User talk:Fixer/Archive/2009

From ISFDB
Jump to navigation Jump to search

Star Wars: Galaxy of Fear

I've no objection to these. But I also have no desire to go add them to any Star Wars Series/Sub-Series/Sub-Sub-Series. Or create a new one. Good work on finding a new-(ish) Author though, I'm just not so keen on fitting this into a series I have no interest in. I did the Pratchetts though. And some other recognisable but uninteresting authors. Try harder on Interest-levels for Mods please? BLongley 23:42, 9 January 2009 (UTC)

The Star Wars: Galaxy of Fear series pubs was easy enough to process but I'm concerned about the number of children's books as they are often hard to evaluate sight unseen. For example, I just approved Hot Stuff by Hazel Townson but as there are no Amazon reviews for this, nor for a pb edition. wikipedia:Hazel Townson has many specfictish titles which is a good sign but for all we know these are "ghosts and aliens are not real" moral stories. --Marc Kupper|talk 06:27, 11 January 2009 (UTC)
True, children's books can be hard to pigeonhole, but I can't think of a simple solution.
As far as Bill's request for "interest-levels for Mods" goes, well, I can do it by author, so feel free to submit requests. Last time we discussed this issue, I believe I found a few hundred Zelazny ISBNs that we didn't have on file. Would you like me to submit all of them? Ahasuerus 06:51, 11 January 2009 (UTC)
How about looking for SF/Fantasy books like you are already doing but only return adult books either for authors we don't have or titles we don't have. Given the Amazon titles are often decorated one thought is for the author(s) you scan down the list of their titles and string/compare the first N characters where N is the lesser of the ISFDB or Amazon title length. --Marc Kupper|talk 09:11, 11 January 2009 (UTC)
Sure, I can identify all authors who are in my local database of 290,000 Amazon imports but not in the ISFDB. I can then calculate the number of books for each author and start with the most prolific ones. There will be quite a few false hits, e.g. we use "Tim LaHaye" and Amazon uses "Tim F. LaHaye", but once we have a "top 50" list, we can review it manually. Ahasuerus 18:15, 11 January 2009 (UTC)
Is Fixer coded in Python? If so, you may want to look at the Levenshtein code that's available. I imagine this will pull in a bunch of paranormal and related romance unless you deliberately exclude that in the first pass. --Marc Kupper|talk 09:11, 11 January 2009 (UTC)
I've learned enough Python to be dangerous, but I doubt I will ever become good at it. Fixer is coded using the stuff that I have been dabbling in for decades -- we will call it OOA (Object Oriented Abacus) for our purposes -- which speeds things up tremendously, but makes it effectively unusable by anybody else. Trade-offs, trade-offs... Ahasuerus 18:15, 11 January 2009 (UTC)
Before doing the few hundred Zelazny ISBNs I'd rather that fixer be able to do the add-pubs to matching titles as that'll save a merge step. If something get's dropped under the wrong title we can always unmerge it. --Marc Kupper|talk 09:11, 11 January 2009 (UTC)
Yes, that's on my list of things to do. I have been very busy the last 6 weeks (6am-9pm days) and didn't have the energy to do much ISFDB work aside from the basics, but I hope things will stabilize after the inauguration. I really need to reshuffle the database structures before I can start working on more advanced areas... Ahasuerus 18:15, 11 January 2009 (UTC)
The large print edition submissions could definitely do with being automerged to start with - I found a lot of the merges had the new title as the default and the one with Series name, Number, Wikipedia links, notes etc as the alternative, so had to switch all the check-boxes. Not fun. Still, no need to code for that yet if you're going for things we don't have. I'd like a look at the Top 50 missing authors list, that sounds much more interesting. I'd suggest some authors, but when I find an interesting one I tend to research them myself - e.g. see all the Author Updates I submitted today. I found those by picking a British Anthology series (Doctor Who short stories) and finding the authors on Wikipedia or finding their home pages, and researching the notable authors rather than the fan authors a bit more deeply. BLongley 20:46, 11 January 2009 (UTC)

Prices

I had not been paying attention to the list price Fixer came up with but was surprised at a $20.95 for a 1997-10-00 work. I've left the Fixer info-blob in place so you can look. Perhaps Amazon gives you the old list price in the web-services view? --Marc Kupper|talk 06:06, 11 January 2009 (UTC)

The price does look high for a large print tp in 1997, but OCLC 36589214 confirms $20.95 while used.addall.com confirms that it's a tp. I am not sure what's going on here, but the data seems to be as good as we are going to get short of a physical verification. Ahasuerus 06:25, 11 January 2009 (UTC)
Thanks - I've updated the pub-notes. What is interesting to me that while this is out of print and this not showing the "list" price on Amazon that it's still available from web services. How did you get the tp binding? Amazon reports pb though the dimensions reported are tp size. --Marc Kupper|talk 06:35, 11 January 2009 (UTC)
Amazon provides "PackageDimensions Length Units", "PackageDimensions Length content", and "Binding" information, which can be massaged to determine whether it's a tp or a pb. Most of the time :) Ahasuerus 06:47, 11 January 2009 (UTC)
Where are you drawing the line? And is it Length alone? Only I know that with publications like this I've disagreed with other moderators on whether it's a tp or pb. I know you can't use the "if it makes Bill's bookshelf look uneven, it's a tp" category, but I go for tp based on more than height. BLongley 20:56, 11 January 2009 (UTC)
Amazon-US uses "hundredths-inches" as "PackageDimensions Length Units", so Fixer uses 740 (18.8 cm) as the threshold value. I doubt Fixer has submitted many pbs so far, what with the emphasis on Large Print books, so if we want to change the threshold, the impact should be minimal. Ahasuerus 22:25, 11 January 2009 (UTC)
While I don't own any I noticed in a bookstore a year or two ago books that were MM pb size as far as width went but were perhaps an inch or two taller. I did not pull one out but the binding quality appeared to be pb or mm/pb. The owner shelved them with pb. If I got one I'd enter it in ISFDB as a pb with a note. --Marc Kupper|talk 05:39, 12 January 2009 (UTC)
Yes, I have seen them as well. They also tend to cost a couple of bucks more than regular pbs just like those Oz experiments from Del Rey some years ago.
As far as binding quality goes, that's a different can of worms. Technically, a "mass market paperback" is a paperback book that gets pulped if it doesn't sell, with the cover returned to the publisher as proof of pulping. "Trade paperbacks", on the other hand, are returned to the publisher for credit. Therefore it's possible to have a small size (<740 and typically well bound) paperback which is technically a "trade paperback", e.g. some White Wolf books. Conversely, it's also possible to have a large size (>740 and often poorly bound) paperback that is technically a "mass market paperback". We considered using these industry terms when we first introduced the pb-tp-hc schema, but decided against it since we figured that it would just confuse everybody :) and we decided to go with dimensions instead. Ahasuerus 08:53, 12 January 2009 (UTC)
I have no idea if you are saving the records you get from Amazon but it may be worth a scan to see if have anything that's under 425 wide and over 740 tall. --Marc Kupper|talk 05:39, 12 January 2009 (UTC)
Amazon will happily send a lot of stuff to you (check their XSD file) upon request, so you have to pick and choose what you want to request and what you what to store. In Fixer's case, only "Length" and "LengthUnits" are currently stored, but I am sure Fixer could be convinced to go back to Amazon and ask for Width information if needed. Ahasuerus 08:53, 12 January 2009 (UTC)
I'd prefer that "A Format" and smaller continue to be "pb" as anything much larger causes shelving problems. So "B Format" and "C Format" and any other oddities can be lumped together as far as I'm concerned. I'd like Width to be a consideration, but I don't want extra fragmentation, so I'll let everyone else demand separations of "tp" that suit them. ;-) BLongley 21:48, 12 January 2009 (UTC)

Author request

Humphrey Carpenter's "Mr. Majeika" books seem clearly in but as there's a least a dozen titles to add I wonder if Fixer has the rest? BLongley 14:36, 15 January 2009 (UTC)

Sure, I'll check tonight when I have access to the database. Ahasuerus 16:52, 15 January 2009 (UTC)
Fixer has 14 "Humphrey Carpenter" records on file. Here they are (including ISBNs):
  • Mr. Majeika (Lythway Large Print Children's Series) 0745115829
  • Shakespeare Without the Boring Bits! 0754050483
  • Mr. Majeika 0745184650
  • The Letters of J.R.R. Tolkien 0618056998
  • More Shakespeare Without the Boring Bits 0754050815
  • Mr. Majeika and the Dinner Lady and Mr. Majeika and the Music Teacher (Cavalcade Story Cassettes) 0862200962
  • Mr. Majeika and the Dinner Lady / Mr. Majeika and the Music Teacher (G K Hall Children's Audio Books... 0745184669
  • Mr. Majeika (Cavalcade Story Cassettes) 0862200792
  • Tolkien: The Authorized Biography 0345327292
  • Mr. Majeika on the Internet 0754063356
  • Mr. Majeika and the Haunted Hotel and Mr. Majeika and the School Play (Mr. Majeika) 0745144365
  • Mr Majeika and the School Book Week & Mr Majeika and the School Inspector (Mr. Majeika) 0745144918
  • Shakespeare Without the Boring Bits 0754065057
  • Mr Majeika and the Dinner Lady 0816175543
I will submit them tomorrow night. Ahasuerus 04:07, 16 January 2009 (UTC)
Hmmm, only half the Majeika series and none of the first editions. OK, I've added those, it'll make later title-dates easier. I think you can skip the Shakespeare though. BLongley 17:38, 16 January 2009 (UTC)
Fixer has now submitted all the outstanding Mr. Majeika pubs. Apparently, the reason that half the titles were not captured is that Amazon.com doesn't consider them speculative fiction. In 5 cases I had to ask Fixer to override his auto-suspend logic triggered by the fact that there were no SF Browse Nodes or Subjects at Amazon. After all, there are many elementary school teachers who are disguised wizards from outer space, so what's speculative about it?..
P.S. Fixer may have better luck when he begins munching on Amazon.uk and Amazon.ca data. Ahasuerus 06:36, 18 January 2009 (UTC)
Fixer will probably acquire indigestion... but yes, there'd be a lot of value in grabbing UK data especially for UK publishers and authors. Unfortunately we've already added a lot of Amazon.com versions of British books and given them strange List Prices that I can only imagine were direct Pound/Dollar conversions. Still, there's lots of BBC books and suchlike that Dissembler missed. BLongley 15:23, 18 January 2009 (UTC)

