shortly. Ahasuerus 00:38, 30 October 2014 (UTC)
- Done! Ahasuerus 01:04, 30 October 2014 (UTC)
- Thanks. Much appreciated! Mhhutchins 03:02, 30 October 2014 (UTC)
Check out Bob's Make Variant that I have on hold. Something very wrong there. --MartyD 02:42, 22 October 2014 (UTC)
- Oops, looks like the last patch introduced a bug. Looking... Ahasuerus 03:02, 22 October 2014 (UTC)
- Fixed! As an added bonus, COVER/COVER VT submissions are now color-coordinated :) 03:17, 22 October 2014 (UTC)
Cleanup report for "Authors that exist only due to reviews"
Re this cleanup report: I've been meaning to ask you this for awhile but keep forgetting. Can you list the author being reviewed instead of the reviewer? That will make it clear to any moderator (other than me) exactly what needs to be cleaned. Thanks. Mhhutchins 03:57, 31 October 2014 (UTC)
- Sure thing. FR 731 has been created. I will see if I can implement it tomorrow. (This whole "sleep" thing is so wasteful if you ask me...) Ahasuerus 05:12, 31 October 2014 (UTC)
- Done! Ahasuerus 20:43, 31 October 2014 (UTC)
Possible bug in a new cleanup report
When I remove a title from this report, it shows up on this report. It only now just started. Look at the top of the second report. I blanked the ISBN fields, removing the ISSNs from the records for issues of Paperback Inferno and they now have "identical ISBNs"? The record for Lord Prestimion was cleared from the report yesterday, so I don't know why it's showing up again as one of these "identical ISBNs". Thanks for looking. Mhhutchins 04:44, 31 October 2014 (UTC)
PS: I could have clicked the "Ignore" link, but I wanted you to see it to check for problems before I did that. Mhhutchins 04:47, 31 October 2014 (UTC)
- Hm, something is definitely wrong there. Let's keep the "blank ISBN" pubs in their current state for now. I'll take a closer look in the morning... Ahasuerus 05:09, 31 October 2014 (UTC)
- Fixed! Ahasuerus 20:16, 31 October 2014 (UTC)
Image Links on Cover Title Records
I'm confused by the new feature that shows a link to the cover image on cover title records. Example Cover: The Tar-Aiym Krang. I'm not sure what the point of having the link is as it doesn't go to the URL, it goes to the publication (same as clicking on the image itself). This is redundant and seems confusing to me as when you click on a URL, one normally expects to go to the URL, not a different location.
However, the bigger issue is that it makes a mess of the display. Instead of a nice listing of covers like on a non-cover title record, they are scattered about. The URLs are not even with the image they go with. If the intent of listing the image URLs was to be able to see duplicates or something else about the URL itself, than I would recommend the following format:
COVER1 COVER2 COVER3 LINK1 LINK2 LINK3
If the intent was to show the image source (for example to see that a non-matching cover came from Amazon and may be a result of a wrong scan instead of a bad merge), than I would recommend the following format:
COVER1 COVER2 COVER3 SITE1 SITE2 SITE3
- Oops! That was not intentional, sorry about that. The other day I installed a big patch ahead of adding support for SFE3-hosted images (they have additional, very specific requirements.) There were no planned user-experienced changes, but I messed up. Let me fix it real quick... Ahasuerus 17:26, 9 November 2014 (UTC)
- Got it! Thanks for identifying the issue! Ahasuerus 18:43, 9 November 2014 (UTC)
Publications with Identical ISBNs and Different Titles
Before retiring last night, I cleared up all but one pair of titles on this report. At the time the Clean-up Report list showed 300+ items outstanding on this list. I knew that I'd not cleared up that many from the night before. So today I notice it says that there are still 349 items on the list, when the only ones visible are the same pair I left last night (before the 2am EST report was regenerated). Are some titles not visible on the list? Mhhutchins 22:09, 9 November 2014 (UTC)