Top authors

Now that Fixer's anti-comics/games AI has been upgraded, the "top authors not in the ISFDB" list can be revealed (with preliminary comments in square brackets):

234  Katherine Applegate [instead of the regular K. A. Applegate]
94   Golden Books [publisher]
81   Steve Jackson [games]
76   Stephen Cosgrove
69   Arthur Conan, Sir Doyle [we use Arthur Conan Doyle and Sir Arthur Conan Doyle]
66   DK Publishing [publisher]
64   Cicely Mary Barker
64   James Gelsey
59   Philip Jose Farmer [instead of Philip José Farmer]
57   RH Disney  [publisher?]
55   Steve Jackson Games [games publisher]
53   BradyGames [publisher?]
51   Fanpro  [publisher?]
49   Barnes
48   Dr. Seuss
48   John Lord
47   Rh Value Publishing [publisher]
46   FanPro [publisher?]
46   Paul Jennings
46   Sigmund Brouwer
44   Steve Kenson
39   Brandon Robshaw
37   Jesse Leon McCann
36   TSR Inc [publisher]
36   Various Authors [sic]
34   Arthur Thomas Quiller-Couch
34   Tsr [publisher]
33   Tim F. LaHaye [instead of Tim LaHaye]
32   L. E. Modesitt [instead of L. E. Modesitt, Jr.]
32   Marilyn Sadler
31   FASA Corporation [publisher]
31   Susannah Brin
29   Arthur Charles Clarke [Arthur C. Clarke]
29   Brian Wilson Aldiss [Brian (W.) Aldiss]
29   Deborah Hautzig
29   Herbert George Wells [H. G. Wells]
28   Barbara Mitchelhill
28   Peter Lancett
27   Dave Barry and Ridley Pearson
27   Michael
27   Random House [publisher]
27   Walt Disney [publisher]
26   Charlotte Bronte [non-sf]
26   Constance M. Burge
25   Howard Schroeder
24   John Snead
24   Katherine A. Applegate [K. A. Applegate]
24   Star Wars [series]

I'll post more once we have a better grip on these folks. Ahasuerus 07:00, 20 January 2009 (UTC)

First thoughts:
"Sigmund Brouwer" has the "Mars Diaries" and "Cyberquest", both look in as YA SF.
"Stephen Cosgrove" has the Serendipity series which looks in, but it's Reading level: Ages 4-8 and not a high priority in my eyes.
"Cicely Mary Barker" has the "Flower Fairies" books, again a bit too juvenile to interest me.
"Paul Jennings" has the "Wicked" series - probably in, juvenile again.
"James Gelsey" - Scooby Doo books and Dragon Booster.
"John Lord" appears to be "Beacon Lights of History" - surely not really SF?
Beware of "Constance M. Burge" - they're probably titles in the "Charmed" series and in the dozen examples I own the actual title-page author is always different. The spine credits Burge and the cover says "Based on the hit TV series created by Constance M. Burge" which is why Amazon miscredits them.
"Steve Kenson" looks too RPG.
"Brandon Robshaw" - "Livewire Chillers" looks in.
"Jesse Leon McCann" - more Scooby Doo.
"Susannah Brin" - looks to have some interest, but Amazon has few details.
"Deborah Hautzig" - "Little Witch" for very young readers.
"Barbara Mitchelhill" - "Livewire Chillers" again.
"Peter Lancett" - graphic novels?
"Dave Barry and Ridley Pearson" - Peter Pan/Neverland sequels?
"Howard Schroeder" - several juveniles of SF interest, but few details on Amazon.
"John Snead" looks too RPG. BLongley 19:31, 20 January 2009 (UTC)
Thanks! I will get back to it in a couple of days once I am done cleaning up Russian vts and synchronizing with Contento. The number of misattributions, misspellings and other headaches is mind boggling (I think at least one of the editors was on drugs), but thankfully it's only a few dozen pubs and I can almost see the light at the end of the tunnel. Ahasuerus 02:47, 21 January 2009 (UTC)
"Paul Jennings" has now appeared here due to another Fixer submission, so might be worth an author run at some point. BLongley 20:10, 10 May 2009 (UTC)

Authors by Genre and library vacuuming

The "50 Top Authors not in the database" approach didn't seem to be too promising, so I tried something else tonight. There are about 700 "Time Travel" ISBNs in Fixer's database that are not in the ISFDB and the first 3 submissions found 2 new paranormal romance authors, Natale Stenzel and Brenda Joyce. I then asked Fixer to submit the rest of their books, which generated a complete specfic bibliography for Stenzel and will generate what appears to be a near complete specfic bibliography for Joyce once all submissions are approved. Not a bad ratio, overall, so I will play with it some more over the weekend.

I am also working on improving Fixer's library parsing algorithms now that I have a little free time during the day. At the moment, Fixer can ask libraries for all kinds of data, e.g. give me all books published in 1986 whose subject list includes "science fiction", which is a good start. Unfortunately, Fixer can't parse everything that libraries return, at least not yet. Authors, titles, publication years and other basic fields are OK, but there are all kinds of other obscure fields that may contain potentially useful information. I don't want to submit incomplete records and it would take a long time to teach Fixer how to handle hundreds of possible fields and subfields, so I'll probably compromise, dump uncommon fields in Notes and leave it up to the approving moderator to sort them out.

Once the logic has been cleaned up, I'll try the Library of Congress and see how much damage Fixer can do. If it works reasonably well, Melvyl will get raided next. The British Library will be a special case since they use a somewhat uncommon record format, but it doesn't look too bad. Ahasuerus 04:35, 23 January 2009 (UTC)

The plot thickens! Your sinister plans have been revealed. Once everyone has determined that everything has been entered, you'll pull out the Master Fixer, and all moderators fall to their knees in supplication to Lord Ahasuerus. (Makes the disciples of L. Ron look like wimps!) MHHutchins 06:13, 23 January 2009 (UTC)
Shhh, don't give Fixer ideas! The other night I noticed that he was running searches on "author = 'Jones' AND title = 'Colossus'". Not sure what that's all about, but it made me vaguely uneasy... Ahasuerus 15:11, 23 January 2009 (UTC)

Fixer's Progress

Completed Authors

Authors currently in progress

Secrets: The Best in Women's Sensual Fiction

Good Find with Chris Roberson

Not for the author directly (I haven't checked those much) but "Tales of the Shadowmen" is interesting. From the covers alone I'd thought "Comic" or "Graphic Novel" but I downloaded some samples and there are real stories (with words and sentences, even!) in there. And by some well-known authors too. I've managed to add contents to three (I got delayed by "Gaslight Grimoire", another good find) but will resist the fourth for now. I haven't approved any - there's a lot of cedillas and accents and probably missing apostrophes and capitalisation too, but I think the spin-offs from those might well be more interesting than all the Paranormal Romance stuff. BLongley 22:16, 26 January 2009 (UTC)

Obligatory moan - if there's over half-a-dozen authors on a publication, this might be a good sign that it's an anthology rather than a novel? Don't stop submitting such - they're far more interesting than the "Paranormal Romance" - but they are hard work and as I've only done four this evening you might want to count such as a bit harder work than normal and limit the number of such a bit (unless other mods are getting interested in what Fixer is doing). BLongley 22:16, 26 January 2009 (UTC)

You should have seen his page before I sorted it out! :-)Kraang 03:24, 27 January 2009 (UTC)

Enough romance for now?

I have been trying to tie up all the loose ends (romantically speaking) and move on to something else, but this sub-genre is so incestuous (in a manner of speaking) that it's easy to get bogged down. How about we approve what's in the submission queue, move the current list to a Project page and do something more interesting with our time? <he said impetuously and fiercely glared at the queue> Ahasuerus 02:45, 29 January 2009 (UTC)

I'd have a look at Fixers submissions but I'm tied up with "Non-SF Cleanup 2004". One of the novels turned out to be an anthology edited by Eric T. Roberts who publishes Hadley Rille Books. This lead to 5 more anthologies edited by him that weren't in the data base. I'm now in the process of adding in all the titles(found on his web page), which has added about 80-100 new authors. Only have one more book to do and I can start on "part 12" of the non sf cleanup.Kraang 04:00, 29 January 2009 (UTC)
80-100? Ouch! Feel free to add any identified authors to Fixer's list -- it takes but a minute to find and submit all books by a given author (unless his last name is Asimov or Silverberg, of course) and it can save quite a bit of time. Ahasuerus 16:36, 29 January 2009 (UTC)
We definitely need a change. I'm fed up with Ellora's Cave, although we'll no doubt have to return to it eventually, with categories like "Futuristic/Sci-fi", "Paranormal", "Time Travel", "Vampire" and "Werewolf/Shapeshifter". I can wait till Fixer learns to trawl that site directly. And a bit longer. I could become VERY good at waiting in fact. BLongley 21:23, 29 January 2009 (UTC)
OK, I have cleaned almost everything up. Once the current batch of submissions has been approved, we will have only "Erin Hunter" and "Jayne Castle" outstanding. Then Fixer can go hunting for juicier prey. Ahasuerus 04:54, 30 January 2009 (UTC)

Library catalogs and SUTRS

BTW, the SUTRS format that the British Library is using is uglier than I realized - guess what the "U" stands for... Ahasuerus 04:54, 30 January 2009 (UTC)

I know it's "Unstructured", but you make it sound like "Unusable". Shame, as the first "S" is supposed to be "Simple"... BLongley 18:25, 1 February 2009 (UTC)
More like "Unacceptably time consuming". Anything can be done given enough time and resources, but that rarely happens... Ahasuerus 02:20, 2 February 2009 (UTC)

At one point (ca. 2000) they tried converting to the MARC standard, but apparently decided that it was hopeless and gave up. Oh well, I'll still try to extract some useful data out of their database at some point, but it will have to wait until other things are done. Ahasuerus 04:54, 30 January 2009 (UTC)

I think they actually had more success than that - unfortunately you have to pay for the MARC21 data. I guess it must have cost a fortune to do. Perhaps I should sue them for cybersquatting and demand compensation in free data? It's obviously MY domain (BL.UK) that they're using. BLongley 18:25, 1 February 2009 (UTC)
It looks like BL's data is available via the COPAC Z39.50 interface, which serves it as XML, GRS-1 or SUTRS. I'll play with the XML to see if it's any better, but for now I will likely concentrate on MARC-21 (and its friends). You have to start somewhere. Ahasuerus 02:20, 2 February 2009 (UTC)

Doing Publishers

Done. I hope "S. M. Stirling" wasn't the juicier prey? As that was just a pain, putting all the full stops back in. BLongley 18:25, 1 February 2009 (UTC)