- It looks like a bug. Let me see if I can find it... Ahasuerus 23:49, 9 November 2014 (UTC)
- 2740! But no publications visible on the list. Mhhutchins 03:06, 10 November 2014 (UTC)
- Earlier this evening I made some changes which solved the problem on the development server. However, they clearly didn't work on the main server, so now I am trying to figure out why the change failed... Ahasuerus 03:12, 10 November 2014 (UTC)
- Got it! This is by far the most complex cleanup report, but hopefully it's finally "done done" :-) Ahasuerus 04:32, 10 November 2014 (UTC)
I just approved a Fixer submission for Elvis Has Not Left the Building. It's "alternate history", where Elvis is still alive and working as a private investigator in California (I added some more description to the title summary). Chavey 17:47, 19 November 2014 (UTC)
- It would appear that the book is "secret history" rather than traditional "alternative history". A "secret history" is outwardly our history except that there are wheels within wheels that we are not aware of, e.g. the world is secretly run by the Illuminati or the Milk Board :-) Ahasuerus 18:20, 19 November 2014 (UTC)
What confuses me is that Fixer added this ebook, but also listed it as NONGENRE. If it thinks it's non-genre, shouldn't it have skipped the book? Chavey 17:47, 19 November 2014 (UTC)
- It was an AddPubs, so the NONGENRE Title record already existed in the database. When I pre-process Fixer's submissions, I try to discard unnecessary NONGENRE AddPubs like reprints of Wilkie Collins' non-genre works. (As an aside, do we really want his NONGENRE novels?) However, it's not always a trivial decision and I don't always have enough information during the pre-processing stage, so I tend to err on the side of caution and leave it to the approving moderator. Ahasuerus 18:20, 19 November 2014 (UTC)
Surely the author is not "above the threshold", unless Fixer is just using raw numbers of publications to determine the threshold. Chavey 17:47, 19 November 2014 (UTC)
- Fixer doesn't make these decisions, he just goes by third parties' (Amazon etc) designations. And, as you know, data entry clerks are not always the best judges of genre :-) Ahasuerus 18:20, 19 November 2014 (UTC)
I've temporarily left the book as NONGENRE, since I don't really think of that type of "alternate history" as being spec fic (otherwise we might have to accept all historical romance books as "alternate history"). Chavey 17:47, 19 November 2014 (UTC)
- My thinking is that we only want to include "secret history" books if the premise is significantly counter-factual. Thus a "conspiracy going back to the dawn of time" a la Dan Brown would be included, but a historical novel in which the commander of British forces in Canada during the War of 1812 was a US spy (there is a Canadian historian who subscribes to this idea) wouldn't be. Ahasuerus 18:20, 19 November 2014 (UTC)
But I don't quite understand what Fixer was up to, and what should really happen to this book. Chavey 17:47, 19 November 2014 (UTC)
- I don't think there is any harm in leaving it as NONGENRE. It's kind of borderline, so my thinking is that having it listed with an explanation of what it's all about (like you have done) can be helpful to our users. Ahasuerus 18:20, 19 November 2014 (UTC)
- Ok, thanks for the explanation. Chavey 22:12, 20 November 2014 (UTC)