I am still looking for a good algorithm that would find "juicy prey". Stirling was just an experiment to see how a more recent author would do. So far, we have tried:
  • top authors not in the ISFDB - not very useful
  • searching by subject ("time travel") - ran into "romance issues"
  • an older author (Alfred Bester) - ran into imaginary ISBNs and bad publishers
  • a more recent author (S. M. Stirling) - identified a pseudonym ("Steve Stirling" confirmed by a cover scan of a UK edition), but otherwise not very profitable
I am thinking about doing it by publisher for 1995-2009, e.g. Tor (350ish hits) or Roc (50ish titles), for 1995-2009, but it would split authors. Do you think it would that work better? Should I ask Fixer to submit Roc since it's more manageable? Ahasuerus 02:20, 2 February 2009 (UTC)
P.S. Roc submitted. A few good finds so far, including one previously unidentified anthology. Ahasuerus 11:33, 2 February 2009 (UTC)
About four actually, depending on how unidentified or incomplete you think they were. I probably messed up a few submissions before I realised how bad the publisher data was - but then again, a realisation that "Roc never had those ISBNs!" led me to fix some existing data as well when I went back. Quite satisfying though, and I'd like a few more like that. BLongley 00:25, 3 February 2009 (UTC)
Sounds good! Pyr and Ace submitted (all 85 of them.) Ahasuerus 05:20, 3 February 2009 (UTC)
I didn't see the Pyr, but most of the Ace were OK, if a little too far in the future for some publications. BLongley 19:27, 3 February 2009 (UTC)
That's because Michael got to most of the bad ones first -- take a look at his battle royal with the 999s in Recent Rejects :) Ahasuerus 02:20, 4 February 2009 (UTC)
(There are some suspicious "07865" prefix ISBNs, but those all look a bit suspicious to me, it's not just Fixer's submissions.) Can Fixer take series requests? It looks as though "The Dennis Wheatley Library of the Occult" needs some attention, for instance. But we'd get that if you did a run on "Sphere" I guess, and "White Lion" for the hardcovers. BLongley 19:27, 3 February 2009 (UTC)
Fixer can only craft submissions based on the information in his database, which at this point comes exclusively from Amazon.com. If Amazon doesn't mention this series in the Title/Publisher/etc fields, then Fixer has no way of telling what's going on. At the moment, I see 3 books in this series in the database: Williamson's classic Darker Than You Think, Dunsany's Curse of the Wise Woman (#40 according to our Notes), Shelley's Frankenstein (ISBN 0722177720, not in the ISFDB). I also see Dennis Wheatley's Second Book of Horror Stories: Tales of Strange Happenings (ISBN 0090863607, Hutchinson Radius, January 1968), which we don't have on file either. We may have better luck with library catalogs once Fixer learns how to do useful things with their data. Ahasuerus 02:20, 4 February 2009 (UTC)
Number 40 for Curse of the Wise Woman seems right according to this apparently complete listing of the Sphere editions. As you can see, there's several ISFDB authors represented, and even more once you find the contents of the three "Uncanny tales" anthologies. But as I say, a run on all missing Sphere books from Amazon UK would probably drag most of these in. I did the White Lion editions manually yesterday, but there may be more. For contents, obviously library catalogues would be better, but the lack of capitalisation and omission of leading articles often means I'm retyping 10-20% of each one after a copy'n'paste. BLongley 20:07, 4 February 2009 (UTC)
Different libraries have different quirks. Some catalogs drop leading articles while others insist on using initials for stories by "I. Asimov" and "R. Heinlein" and all of them store Contents level data in 5XX fields, which are poorly structured. I won't even try to create Contents entries automagically, at least not on the first pass, I'll just ask Fixer to dump everything in Notes. Ahasuerus 02:55, 5 February 2009 (UTC)
P.S. I hope to get to Amazon UK over the weekend, but there is a lot of stuff going on at the moment, so I can't be 100% sure. On the plus side, I finally figured out why Fixer was having trouble with certain types of catalog queries. It was so obvious, I sent him to bed with no supper... Ahasuerus 03:04, 5 February 2009 (UTC)
Take a look at VRLGNDFPST2007. Did you intend to dump all the extra info into the notes? This was also a duplicate that appears to have been double submitted, once with all the notes, and once with the link to amazon alone. I put Cover artists on all the new Pyr. This is the only Pyr record without a cover artist now (assuming it's a mistake or an experiment) Kevin 00:43, 4 February 2009 (UTC)
Hm, this is rather peculiar. This ISBN is not in the backup that I have installed locally, so both pubs were quite recent. I have also checked Fixer's database and he created only one submission for this title yesterday. I see that Michael approved and massaged the submission at 12:57am ISFDB time and that was it until you updated one of the records today. I am not sure where the second record came from, but I have now deleted it, so we are back to normal. Strange, though...
As far as Moderator Notes go, Fixer always creates them to help moderators determine whether the book is specfiction or not, but moderators generally delete them during the cleanup process after approval. Ahasuerus 02:20, 4 February 2009 (UTC)
I figured as much re: the Moderator note, which is why I pointed out that both entries had the 'smell' of fixer submissions. I can also confirm that neither record was there yesterday afternoon as I was already working on Pyr and had completed that year/section of entries and then went back today to double check before calling ti quits on that tangent. Kevin 03:25, 4 February 2009 (UTC)


Processing old Amazon.com data

I submitted Alfred Bester earlier today and it worked reasonably well. Fixer found a few Spanish pubs, a few recent reprints and a couple of UK editions from the 1970s/1980s, which were later confirmed by OCLC. Unfortunately, older books -- as listed by Amazon.com -- are a separate can of worms since Amazon uses the publisher's current name. Thus, all NALs become Rocs and Signets are anyone's guess. Amazon also loves to concoct ISBNs for 1960s books out of thin air (or a close approximation thereof) and it can be hard to tell whether we already have the pub on file.

Oh well, at least it was educational. I think I'll ask Fixer to leave pre-1990 books alone for now. Wouldn't want our moderators to spend the next week trying to sort out 400 Zelazny pubs, most of them wrong. Ahasuerus 04:41, 1 February 2009 (UTC)

Hmmm. I was hoping we could go back to nearer 1970, when at least British pubs were likely to have good (I)SBNs. I think we probably could if it's only the publisher that's in doubt - those are often fixable with enough knowledge of when a publisher existed, and Amazon UK at least often puts the correct Imprint after the title even when it's mangling the publisher field. E.g. I've done such things as search for "(Tandem Science Fantasy)" on Amazon UK and usefully used the results. This might require more knowledge than most moderators currently have though. Still, if Fixer can read Amazon UK, then this sort of search should be great for filling in gaps in known British SF Publishers, and knowing the Imprint/Publisher you're trying to find should make it easier to correct the duff publisher data. BLongley 18:55, 1 February 2009 (UTC)
No worries, even though I am using 1995 as the cutoff date for now, the earlier stuff is not going anywhere and will be processed in due course of time. Resistance is futile, you WILL be cataloged!! Ahasuerus 05:22, 3 February 2009 (UTC)
By the way, even 1990 is too far back to trust Amazon publisher data. For instance, see all our "Wizards of the Coast" data for 1992 and 1993 - for TSR books they didn't acquire until 1997. I'm not sure where the cut-off should be, but I wouldn't think it's before 1995, when Amazon started up. Presumably they got their new stock 'correctly' classified, but all their historical data is questionable. BLongley 18:55, 1 February 2009 (UTC)

999 ISBNs

Hey, Fixer! Get with the program! Everybody knows books with 999 ISBNs stink to high heaven. Do I have to send my cousin Louie around to rough you up a little? And I don't mean the gentle massaging I've been doing with your other submissions! (:D) MHHutchins 06:57, 3 February 2009 (UTC)

Fixer is very sorry about the 999/555 mess, he really should have known better! He was still recovering from taking an advanced "Catalog Raiding 606" class at the Academy for Data Transmogrification, so at least he has an excuse. He promises not to do it again. (And he can't talk at the moment since he is cramming for his "Amazon UK for bots 404" exam.) Ahasuerus 02:35, 4 February 2009 (UTC)
P.S. Make it "Fixer won't do it again once all those Baen submissions from today have been processed" :( Ahasuerus 03:31, 4 February 2009 (UTC)
What I'm finding with these Baens are the multi-paks. Luckily, they're very easily spotted. MHHutchins 04:11, 4 February 2009 (UTC)

And another thing. How come you get to make submissions without authors and us oxygen-breathers get a big knock upside our heads when we do the same thing? Looks like favoritism, if you ask me. You got a cousin Louie, too? MHHutchins 07:02, 3 February 2009 (UTC)

Fixer uses our official backdoor, aka the Web API, which performs limited validation (and doesn't know how to handle apostrophes), so I suppose we could call him Louie... Ahasuerus 02:35, 4 February 2009 (UTC)

But I gotta give you your propers, man. Your last four submissions have been smooth as butta. How'dya learn to merge with title records AND place records into a series with one submission? Just between you and me, fella, I'm gettin' to like you. (But keep it on the down-low, will ya? Wouldn't want to start some ugly rumors now. You never know who might be lurking around the corner.) MHHutchins 07:08, 3 February 2009 (UTC)

If the pre-existing Title record was in a series, adding another pub to it doesn't affect its series status :) Ahasuerus 02:35, 4 February 2009 (UTC)

China Miéville

Can I take it that Fixer has problems with accents? These are being submitted as China Mieville and not attaching to the existing titles. Of course, the approval picks up the canonical name with accent and the merge is simple. BLongley 18:07, 7 February 2009 (UTC)

Amazon lists this author as "China Mieville" and Fixer has touching (if misplaced) trust in Amazon's data. I suppose what I need to do is to adjust Fixer's look-up logic, which currently uses "aaronallston" for "Aaron Allston" and so on. Once the logic uses "chinamieville" rather than "chinamiéville", auto-merge should work fine for accented names. Ahasuerus 18:44, 7 February 2009 (UTC)
Done. Ahasuerus 03:41, 11 February 2009 (UTC)

A worse one was by Kathleen O'Neal Gear and W.Michael Gear - the submission with " rather than the ' in O'Neal meant that when adjusting the author names (to put the space back after the "W") the edit screen left Kathleen's name cut off at the "O". Is Fixer going to learn to handle apostrophes better? BLongley 18:07, 7 February 2009 (UTC)