Since the recent conversations regarding deep linking to SFE3 went to all Mods, I did pay attention and it was quite interesting. Beyond me programming-wise, but that's the way it is. My first encounter with same comes from [here]. I set my Preferences to display all images from any title's bibliographic page. The Tor edition, first one on top, does NOT display in this mode. Opening the record does display the image though it loads a little slowly. I have no idea if this is a bug or just the way the linking to SFE3 will work. I am merely the messenger!! Cheers! --~ Bill, Bluesman 01:45, 23 November 2014 (UTC)
- Good catch, fixed! Ahasuerus 02:05, 23 November 2014 (UTC)
- Not really. Now there isn't even the rectangle with the ? in it on the main page and the record now has no image .... ?? And it took over a minute for that to download. --~ Bill, Bluesman 03:23, 23 November 2014 (UTC)
- Hm, that's strange -- seems to work for me. I wonder if the Internet connection between your PC and the SFE3 server was having issues when you last looked at the page. Could you please try again and reload all images? I am not sure what key combination you need to use in your browser, but in IE, Firefox and Chrome it's Control-F5. Ahasuerus 04:02, 23 November 2014 (UTC)
- Hmmmm, now it comes up fine. Gremlins! --~ Bill, Bluesman 05:04, 23 November 2014 (UTC)
- It stands to reason that gremlins would attack an SF site! :) Ahasuerus 05:12, 23 November 2014 (UTC)
Pending submissions affecting one of your verified
Hi, of the active editors you seem to be the first in line as verifier for this book for which there is a submission pending as by Thync. I would have welcomed him/her and taken the other one pending, but thought it better to have one and the same person doing it at the beginning. Stonecreek 07:43, 27 November 2014 (UTC)
- Got it, thanks! Ahasuerus 22:21, 27 November 2014 (UTC)
Robert A. W. Lowndes duplicate titles
Hi. Several duplicate titles for above author show up in this search. Most of them (except for the 2 Dracula intros) are your pv'd magazines and they are departments which haven't been disambiguated with the title and date. One is SWFritter's and one is Seniorlady's. For some reason the Inquisitions column in Bizarre #2 doesn't show up in the search but is also not disambiguated. In general, when this type of situation occurs, is it worthwhile following through and straightening them out? Thanks, Doug / Vornoff 17:32, 2 December 2014 (UTC)
- Good points, thanks! I have merged the two Dracula intros and disambiguated the rest of my verified publications. They were some of my earliest verifications, back when the data entry standards for magazine departments were still being hammered out. As a general rule, I see no harm in bringing these cases up to the verifiers' attention. What's the worst that can happen? :-) Ahasuerus 23:20, 2 December 2014 (UTC)
- Okay - Thanks. Doug / Vornoff 23:48, 2 December 2014 (UTC)
Exerion / Exerión
Hi. I saw that you added an accent to Exerion, but there is apparently none, according to this source which, being Spanish, seems more likely to be correct. I have removed your reference to Cosmos Latinos to replace it by one to tercerafundacion, which identifies the first publication. Cheers ! Linguist 10:48, 6 December 2014 (UTC).
- My source was the title page (p. 294) in Cosmos Latinos (courtesy of Amazon's Look Inside), which reads:
Exerion Exerión, 2000 by Pablo A. Castro translated by Andrea Bell
- I'll update the Note field to indicate that secondary sources differ. Thanks! Ahasuerus 16:19, 6 December 2014 (UTC)
Removing series data from a CHAPTERBOOK record
I don't know how he did it, but an editor was able to figure out how to enter series data to a CHAPTERBOOK title record. Now I can't figure out how to remove it! Mhhutchins 07:37, 8 December 2014 (UTC)
- How devious! :) I had to change the title type to COLLECTION, delete the series, then change the title type back. Ahasuerus 16:22, 8 December 2014 (UTC)
Five editions in one?