Oh, the old double quote bug still lives! That's pretty bad :( Ahasuerus 18:44, 7 February 2009 (UTC)
It's livable with, so long as people know it's there. I've had to correct some " entries back that other Mods have let through though. BLongley 20:37, 7 February 2009 (UTC)
Unfortunately, Fixer is limited by the Web API, which refuses to recognize "&apos", so Fixer has to either accept that all new apostrophes-afflicted records will have "&apos" embedded or use some kind of workaround like the double quote that he has been using for the last couple of weeks. I'll play with various Unicode permutations to see if there is anything else I can do.Ahasuerus 18:44, 7 February 2009 (UTC)
Have you tried getting a \' through somehow? I know that's how the final MySQL update escapes it. Or double single-quotes? BLongley 20:37, 7 February 2009 (UTC)
Doubling single quotes did it, thanks! :) Ahasuerus 04:06, 11 February 2009 (UTC)
In other news, Fixer has finally mastered the fine art of catalog munching and is done downloading all readily identifiable speculative fiction records (under 60Mb, so not too bad) from the Library of Congress. Melvyl, the Merril collection and a bunch of other major and/or SF-related libraries are next as soon as I am done teaching Fixer how to multi-thread. Once he is up and running in that area, I'll let him feast -- it will take a few days to download everything from the big boys -- and work on Amazon UK. Ahasuerus 18:44, 7 February 2009 (UTC)
Good news indeed - might be a bit late for Amazon UK in some cases, but I'll take that to a separate thread. The Tor run has taught me some things. BLongley 20:37, 7 February 2009 (UTC)

Tor

This might be the first Publisher where we've run into this problem: Tor is a UK Publisher/Imprint too. But by concentrating on Amazon.com, I think you're letting in [IMPORT] data with duff prices. (Don't worry too much, it seems Dissembler has too.) Some of these were fairly easy to spot - e.g. I'm pretty sure no US edition has really had a list price of "$17.62" for instance. I'm pretty suspicious of "$37.20" as well. And "$12.39". And "$11.39". And "$14.15". And "$26.85". In a few cases where the Tor price looked extremely dodgy in US dollars but fine from Amazon UK in Pounds, and it was a UK author, or a pseudonym mostly used for UK editions, or had a Pan ISBN prefix, and Look-Inside confirmed it, I've adjusted Fixer's entry to take the UK price instead and stated it came from the UK site. I don't think I've fixed as many as I should have, but then I don't know US Kibblesworths [1] and if I had no better data to work from I've left it alone, duff price included. The Tor books where this occurred were identical in all other respects down to publication date. Sometimes the clue is the [IMPORT] suffix on the US site, but that wasn't always present. But I'm pretty sure there are Tor UK titles being misrepresented as Tor US ones, where the current US publications of the time were actually still with other US publishers/imprints such as "Bantam Spectra". BLongley 21:36, 7 February 2009 (UTC)

So, a few questions: can Fixer spot the [IMPORT] and leave those for later? You'll pick up a lot of those from Amazon UK with better data, I think. Maybe Amazon Canada too, although the current dual-ISBN discussion makes me think that Fixer will either give us a headache or he'll abandon the idea. The Tor Australia ones are probably worth keeping. BLongley 21:36, 7 February 2009 (UTC)

Let me see where that [IMPORT] label comes from... Ahasuerus 00:36, 8 February 2009 (UTC)
I see (elsewhere) that you found it, and found it wanting. 50% is better than nothing though. BLongley 22:55, 9 February 2009 (UTC)

If not, can Fixer spot silly prices that can realistically only have come from direct currency conversions? (Just spot the silly prices for now, I don't think we've gathered enough data to guess Amazon's exchange rate for various dates.) BLongley 21:36, 7 February 2009 (UTC)

Unfortunately, this isn't as easy as it sounds since "library binding" books can have "silly prices" like $17.62 :( Ahasuerus 00:36, 8 February 2009 (UTC)
I did a Human check on "silly prices" yesterday, from my last backup. Got a bit OCD even, and adjusted all the prices with a "Can" in the price field. I will probably have to do so again after the next refresh, I suspect that this isn't an old problem, it's an ongoing one. But I found a load of Floor Displays too. There is some value in "Silly Price" checks even if we can't do it programatically yet. And no, "$17.62" doesn't look like a valid price even for Library editions. BLongley 22:55, 9 February 2009 (UTC)

[1] I should explain "Kibblesworth" in case you haven't read the Meaning of Liff. It's the amount below a nice round number that a price is set at. So if a price is $5.99, the Kibblesworth is a cent. If a price is £5.95, the Kibblesworth is 5 pence, or a shilling for us older folk. These days, I'm suspicious of any UK Kibblesworth that isn't 0, 1, 5, or 50 on a single book. It seems 2 is a valid Kibblesworth on some pairs of books, 3 on trilogies, etc. Given series like "Wheel of Time", I can believe a Kibblesworth of up to and including 10 on boxed sets. But some prices are just alarm bells to me. However, I'm aware that US prices on some goods are designed so they look good BEFORE tax is added so can't guess what US Kibblesworths might make sense to anyone there. How much does the Sales tax on books vary? BLongley 21:36, 7 February 2009 (UTC)

Since our published prices are before tax, (and your published prices are presumably tax inclusive) but both are 'prices the public sees'... the kibblesworth warning flag works on both sides of the pond with the same numbers... though you might add 25 to your list of likely kibblesworths.Kevin 21:53, 7 February 2009 (UTC)
25 might be a good number now, I'd have to check some more recent UK books to be sure. It certainly was a good number in older times, but then again any 5p price point was good when books cost under a pound. I guess I'm asking for some common sense: which is not actually that common, and I don't know how best to impart it. For instance, I know instinctively that an old 3/5 British price is WRONG and should be 3/6. Real example: a paperback at £4.83 on Amazon UK is WRONG, as is the £5.52 "List price" that Amazon UK tell me it is really. "US List Price: $7.99". Ah, I can believe that one. Does Amazon US quote UK list prices and convert and then discount? BLongley 23:30, 7 February 2009 (UTC)
And yes, UK prices are Tax-inclusive for consumers. No variation across England at all. Probably not across Wales either. Scotland may go weird on us. Northern Ireland - not going to go there. BLongley 23:30, 7 February 2009 (UTC)
And I heard that Northern Ireland was a wonderful place to visit.....Kevin 23:37, 7 February 2009 (UTC)
US sales taxes are imposed by states, which also determine what goods are subject to taxation. Some states have no sales tax at all while most hover between 4 and 7%. Some expensive areas like New York City have additional taxes, which can be fairly high, and tourist-oriented industries (hotels, rental car companies, etc) have extra taxes slapped on top, which can add up to 20-30%+, I would guess. Ahasuerus 00:36, 8 February 2009 (UTC)

Amazon "formats"

Now that I know where to look, I see that Amazon.com has all kinds of "format" information. Here are the formats that we have on file for books that have not been auto-rejected or auto-suspended:

*.exe                         21
*3.5 inch diskette            2
*Abridged                     1554
*Adult                        2
*Audiobook                    4799
*Bargain Price                33
*Black & White                6
*Box Calendar                 1
*Box set                      257
*Braille                      11
*CD                           1053
*CD-ROM                       2
*CLV                          12
*Classical                    8
*Collector's Edition          8
*Color                        2
*Compilation                  10
*Deluxe Edition               1
*Engagement Calendar          2
*Enhanced                     1
*Facsimile                    10
*Folded Map                   1
*Full Screen                  1
*Illustrated                  190
*Import                       9739
*Large Print                  2482
*Limited Collector's Edition  5
*Limited Edition              4
*MP3 Audio                    543
*NTSC                         1
*Print                        1
*Single                       2
*Soundtrack                   2
*Special Edition              24
*Special Limited Edition      29
*Student Edition              19
*Teacher's Edition            2
*Unabridged                   3216
*Unknown format               6
*Wall Calendar                4
*Wall Map                     2

I plan to do the following:

  1. reject all maps, "*.exe"s, calendars and NTSC
  2. suspend all Imports
  3. use "MP3 Audio" and "CD" as the binding if no other binding information is available
  4. reject 3 non-genre "classical" books (Schubert, Tchaikovsky and The Romantic Era: 1820-1900) and submit the 6 that are genre-related
  5. teach Fixer to ignore "Bargain Price" and to add any other format to Notes

Anything else that we need to do with formats? (And what the heck is "CLV" anyway?) Ahasuerus 00:03, 10 February 2009 (UTC)

I think in this case CLV = "Constant Linear Velocity" (as opposed to "Constant Angular Velocity") and would apply to optical discs of some sort - probably DVDs or CD-ROMs now, but possibly the older laser-discs too. Of course, Amazon may have decided that any series that reached Volume 155 is worthy of note with a separate binding. Or they are selling spaceships now and it stands for "Crew Launch Vehicle". BLongley 19:38, 10 February 2009 (UTC)
Upon closer examination, it turns out that Amazon often slaps the "Import" label on what appear to be legitimate US edition, so it's even less usable than we thought :( Ahasuerus 20:02, 10 February 2009 (UTC)
OK, how about doing what the usual Amazon display does and show the [IMPORT] after the pub title? That should be a nice warning to Mods to check other references. BLongley 20:57, 10 February 2009 (UTC)
That's a good point, I'll try to make IMPORTed records more visible. In the meantime, Fixer has learned how to multi-thread catalog requests and is merrily absorbing the collective wisdom of dozens of libraries. Unfortunately, not all library catalogs support the same search criteria (subject, publication year, etc), so Fixer may need multiple passes to get 95%+ of the eligible data, but it looks promising at the moment. Converting the raw data into useful submissions will be a whole different headache, though. Ahasuerus 21:20, 10 February 2009 (UTC)
I look forward to the next Fixer run then. I can't promise it as much attention as some of the last, if it comes before the weekend - I've had three "snow days" off work recently ("ice days" really - I can drive through snow OK, it seems, and through the floods that have ensued since the thaw started, but two inches of black ice were a bit too much for my car) - but I'm back to travelling again and can't promise Fixer approvals even though they're getting interesting. Even if the Tor run introduced a load more "Paranormal Romance" authors. :-/ BLongley 22:55, 10 February 2009 (UTC)
Fixer is definitely not ready to submit the library data that he has accumulated. There is a lot of useful information there, but the more I look at the data, the more daunting the task looks, so I'll probably put it on the back burner while letting Fixer have his fun with the downloads. Fixing the upload logic to handle Imports, apostrophes and what have you is next on the list, then I'll need to take another swipe at Amazon UK. I have some free time right now (and, Cthulhu willing, for the next couple of weeks), so hopefully it won't take too long. Programming is quite therapeutic if you don't do it too often :) Ahasuerus 01:26, 11 February 2009 (UTC)
P.S. An "Import" warning has been added. Ahasuerus 03:30, 11 February 2009 (UTC)

Amazon UK

Well, it wasn't as bad as I had feared. There are some differences between the two locales, but most of them do not affect Fixer, so now he is happily downloading all kinds of information about Amazon UK's Browse nodes. You can only get up to 4,000 items in response to a query, so large nodes need additional TLC, but I already have that in place for the Amazon US side. I expect the nodes to be downloaded overnight (1 request per second to comply with Amazon's rules) and then the real fun will start. Ahasuerus 03:17, 12 February 2009 (UTC)

Fixer has been on hold for the last week since I have been extremely busy with non-ISFDB issues :( Ahasuerus 02:02, 20 February 2009 (UTC)
Fixer has learned how to process UK data and is now busily downloading everything that he can get his paws on. It will likely take a few days to get everything since Amazon imposes a 1 second limit on the frequency of data downloads and there are some 150,000+ records (times 2) to massage. Ahasuerus 06:32, 22 February 2009 (UTC)
The basic browse node data, i.e. the most prominent 150,000 specfiction records, has been imported. Fixer is now munching on the fringes, which will take a while since he has to work around Amazon's limit of 4,000 books per BrowseNode request. It's doable, but not exactly pretty and will take a few days -- and, of course, the supported sort orders and such are not the same in the US and the UK. Then he will have to do the same thing with "known speculative fiction subjects" and then we should be done. I know, I know, diminishing returns and all that, but we wouldn't be into genre bibliography if we were not irredeemable perfectionists :) Ahasuerus 01:17, 25 February 2009 (UTC)
All browse nodes processed for Amazon UK and we are up to 200,000 records. Starting subjects... Ahasuerus 03:24, 1 March 2009 (UTC)
The brain is back to "mostly functional", resuming the work on Subjects. Ahasuerus 02:04, 6 March 2009 (UTC)
First Subject pass done. Amazon burped on Michael Bishop for some reason (what do they have against Michael Bishop anyway?), but it's been resolved and now the second pass is in progress. This one will take longer since it requires setting up price buckets and other ugly things to get around the 4,000 limit. Ahasuerus 03:00, 7 March 2009 (UTC)

Star Wars

Just a comment - so much Star Wars at once is a bit of a turn-off. I know it's not your fault that the publisher did it. But there might be better ways of choosing and/or ordering submissions. One author should probably be submitted in one go, as it's annoying to clear the queue and do a dozen Brooks or Eddings or Jordan merges, and then do the same merges again the next time, and the next, and the next. Similar with Star Wars - grouping/submitting by author might make them more manageable tasks. BLongley 23:44, 27 February 2009 (UTC)

A few weeks ago Fixer tried submitting by author. Bester was quick and easy and Zelazny (~500 books) was next in line. However, after chatting with Fixer, I realized that this author-centric approach would have resulted in submitting all of Zelazny's UK editions from the US database, messing up (at least) all UK prices. I then asked Fixer to stick with "known US publishers" like Del Rey and Roc until we had all of Amazon UK's data on file. We know that Tor publishes books on both sides of the Atlantic, but the US data should still be correct for the US version of the book. Hopefully :) Ahasuerus 03:55, 1 March 2009 (UTC)
Once the UK data is ready, I can teach Fixer how to use both databases at the same time and add any discrepancies to Notes. It would help to have a list of "primarily UK publishers" so that Fixer would know which price to put in the Price field and which one in Notes. Ahasuerus 03:55, 1 March 2009 (UTC)
Here's a few to get you started: Aldine, Allen & Unwin, Allen Lane, Allison & Busby, Armada, Armada Lions, Arrow, BBC, Badger, Beccon Publications, Big Finish, Bodley Head, Boxtree, Brown Watson, Cape, Cassell, Century Hutchinson, Chatto & Windus, Collins, Corgi, Coronet, Dobson, Eyre & Spottiswoode, Eyre Methuen, Faber, Fontana, Four Square, Futura, George Allen & Unwin, Gollancz, Grafton, Granada, Hamlyn, Hart-Davis, Heinemann, Hodder & Stoughton, Hodder Headline, Howard and Wyndham, Hutchinson, J. M. Dent, John Spencer, Jonathan Cape, Legend, Lions, Magnet, Mammoth, Mandarin, Mayflower, Mayflower-Dell, Methuen, Michael Joseph, Millennium, NEL, New English Library, Orbit, Orion, Paladin, Pan, Panther, Panther Granada, Peacock, Penguin, Picador, Piccolo, Puffin, Quartet Books, Rupert Hart-Davis, Scion, Sidgwick & Jackson, Sphere, Star, Tandem, Target, Telos, The Science Fiction Foundation, The Women's Press, Titan, Triad, Triad Grafton, Triad Granada, Triad Panther, Unwin Hyman, VGSF, Venture SF, Virago, Virgin, Vista, Voyager, W. H. Allen Star, W. H. Allen, Weidenfeld & Nicolson, William Kimber. You have to be careful with some: e.g. "Pan" alone is probably British, but "Pan Macmillan" is global. BLongley 12:26, 1 March 2009 (UTC)
Thanks! I am still down for the count and Fixer is getting fidgety, but we do what we can... Ahasuerus 01:17, 3 March 2009 (UTC)

I can't really define what I want, but this isn't it. Likely Omnibuses, Collections and Anthologies need to be done in small doses. Author runs of novels can be done in larger numbers. But dealing with a Fixer queue ordered by one categorisation alone (publisher) isn't actually much fun. Just my opinion, and I'm happy to discuss this with all the other people dealing with Fixer entries. (I hope there are some others. At times it feels a bit lonely when I'm looking at submissions from days ago that nobody else has touched. I actually have no interest in "The Art of Star Wars" submitted Wednesday, for instance. I didn't really enjoy the Star Wars "Essentials" either - I've found them neither Essential or interesting.) BLongley 23:44, 27 February 2009 (UTC)

Michael approved and massaged quite a few Fixer submissions last week and I dabbled in it whenever I had a few minutes, so you are not alone. Ahasuerus 03:55, 1 March 2009 (UTC)
Going forward, we can do two things. First, we can make Fixer smarter about using US/UK data and start submitting by author -- see above. Second, we can switch to the library data that Fixer has been slowly accumulating in yet another database. It's not perfect, but it's generally much cleaner than what Amazon has and should be noticeably easier to approve/massage. Unfortunately, the code to process/submit this library data is not ready yet since I have been busy to the point of exhaustion over the last two weeks. I am still trying to recover and hope to be back to normal by early next week. As John W. Campbell Jr. once observed, nobody is getting any younger... Ahasuerus 03:55, 1 March 2009 (UTC)
P.S. I plan to submit the rest of Del Rey later tonight. There are fewer than 200 eligible books left and by the time we are done with them, Amazon UK should be ready. Ahasuerus 03:55, 1 March 2009 (UTC)

Warning about Amazon UK prices for Ebooks and Audio Books

I'm not sure what the API is returning for Recommended Retail Price currently, but what they're showing on the UK site is a bit misleading. For instance, a lot of audio books I entered today are showing a RRP of £9.77. Which is sort of correct, for now, but probably isn't what is shown on the book itself. The reason is that the government dropped the rate of VAT (sales tax) from 17.5% to 15% last December, and are going to keep it that way through December 2009. Amazon seem to have passed on the full rate cut to customers, so a £9.99 audio book is shown as £9.99/1.175*1.15 = £9.77. I don't know when they changed (most IT departments couldn't get such a change made in time for the start) but if the API is returning 15% VAT RRPs then it's probably wise to adjust them back to 17.5% as that'll be what's shown on the publication. Or just leave e-books and audio books for now - good old paper books are still 0% VAT and didn't change. BLongley 20:15, 15 March 2009 (UTC)

Thanks, I'll check the data that Amazon UK sent back to me! Ahasuerus 21:13, 15 March 2009 (UTC)
Thinking about it - Amazon UK are being a bit silly. If they quoted the printed RRP, then their discounts would look bigger! BLongley 21:33, 15 March 2009 (UTC)

Amazon UK done

Amazon UK is done, now I need to teach Fixer how to combine US and UK records and we should be in good shape. The only thing stopping me is that I am tired and somewhat burned out, but hopefully I'll recover sooner rather than later... Ahasuerus 01:57, 31 March 2009 (UTC)