- Oh, that. I have seen this problem from time to time. Apparently the user interface that Amazon makes available to publishers includes check-boxes for "abridged", "audiobook", "box set", "illustrated" and "large print edition". From time to time some imperfectly-clued-in data entry person checks all or some of them for no good reason and we get a mongrel pub :) Ahasuerus 01:48, 11 December 2014 (UTC)
- I added another note about the OCLC listing for this ISBN which might explain further the situation. Thanks. Mhhutchins 02:55, 11 December 2014 (UTC)
ebook of Macroscrope
Re this publication record: priced at $0.00, giving B&N as the source. But checking B&N just now shows it's not available. It is available at Amazon for $6.00. Should we change the record to reflect the current price? Mhhutchins 05:45, 19 December 2014 (UTC)
- Hm, let me see... When I go to the B&N page and display the HTML code (Ctrl-U), I see that it says:
- "listPrice" : 6.0
- so B&N matches what Amazon has. I am not sure why Fixer thought that the list price was $0.00 on 2014-03-14 -- perhaps it was on sale and something confused Fixer's parsing logic? Ahasuerus 05:52, 19 December 2014 (UTC)
- Or maybe some holiday eggnog? I've changed the price. Thanks. Mhhutchins 19:56, 19 December 2014 (UTC)
Retro-designation of publisher credit
Here is a good example of how the publisher (or Amazon) is back-crediting publications for the current publisher. Gallery Books, an imprint of Simon & Schuster, didn't exist in 2007 (see this announcement dated October 1, 2009). Or the ebook is being dated the same as the first S&S edition, regardless of when it actually appeared. Either way, we'll have to correct the publisher credit or the publication date of this ebook edition. The Amazon "Look Inside" shows the publisher as Simon Spotlight Entertainment, but typically not dated. Another concern: that ISBN and ASIN comes up as unknown on Amazon.com. Strange for what is essentially a US-based publisher (Simon & Schuster). I couldn't find any US published ebook edition on Amazon.
Sorry to bring this to your attention after all the efforts you're putting into getting these ebooks into the database. But it seems that the further back we go (seven plus years in this case), there will be less reliability of the data that Fixer is acquiring. The question becomes when do we simply give up trying to create publication records for older ebooks based on current data? Mhhutchins 17:22, 21 December 2014 (UTC)
- That's a very good question. Using Amazon as our source for older books has always been problematic even when dealing with traditional publications. To make matters worse, e-books present additional challenges since so much about them is ephemeral and can be easily changed retroactively. I was kind of hoping that we could use Amazon's data going back to at least 2005, so your findings are a bit disappointing if not surprising.
- On the other hand, there aren't many other sources of information about older e-books and I think they are fairly important. For a while publishers like Mushroom eBooks were the only source of certain older, our-of-print books. Omitting these ISBNs could significantly distort certain books' publication history.
- BTW, here is how many older e-AddPubs Fixer knows about:
- 2007 - 703 (January - July)
- 2006 - 586
- 2005 - 519
- 2004 - 438
- 2003 - 389
- 2002 - 392
- 2001 - 289
- Going forward, I can think of the following options:
- Stop e-book submissions once we wrap up 2007
- Create skeleton records for pre-2007 e-books, omitting price/publisher data since it can be unreliable
- Do additional research for pre-2007 e-books to ensure that the data is correct (time-consuming, but may be viable given the relatively low number of e-books before 2007)
- Should we move this discussion to the Moderator Noticeboard to see what other folks think? Ahasuerus 17:58, 21 December 2014 (UTC)
- Yes, but it should be presented to the group on the Community portal, even if most non-moderating editors won't get involved. The question involves more than just moderators. I personally feel we've gone back about as far as we can go when it comes to ebooks, but I'd hate to drop the effort entirely. I'd go for #2 which seems a viable compromise, i.e. only create stub records for pre-2007 ebooks, and allow primary verifiers (and interested parties) to do further research to fill out the data. Thanks. Mhhutchins 18:46, 21 December 2014 (UTC)
The title above shows up on Heinlein's Summary page under the series Lazarus Long. On the alphabetical version of his page it doesn't show up at all other than in the Serial section. On the chronological version of his page the title doesn't show up at all .... most odd. --~ Bill, Bluesman 22:57, 21 December 2014 (UTC)