The next week looks promising. With luck, I should be able to teach Fixer how to create composite submissions from both stores' data. Then I can ask him to create submissions for, say, all of Zelazny's books sorted by title. Does this sound reasonable? Anything else that I need to improve while I am at it? Ahasuerus 19:35, 2 April 2009 (UTC)
I'd prefer a Publisher submission rather than an Author one (I find more interesting stuff that way), but at the moment I'll have a look at anything different. I'll let you know the next set of required improvements when I see what needs improving. :-) BLongley 19:52, 2 April 2009 (UTC)
As long as the publisher in question publishes books in one country and one country only, Fixer can create submissions using just that country's data, so we can create a bunch as early as tonight. However, as you wrote above, "One author should probably be submitted in one go, as it's annoying to clear the queue and do a dozen Brooks or Eddings or Jordan merges, and then do the same merges again the next time, and the next, and the next." If Fixer were to submit a bunch of, say, Gollancz records, we would have to do a lot more merges compared to submitting all Zelazny records at the same time. I can do it either way, of course, just let me know what's easier :) Ahasuerus 20:16, 2 April 2009 (UTC)
"By Publisher" offers the chance of finding a totally New Author, and I like those. For instance, I was quite pleased to discover Richard Saxon this week - we'd had no record of him before, and I could wander all over the web trying to find out more of his works, pseudonyms, biography, etc. (I've ordered copies of all his works as well, so those should have a few Primary Verifications soon too.) BLongley 21:04, 2 April 2009 (UTC)
Well, we have tried this approach with "Subject = time travel" and "Publisher = Roc/Tor/etc" and the results were somewhat mixed. I hope that a scan of UK publishers will be more productive since they are under-represented in the database compared to US publishers. Ahasuerus 22:04, 2 April 2009 (UTC)
"By Author" might find a missing title, but that's rare. I have encountered such and increased my Amazon shopping-list accordingly though. I do feel that Author submissions, if they have to be broken up into smaller batches, should at least do all versions of the same title in one go to minimise merges, but how well Fixer can determine possible variations I don't know. When the variation is a suffix it shouldn't be too hard, but so far I've just not found such submissions as of much interest. BLongley 21:04, 2 April 2009 (UTC)
That's a good point. I'll try to submit "similar" title together, although it may require a little manual massaging before submission. We'll see how it goes. Ahasuerus 22:04, 2 April 2009 (UTC)
And we can still try and keep records together - whether it's Author or Publisher first, order by the other second, and title third. I suspect "By Publisher" won't cause too many variants now that Fixer can "Add Publication to this title". BLongley 21:04, 2 April 2009 (UTC)
If people are worried about the number of needed merges created, start small - e.g. I've had a good stab at entering missing "Tandem" entries already, it would be interesting to see how many I missed. BLongley 21:04, 2 April 2009 (UTC)
True, smaller UK publishers are probably a good place to start. Ahasuerus 22:04, 2 April 2009 (UTC)
Or "Corgi" or "Orbit" or "Sphere" - but I suspect those are bigger and I've not kept up to date. I'm sort of researching "Consul Books" and "World Distributors" at the moment, but half are pre-ISBN and the problems of a Manchester company with a London address makes the Library searches awkward. :-/ I don't expect Fixer to handle those - yet! BLongley 21:04, 2 April 2009 (UTC)
Oh, and not tonight, unless someone else wants a stab at them. Give it 36 hours and I'll have something to look forward to at the weekend. BLongley 21:12, 2 April 2009 (UTC)
I'd like to submit something tonight just to see how Amazon UK books go, but it doesn't have to be a massive 400 book run a la Tor and we don't have to approve all submissions right away. On the other hand, first I need to refresh the backup and the full zipped version is now over 1Gb, so it will take some time to download and import into MySQL. (Don't panic, I also have today's backup file saved on another computer as well as on DVD.) Once there is a bunch of Fixer submissions in the queue, I can start working on setting up "composite submissions". Hopefully, this will be the last major change to Fixer's Amazon-specific logic for a while so that I could concentrate on processing library records, which look much cleaner (but also significantly trickier to process). Ahasuerus 22:04, 2 April 2009 (UTC)
Amazon UK is aware of over 1,500 books published by Tandem, Tandem - Library, Universal - Tandem, etc, so I limited the run to the first 10 eligible books published by "Tandem". There are a few obvious problems with the submissions, which will be fixed tomorrow night. Ahasuerus 04:22, 3 April 2009 (UTC)
Well, the price prefix looks funny and the Mod notes should point to the UK site rather than the US one but otherwise they look reasonable. The fact that one of the first examples tells us the US site thinks the publisher is "Award" shows this is going to be of serious help. BLongley 18:26, 3 April 2009 (UTC)
It's definitely useful information, but it has to be approached as carefully as any other source. Apparently, Amazon UK tend to drop leading articles and even insert erroneous articles in the middle of titles, which can be sometimes corrected by cross-checking against OCLC. They also use post-decimalization prices for pre-1970 books, which I move to Notes. OCLC is not perfect either, of course, but between the two of them it's often possible to recreate the data. Ahasuerus 03:10, 5 April 2009 (UTC)
Of the choices, the "Nick Carter" looks a bit suspicious - "Werewolf" may just be a code name and not indicate any actual lycanthropy. BLongley 18:26, 3 April 2009 (UTC)
Yup, it turned out to be a standard issue post-James Bond Nick Carter novel with no SF elements. Ahasuerus 03:10, 5 April 2009 (UTC)
Another one for the "Human submissions only" list maybe, as there's 260 more in that "Killmaster" series and yet we've not had a single one entered yet. BLongley 18:26, 3 April 2009 (UTC)
There were 12 of them on file, I have suspended them for now. According to Wikipedia, some of them have SF-sounding titles like The Solar Menace, but I don't see anything that looks like real SF. Ahasuerus 03:10, 5 April 2009 (UTC)
The two Doctor Who titles are probably not needed - IIRC, the "Doctor Who Discovers ..." series are educational books presented by the fictional character. But there weren't many and they can be rapidly hidden away in some corner of the Whoniverse. BLongley 18:26, 3 April 2009 (UTC)
OK, further research suggests five, and there may be framing fiction. One refers to fictional creatures as well, so may be valid NONFICTION. Apparently they sold badly so we're spared another 24, and it's unlikely they'll ever be reprinted. BLongley 19:41, 3 April 2009 (UTC)
I'll leave these two submissions to your tender mercies then :) Ahasuerus 03:10, 5 April 2009 (UTC)
OK, massaged into shape and hidden away where no normal editor need encounter them. :-) BLongley 12:17, 5 April 2009 (UTC)

Amazon UK corrections

  • Changed Notes to read "Data from Amazon UK as of 2009-03-00" instead of "Data from Amazon.com as of 2008-12-00". I wonder if US-based submissions should be changed from "Amazon.com" to "Amazon.com (US)" to clarify the source?
  • Fixed the way "£" is displayed. Apparently, the Web API expects all non-ASCII characters to be UTF-8 by default, but it can be overridden by changing the "charset" header to "iso-8859-1". I think I am beginning to see where Al's hatred of Unicode comes from.
  • When Amazon's "Edition" field says "n.e.", Fixer no longer adds it to Notes since it's apparently a placeholder of some sort and is not informative.
  • What's a good cut-off date for decimalization? Is it valid for an alleged 1971-07-21 edition to be priced at "£0.25"?
Note change is good, but link is still to .com rather than .co.uk - i.e. still pointing to garbage rather than the source in some cases. I think we all know Amazon.com is actually a US site though, no change needed there. BLongley 21:21, 6 April 2009 (UTC)
I too hate Unicode. But thanks for the "£" effort, the "L" workaround doesn't work on some displays. BLongley 21:21, 6 April 2009 (UTC)
I think they meant it to mean "new edition" but it's rarely been a sign of newness - maybe new coverart at times. Not necessarily any change in page-count or text. Obviously, the price usually goes up. BLongley 21:21, 6 April 2009 (UTC)
Oh yes, July 1971 is fine. 15th Feb 1971 was the official changeover, but books were dual-priced both before and after the change - anything 1969-1973 with two UK prices is OK. I don't think we've got a standard for which to use, and as it's a text field we're probably never going to use it programatically anyway. BLongley 21:21, 6 April 2009 (UTC)
By the way, I can work with the UK Fixer submissions as they stand, if you want to submit a few even while they're not perfect. BLongley 19:48, 7 April 2009 (UTC)
I noticed you'd held a couple of Tandem submissions that might have been because of the Tandem Fantasy or Sci-Fi suffixes on the title - I'm not sure what you did with those? If they're controversial as to where they end up, just choose another publisher to work with. I can live with Tandem Fantasy, Tandem Science Fantasy, Tandem Science Fiction for now, and wait for publisher regularisation / publisher series to get addressed eventually. After all, some of the first submission needed moving to "Target" which is not exactly intuitive. BLongley 19:48, 7 April 2009 (UTC)
Sorry about the delay, I was busy fighting other fires. The URL now points to the UK site for UK-based submissions and there are over 25 entries in the queue for your approval pleasure. Amazon UK's data looks about as dirty as Amazon.com's, though -- one of their 1971 "Tandem" record turned out to be a US audio book recorded by Roger Zelazny -- so all usual Amazon caveats still apply. Once this batch has been approved, we can check if there are still any outstanding issues with Fixer and how we want to structure future submissions, e.g. by author/publisher/etc. In the meantime, I'll start working on teaching Fixer how to create composite US/UK submission.
As far as Tandem derivatives (Fantasy, etc) go, the submissions were on hold for other reasons as I was cleaning up the submission logic. Following your lead, I ended up changing the Publisher field to "Tandem Fantasy" etc. Ahasuerus 20:38, 7 April 2009 (UTC)
Cross-checking a few submissions, I see that "Tandem Library" may be in the business of repackaging US editions, at least in some cases. If so, we could end up with a number of incorrectly attributed pubs. Is there a way to find out more about Tandem Library's devious designs before Fixer creates more submissions? Ahasuerus 20:53, 7 April 2009 (UTC)
Food for thought there - "Tandem Library" seems to be an aberration, it doesn't fit with the UK Tandem publications. "0613" ISBN prefixes don't fit with the rest, there's a different issue there. Let me go find my physical meal and I'll whack them about a bit when I've eaten. BLongley 20:58, 7 April 2009 (UTC)
OK, [1] and [2] don't fill me with any confidence that I can apply any specialist UK knowledge to "Tandem Library" - they're in no way associated with UK editions by Tandem Books/Fantasy/Science Fiction/Science Fantasy as I'd hoped for. I'd expected a few non-SF titles but not a big gap like this. BLongley 22:02, 7 April 2009 (UTC)
I'd reject ALL these if the UK-ness was the reason for submission - on a more general note though, the Spanish 1984 and the "Alien (Animorphs (Sagebrush))" probably should be in on other grounds. The children's books are sort of split - I'm not a fan of adding them just because they're juvenile nearly-SF but when we have something like "Pinocchio" already it's hard to reject other versions. But I think all these can be rejected and resubmitted from the US site data, and "Tandem Library" from the UK site is probably not worth looking at. BLongley 22:02, 7 April 2009 (UTC)
I think "(Hello Reader! (DO NOT USE, please choose level and binding))" or even just "DO NOT USE" might be a warning Fixer might usefully recognise in future. I will, of course, use that title for my first SF Novel just so I get to enter it here myself. ;-) BLongley 22:13, 7 April 2009 (UTC)

US/UK stats

At last count, the breakdown of (non-suspended, non-rejected, non-submitted) books in Fixer's Amazon database is as follows:

  • 21,331 US-only books
  • 42,236 UK-only books
  • 106,521 US and UK both

Ahasuerus 04:39, 23 April 2009 (UTC)

Author with more spec-fic

Can you run Fixer on this author "H. J. Ralles"[3], there appears to be a lot more titles aimed at young adults. Thanks!

Not as many as I expected, but the 8 matching records have been submitted. Ahasuerus 01:31, 27 April 2009 (UTC)
P.S. Her Web site has additional information and the series order. Ahasuerus 01:54, 27 April 2009 (UTC)
Everything seems up to date now. Thanks!Kraang 02:23, 27 April 2009 (UTC)
Looks great, thanks! :) Ahasuerus 03:07, 27 April 2009 (UTC)

C(hristine) (M.) Morgan

Found this author C. M. Morgan[4], we have one book in a series of 6. This author also publishes under her given name of Christine Morgan, there maybe others under that name. Some of the Abebook sellers also have her as Christine M. Morgan just to to make things a little bit more confusing! Thanks!Kraang 01:37, 29 April 2009 (UTC)

Christine Morgan, C. M. Morgan and Christine M. Morgan all submitted. We'll pin her down yet! :) Unfortunately, even though she is apparently published by a US-based publisher (her own, as far as I can tell), many of her books were only listed by Amazon UK, so they have prices like "£4.83", which will need to be corrected. Her Web site and her online bookstore have additional information about the books, which may help whip them into shape. Ahasuerus 02:10, 29 April 2009 (UTC)
I found the right book by her to help whip it into shape[5]! Thanks!Kraang 02:20, 29 April 2009 (UTC)
Hm, her blog mentioned that she was an avid gamer, but I didn't realize what kinds of games she was into! :) Ahasuerus 02:31, 29 April 2009 (UTC)

Moorcock Millennium editions

I should have mentioned this after the last lot - but you're putting them almost always under the wrong title. Is it possible to check for existing Omnibus and Novel records, and if both exist, don't automerge the submission? (I won't get started on the fact that there can be White Wolf and Millennium Omnibuses under the same title too. Oh, come back Unapersson, all is forgiven!) BLongley 18:37, 29 April 2009 (UTC)

Logic changed as follows: If Fixer finds more than one eligible Title (i.e. a book length title with the same Authors) on file, he no longer tries to auto-merge. Why is it that I am not surprised that it was Moorcock who brought this issue to light? :) Ahasuerus 03:20, 30 April 2009 (UTC)
Ta muchly! That should make the next set less fearsome. I guess Moorcock would be the most likely trigger, but there's a lot of trouble with anyone getting even a single "complete works" reprint series. Robert Silverberg and Philip K. Dick come to mind: e.g. LNTMTMBKT2000 caused troubles and 37576 is a nightmare. We could create "Journeyman Projects" for a single title alone at times. And sorting out NESFA entirely should be worth a Ph.D. in ISFDB library science. :-/ BLongley 20:33, 30 April 2009 (UTC)

"Wizards of Waverly Place" young adults 9-12

Found this series "Wizards of Waverly Place"[6] put out by Disney press for young adults 9-12. The books appear to be by different authors. Is this one that should be done manually? I have a sneaking suspicion there's a lot more of this stuff out there. Thanks!Kraang 02:34, 30 April 2009 (UTC)

Unfortunately, Fixer doesn't capture series information since I couldn't find a way to get to it :( A quick OCLC search suggests that "Wizards of Waverly Place" is a series of YA books based on a TV series and that they are done by a group of authors which includes Heather Alexander, Beth Beechwood and Sarah Nathan. I have submitted all titles in the Fixer database that had "Waverly Place" in the title and a few other ones by related authors, but we may need to trawl OCLC for additional information. Do you know how to do series searches in OCLC? Ahasuerus 03:10, 30 April 2009 (UTC)
No, but that won't stop me from figuring it out. :-) Thanks!Kraang 03:31, 30 April 2009 (UTC)

14-digit ISBNs

Fixer is finding some in the UK data, e.g. 97805750773010. Is it possible to catch those? They don't get flagged up as invalid like 10-digit ones do, the only clue is a subtle lack of formatting. BLongley 20:36, 3 May 2009 (UTC)

Hm, let me take a look... Ahasuerus 20:39, 3 May 2009 (UTC)
I can't find 97805750773010 at Amazon UK, so I wonder if it's something that Fixer does when converting ISBNs. Do you happen to remember which publication it was associated with? Ahasuerus 22:53, 3 May 2009 (UTC)
I think I have found it. Fixer was using "10" as the checksum digit for ISBN-13s, silly bot! Ahasuerus 22:58, 3 May 2009 (UTC)
Sorry, I should have made it clearer that 97805750773010 was what was here rather than on Amazon. Still, a search on "978___________" should find the rest. BLongley 23:53, 3 May 2009 (UTC)
I think I got all of them now. Ahasuerus 00:49, 4 May 2009 (UTC)
But that finds "Crystalwizard", and that looks a nice distraction to clean up... "Kelly Christiansen" watch out, I'm investigating! BLongley 23:53, 3 May 2009 (UTC)
Checking the back cover on Amazon.com, the author seems to be credited as "Crystal Wizard". Since the publisher is AuthorHouse, a well known vanity press ("Since 1997, AuthorHouse has helped more than 40000 authors put their voice in print!"), the author can presumably choose his own pseudonym, no matter how unmarketable it would be in the commercial world... Ahasuerus 00:49, 4 May 2009 (UTC)

"All versions of the same title in one go to minimise merges"

We've discussed this before. I'm not sure it's being implemented: "The Larion Senators" looks rather familiar, "The Cold Kiss of Death" too, and I'm pretty sure "Red Gloves" has been submitted one title at a time in three separate Fixer runs today. The last being particularly annoying as it's by a pseudonym and I'm not even sure if it's a "Swords and Sorcery" novel we want or just "Swords" we can maybe ignore. When they're submitted closely together I can spot useful things like the hc, tp and pb versions and add bindings when Fixer doesn't know them - when they're apart it means more work. Are there some more sorting improvements lined up? BLongley 00:21, 4 May 2009 (UTC)

There are indeed sorting improvements in the works, but they are all dependent on the same issue that needs to be addressed first. Fixer has a quarter million records from Amazon.com and another quarter million records from Amazon UK. Even after auto-rejecting or auto-suspending all ISBNs that already exist in the ISFDB database or look like manga/comics/western/accessories/RPG modules etc, we still have about a hundred thousand records for each source. These records overlap to a significant extent -- see the stats posted above -- and Fixer really needs to know how to create composite submissions using both sources for the same ISBN before better sorting can be implementing. Unfortunately, composite submissions are not as simple as they may sound since they require reconciling all kinds of fields and I am still working on it. In the meantime, I am using Gollancz as a guinea pig to find minor problems like the one that you discovered with 13 digit ISBNs and, when I get tired, I write scripts to fix URLs, download biblio Web sites, etc. Hopefully, composite submissions will go live within the next week or two. Definitely in May since my ISFDB time in June is likely to be low. Ahasuerus 00:59, 4 May 2009 (UTC)
Gollancz has many problems - there was a lot of vapourware created when they were taken over and editions were cancelled, and a lot of ISFDB entries still credit Gollancz where the imprint should rightly be "Millennium" for instance. Anything between 1998 and 2002 is a bit suspect. For newer editions it would help to keep the publications together. BLongley 01:48, 4 May 2009 (UTC)
I think there are only about 100 Gollancz books left on the UK side, so we are getting close, although I should probably also run the US side to see if there is anything interesting there. Any suggestions re: the next victim while I am working on composite submissions? Ahasuerus 04:05, 4 May 2009 (UTC)
My ISFDB time may be increased soon, for several sorts of bad reasons I won't go into, but I think we could be a little more productive in the meantime. Feel free to badger me for suggestions. BLongley 01:48, 4 May 2009 (UTC)
One relatively painless thing to do would be to create two subsets of records, one for all "Amazon UK-only" records and one for all "Amazon US-only" records and then sort them by author/title. That should give us many thousands of records to work on while I am polishing composite submissions. Ahasuerus 04:05, 4 May 2009 (UTC)
Gollancz done, working on the approach outlined immediately above. Ahasuerus 00:59, 5 May 2009 (UTC)
After playing with the data for a little while, I don't think I like the "Amazon UK-only"+"Amazon US-only" approach much. It's easy to do, but it gives you less than half of an author's output. On the other hand, sometimes it creates nearly complete biblios for obscure country-specific authors, especially the vanity set. Decisions, decisions... Ahasuerus 02:32, 7 May 2009 (UTC)

(Unindent) Oi! What are all those "The Woman in White" submissions doing in this run? They should have been in the last one with the rest! I haven't even finished the last one yet - "Great Tales of Terror from Europe and America" could need hours of work alone. Do you know how many Anthologies you added last time? BLongley 20:22, 10 May 2009 (UTC)

The sorting code is not live yet. Later tonight, hopefully... Ahasuerus 20:44, 10 May 2009 (UTC)

Amazon UK links

I've recently noticed that the "Amazon UK" link beside a publication doesn't work for pubs with an ISBN-13 rather than an ISBN-10. I'm working around it for the moment by leaving the "Detailed information available here" link from the Mod notes in, e.g. Hater, but if we don't get the software fixed soon then it might be better for Fixer to move such links above the "MODERATOR NOTES:" section to encourage mods to leave such in? BLongley 11:52, 4 May 2009 (UTC)

It looks like Marty will add it to his list of things to fix shortly. (Hopefully.) Ahasuerus 03:36, 7 May 2009 (UTC)

James Dashner

Found another one, James Dashner[7]. Thanks!Kraang 02:15, 5 May 2009 (UTC)

Looks like he has two short series, The Jimmy Fincher Saga and The 13th Reality Series, both submitted now. Ahasuerus 03:28, 5 May 2009 (UTC)


George R.R. (Tuf Voyaging)

Hi, I was idling looking at a title I recognised for "Fixer" in the moderator queue & have put it on hold 'cos there's no space after the first R. in the author. Not sure if it's something the scripts can deal with; just thought I'd bring it to your attention, because I can't see anything above specifically on spacing of initials. ...clarkmci/--j_clark 04:24, 5 May 2009 (UTC)

Unfortunately, this is one of many problems with Amazon's data and I am a little hesitant to start changing their data programmatically. For example, they have records for "H. G. Wells", "H.G. Wells", "H G Wells", "H. G. (Herbert George), 1866-1946 Wells", "H Wells", "H. G, Wells", "H.G Wells", "H., G. Wells", and "HG Wells". Trying to regularize all of these permutations in Fixer's logic would be a huge headache, but I suppose I can try to handle the low hanging fruit, e.g. change "H.G. Wells" to "H. G. Wells". I'll add it to the list, thanks! Ahasuerus 16:32, 6 May 2009 (UTC)
"J.R.R. Tolkien" was another I got a bit tired of respacing. And "H.P. Lovecraft". I was worried about "A. R. R. R. Roberts" too, but it seems Amazon changed a lot of those to "Adam Roberts" and the hassle was even worse changing them back. (He's also published as "A3R Roberts"!). But I'm sure we can have an exception list for the most prolific double-initial-or-more SF authors - just don't let "R.L. Stevenson" through please! BLongley 18:12, 6 May 2009 (UTC)
I let this through and fixed the spacing, just to clear the Gollancz: your spacing concerns are noted, I just want to see what the next is! BLongley 19:49, 6 May 2009 (UTC)
Code fixed so that authors with no spaces between the leading initials now have spaces added at submission creation time. This works for up to 4 leading initials, all at no extra charge! Middle initials are not affected since there are entirely too many permutations there. Even leading initials required a bunch of exceptions to process authors like "W.i.t.c.h" and "B.L.U.N.T". Leave it to the SF community to come up with bizarre names! Ahasuerus 22:22, 9 May 2009 (UTC)
Why did you bother with those exceptions? "W.i.t.c.h" just sends you to "unknown", and loads of stuff you'd never want to see. Only a series search for "W.I.T.C.H." gets you quickly to the relevant titles. And an improperly structured series. And things like "Chapter Book" titles that aren't entered that way. Has anyone here ever wanted to see those? Are any verified? Might they be manga? I can't even find "B.L.U.N.T". Exceptions for things we don't want and can't find are a bit over-the-top, IMO. BLongley 00:17, 10 May 2009 (UTC)
These authors may not be in the ISFDB database, but they exist in Amazon's data, which Fixer faithfully reproduces in his database. And when the time comes for him to create an ISFDB submission based on what he has locally, I don't want Fixer to credit "W. i. t. c.h" rather than "W.i.t.c.h", which would likely confuse the approving moderator even more than he would be normally confused by a submission like that :) Ahasuerus 00:45, 10 May 2009 (UTC)