- Yes, it's a brand new (shiny! :-) bug introduced earlier this week when the Summary Biblio logic was reworked (reported by Hauk earlier today.) It will be fixed in the next patch which is currently being tested. Ahasuerus 23:51, 21 December 2014 (UTC)
- Thank you. If it weren't for bugs and gremlins ..... Merry X-mas! ;-))) --~ Bill, Bluesman 23:55, 21 December 2014 (UTC)
I see that all books which were typed as NONGENRE before the change are now typed as NOVEL. Since 99% of them were novels (as the standards explicitly stated only novels should be typed as NONGENRE), I suppose it was the best way to do the conversion. That left about 75 titles which had types that mismatched their publications. Really not bad, considering the enormity of the change. They were all either nongenre collections (such as those by Harlan Ellison, Isaac Asimov and Arthur Conan Doyle), or the publications were typed as NONFICTION and their title records were typed as NONGENRE. (Not sure how those got through to begin with.) Thank goodness, there was already a cleanup report in place that found these. I'm working on those now. This is just a short note to tell you how much I appreciate the effort that went into this. Thanks, as always. Mhhutchins 03:25, 23 December 2014 (UTC)
- Glad to hear things are working out! (And hey, that's why I retired two years ago -- so that I would have enough time to work on these kinds of things :-) I was a bit nervous about this change since it touched so many different areas, but it looks like nothing major got broken (knock on wood.) Ahasuerus 03:48, 23 December 2014 (UTC)
Just figured out how those dozen or so pub records were entered as NONFICTION. It must have been a slip of the mouse, because NONGENRE was only a fraction of a click away from NONFICTION. :) Mhhutchins
Just figured out something else. We can make NONFICTION into NONGENRE as well. Fantastic! For example, we can move Frank Herbert's computer book down to the NONGENRE section. Was that an intentional side benefit of the change, or should there be a discussion about which NONFICTION is genre and which is nongenre? Mhhutchins 03:34, 23 December 2014 (UTC)
- Well, the intent was to allow all title types (except REVIEWs and INTERVIEWs) to be designated "non-genre", but I didn't necessarily think through all the implications. It wasn't until after I installed the patch that I started thinking along these lines. Some ESSAYs may also be arguably "non-genre", but NONFICTION is probably our biggest gray area. We may want to discuss this issue on the Rules page before different editors take it in different directions. Ahasuerus 03:48, 23 December 2014 (UTC)
Cleanup script for "Authors That Exist Only Due to Reviews"
I've been meaning to bring this to your attention, but kept forgetting. Once the problem has been fixed for records that show up on this list (by adding a new title/pub record or correcting the spelling of the unknown author's name), the records don't clear from the list, as they do on all other cleanup scripts. I've corrected the three problems currently listed but they will remain on the list until the script is ran again the next night. (Actually there were four records originally listed, but one of them did disappear from the list. Go figure!) Thanks. Mhhutchins 19:23, 27 December 2014 (UTC)
- Thanks, I'll take a look. I think it was one of the first cleanup reports that I tackled earlier this year and it may not be using quite the same logic as the rest of them. Ahasuerus 19:29, 27 December 2014 (UTC)
- Fixed! Ahasuerus 20:29, 27 December 2014 (UTC)
- Great! Thanks. Mhhutchins 20:59, 27 December 2014 (UTC)
- Thanks, I will try to correct it later today. Ahasuerus 15:08, 28 December 2014 (UTC)
- The bug has been exterminated with extreme prejudice :) Ahasuerus 18:51, 28 December 2014 (UTC)
HTML Character Code Clean-Up
Unless I'm missing something, it seems we could use a clean-up of HTML character codes in titles. I know people used to use them prior to the database being improved for other languages, but now they seem to just cause search problems. Consider the following advanced title searches:
- Title = "Aurelia", Author = "Gérard de Nerval" returns 6 results but NOT 1380959
- Title = "Aurélia", Author = "Gérard de Nerval" returns same 6 results but NOT 1380959
Looking at 1380959 in yesterday's database dump, it is entered as "Aurélia"
Querying "́" returns 180 title records & 102 publication records in the dump.
_title like '%&#%' returns 3350 title records & 809 publication records.
- In most cases, "HTML character codes" are valid because our data is currently stored as Latin-1, which has very limited support for non-ASCII characters. Eventually we will migrate from Latin-1 to UTF-8, but for now we have to live with Latin-1 (even though it has unfortunate side effects) and that means using HTML encoding for non-Latin-1 characters.
- However, after reviewing the data in the database I see that "́" is a separate case. It's a special "Combining Acute Accent character, a part of the "Combining Diacritical Marks" family of characters. These characters are used to modify the preceding character, so what's actually stored in the database is strings like "Sala de Autopsias Número 4", "Harry Potter agus an órchloch" and "Exilé". This can happen during copy-and-paste operations when using third party Web sites as a source -- see Bug 423.
- I think we need to do two things to address this issue. First we need to modify the data entry logic to convert "known culprits" like "ó" and "é" to their Latin-1 equivalents. Second, we need to create one or more cleanup reports to find "&#769;". I'll see what I can do. Thanks for providing additional information about this issue! Ahasuerus 02:42, 29 December 2014 (UTC)
Non-Genre Cleanup Reports
Now that we have the non-genre flag, seems like we could use some clean-up reports to go with it:
- Parent/variant is marked non-genre, but variant/parent isn't
- Container is/isn't marked non-genre, but contents aren't/are marked non-genre
The first one should be pretty straightforward. I was setting the non-genre flag on some records that were pseudonyms when it occurred to me that I needed to do it to the parent as well. While I think I got all of those, this is probably something that will pop-up in the future.
The second one is a bit more complex. In general, there should be agreement between the container and the contents as to whether they are genre/non-genre. Obviously there will be mixed cases (Dangerous Women is an example being I consider it a genre anthology, but it has a number of non-genre stories including two by authors that meet threshold so are included in the publication listing). However, as a start, it could look for cases where it is a fiction container and all the fiction contents are the opposite. This would catch some cases. As more stuff gets the flag set, I'm sure additional patterns will be found.
- The first report already exists. Mhhutchins 17:20, 30 December 2014 (UTC)
- And the non-genre contents of non-genre publications should only be by authors "above the threshold". I don't think a script can make that determination. Mhhutchins 17:20, 30 December 2014 (UTC)
- Agreed, that's why I entered Dangerous Women the way I did. The script wouldn't try to make a threshold determination. It would only look for obvious disconnects - somebody set the publication title record to non-genre, but all the contents are marked genre; somebody set a shortstory to non-genre, but it's also in a chapbook that is marked genre. The clean-up based on the script could either be to set the appropriate flag or delete the record if the "threshold" isn't meet. Maybe I'm being too pessimistic and this won't be needed, but I'm guessing that as people start using the flag, some will forget that both the container and the contents need to be set. -- JLaTondre (talk) 20:10, 30 December 2014 (UTC)
COVERART w/o Cover:
In the last database dump, there are 16 COVERART records without 'Cover:' at the beginning of the title. Quite awhile back, I cleaned some of these up so they are growing, but not fast. I can add it to my list of things to routinely check and can easily keep up with it. As such, I'm not sure a clean-up script is necessary, but rather add a bug ticket to prevent them from happening? I believe it is coming from two sources:
- Title edits - people editing the cover title to match a changed publication title
- Variants - people making a new variant
And in both cases forgetting to add the "Cover:" at the beginning. Seems like the applicable submission error checking could be modified to check for the presence of "Cover:" when the title type is COVERART. Thanks. -- JLaTondre (talk) 14:20, 31 December 2014 (UTC)
- That's a good point. I have created FR 756, which currently reads:
- Enforce the "Cover:" prefix for COVERART titles. The prefix can be dropped when editing title records and creating variant titles. The following safeguards could be implemented:
- 2. Add moderator warnings
- 3. Create a cleanup report.
- Thanks! Ahasuerus 16:03, 31 December 2014 (UTC)