Approving and improving Fixer's submissions

Something like this pub makes me question how thoroughly moderators are researching when accepting submissions by Fixer. The dates of the title record were changed to 1992-05-18 as well (I just changed it back to 1995-04-00). Going to OCLC would have revealed no records for this ISBN. Abebooks shows only three listings (amazingly small for a McCaffrey title), and a Google search comes back with 15 hits, all from websites feeding off Amazon's bad listing. A simple search on Locus1 would have given info that the first edition was in 1995. I would not have accepted this submission, and wonder how many other bad pubs have slipped by. Would any moderators who work on Fixer's submissions want to explain their methods of acceptance? Or am I wasting my time and should rubber-stamp all of Fixer's submissions? MHHutchins 18:29, 6 May 2009 (UTC)

Oh, please don't rubber-stamp them all! A simple look at the link provided often gives a "Currently unavailable. We don't know when or if this item will be back in stock." message and no "Available from these sellers." links. That should be a big warning that this is vapourware and shouldn't be allowed in. (Or allowed in but immediately changed to 8888-00-00, if we want to indicate an ISBN has been "used up".) Maybe Fixer can spot those though and improve the submissions? BLongley 19:41, 6 May 2009 (UTC)
The other thing that's easy to check is that if the title is going to be auto-merged, click that link and look to see if we already have that title in that format at about that date, for a similar publisher or imprint. A lot of the latest Gollancz submissions were for Gollancz vapourware that appeared under a different imprint. Some never actually appeared in that format, due to the publisher changeovers around the turn of the century. I wouldn't expect Fixer to know those though. Not all Mods either: but for those familiar with a publisher, it should be easy (for instance) to spot the hc, tp and pb version from the same overall publisher: I know I've used that knowledge to assign tp or pb classifications to the unknown bindings. BLongley 19:41, 6 May 2009 (UTC)
I'd be interested to know what other mods are doing. But I'd prefer they 'fess up themselves rather than go look at the Recent Integrations list and comparing with the Recent Rejects list - there should be some rejects for almost every Fixer run there has ever been. Although I appreciate I should probably put more effort into explaining my rejects at times. BLongley 19:41, 6 May 2009 (UTC)
Fixer's submissions clearly need more TLC than regular human-generated submissions, in part because the Web API doesn't do much validation of the incoming data (e.g. it lets you create submissions without Authors) and in part because Fixer's data comes from potentially unreliable sources like Amazon. At one point I was going to create a new section about bot submissions and add it to the Moderator Help page, but I got sidetracked by everything else going on. Any volunteers? :)
Also, I am actively investigating a number of other ways in which Fixer can help, e.g. by filling in page counts and similar data elements from the Library of Congress and possibly even using the newly open "xISBN" gateway at OCLC, but we need to make sure that all moderators understand the ramifications of careless approvals before we can start processing more Fixer submissions. Perhaps we should move or copy this discussions to the Moderator Noticeboard? Ahasuerus 02:28, 7 May 2009 (UTC)

"Gollancz US"

There is a surprising number of eligible "Gollancz" books on the Amazon.com side, probably over 200. I am not sure how many of them are (in)valid, but I have submitted the first 10 to see what happens. Ahasuerus 02:41, 7 May 2009 (UTC)

...and another 10 for full measure. Ahasuerus 03:15, 7 May 2009 (UTC)
Oh, I think I see what's going on! When Fixer creates a submission based on, say, an Amazon UK record, he doesn't mark the associated Amazon.com record as "submitted" and vice versa. Then, if the submission is rejected, Fixer has no way of telling that he shouldn't be creating a submission for the same ISBN from the other store. Oops! Ahasuerus 03:40, 7 May 2009 (UTC)
All Fixer submissions have been cleaned up; most were vaporware. Fixer's database has been updated so that any ISBNs submitted from the UK side also result in the matching US records marked as submitted. Fixer's submission logic has been changed to avoid these problems in the future. Ahasuerus 04:22, 8 May 2009 (UTC)

£0.17 Price from UK = 3/6

Bill noticed this, and then I saw another submission with this problem. Pre-decimal prices are getting mangled. One price point is 3/6 which is being converted to £0.175 by amazon, and then reported as £0.17 (With the trailing halfpenny either rounded by fixer, or by amazon). Could we let fixer know about that and convert all £0.17 and £0.175 prices to 3/6? Just a thought. Thanks Kevin 18:50, 10 May 2009 (UTC)

Good point, let me see what Fixer has on file... Ahasuerus 20:28, 10 May 2009 (UTC)
There are 15 occurrences of "£0.17" (3/6), one "£0.13" (2/6) and 4 "£0.10" (2/-). Not too bad, but I will teach Fixer how to convert them. There are another 39 "£0.01" prices, which will be discarded. Thanks! Ahasuerus 20:43, 10 May 2009 (UTC)
Interesting. Why would 17.5 get rounded down and 12.5 get rounded up? I'm interested in those 10p and 1p titles though. 1p might be a valid promotional price. I was an avid (if impoverished) paperback-buyer at the time of the change-over so recall that the common 2/6 or 3/6 prices disappeared and most pbs became 25p - I don't recall ever buying one at 20p but suspect they might have existed, which is why I'd put the cut-off below that. BLongley 20:53, 10 May 2009 (UTC)
OK, all 1p titles have been submitted. Most of them were already in the ISFDB and a few were suspended due to comics browse nodes, so only 14 submissions were created. 3 were clearly non-SF and I rejected them, the rest are patiently waiting in the queue :) Ahasuerus 21:08, 10 May 2009 (UTC)
I can't see a single one that merits a 1p price. But I can't see many that would even qualify for inclusion. Why 9688808989 for instance? Spanish sports book with a UK price? Why did Fixer choose that? And 1416509585 - is there a "CSI: NEW New York, Mars" spinoff or something? (Nothing against CSI, but as far as I know they've only autopsied humans so far, and no extraterrestrials have been found yet.) BLongley 22:13, 10 May 2009 (UTC)
Well, Teller's War: The Top-Secret Story Behind the Star Wars Deception was grabbed by Fixer because Amazon had helpfully labeled it "Star Wars", Ghost Ship: A Cape Cod Story was actually a legitimate ghost story for children, The Gumdrop Ghost and The Crime Lab Case were Nancy Drew stories with the usual "faux supernatural" elements, etc. On the other hand, I guess "£0.01" books are likely to have more than their fair share of bad records... Ahasuerus 22:23, 10 May 2009 (UTC)
OK, but why "Fisica" and "Csi New York 02 Blood on the S (CSI: NY)"? Nothing in the browse nodes listed suggest any reason for selection. BLongley 22:49, 10 May 2009 (UTC)
Oops! That was a (relatively) recently introduced bug with adding/removing the auto-suspend flag when there are no SF Browse nodes and no SF Subjects. Thanks for catching it! :) Ahasuerus 23:02, 10 May 2009 (UTC)
Actually I documented The CSI book as also having an 'in' author, but with only 5 short works (over 50 novels in Mystery, but just playing at the edges of SpecFic) Kevin 23:10, 10 May 2009 (UTC)
I think there may be two separate questions here. The first one is why Amazon.com/UK sent this record to Fixer when Fixer asked for a certain Brose Node or Subject. Amazon's internal algorithms are complex and not terribly reliable (see their Associate forums for numerous complaints and bug reports), so anything is possible. The second question is why Fixer created a submission based on this record even though he found no SF Browse Nodes and no SF subjects. The answer is "it was a bug" :) Ahasuerus 23:19, 10 May 2009 (UTC)
You could probably do "£0.12" to "2/6" as well, but I wouldn't try higher than that. E.g. "£0.20" might be a valid post-decimal price, it wouldn't always mean "4/-". BLongley 19:55, 10 May 2009 (UTC)
It so happens that I verified a 1970 UK pb the other night and the stated price was "5/- (25p)", so "£0.20" may be a valid price. Ahasuerus 20:28, 10 May 2009 (UTC)
I'm not seeing many of those though, and if the programming is hard then we could just add another note on how to moderate Fixer's submissions to the notes that we have on that topic already. (We do have some don't we? I just can't seem to find them...) BLongley 19:55, 10 May 2009 (UTC)
I have been meaning to write something up, but I haven't had much time lately, what with the Fixer changes and other things going on. Are you volunteering? :) Ahasuerus 20:28, 10 May 2009 (UTC)
Nope, I hate writing long articles. I'm supposed to be updating my CV right now and am using ISFDB work pressure as self-justification to avoid such. Somebody should be able to cobble something together from my various moans though. ;-) (Not my CV - the Fixer article!) BLongley 21:05, 10 May 2009 (UTC)
Actually, thinking more, it's probably best to resurrect my price crib-sheets from wherever in the Wiki archives they are. We need a more general sanity check on prices: e.g. Fixer submitted a 1973 paperback with a £3.50 price on it. I know that's wrong, and in the end found it was £0.60. (And I found the contents too.) But most non-UK mods, or younger UK mods (if we have any), wouldn't spot that. BLongley 20:03, 10 May 2009 (UTC)
I thought about dropping prices for older books, but then I found that Amazon US/UK's (especially UK) "list price" field tends to be surprisingly accurate even for 1970s books. I suspect that they purchased and imported one or more price lists at some point. Still, mistakes do happen. Ahasuerus 20:28, 10 May 2009 (UTC)
Yes, most are believable. I just think the Mod's alarm bells won't always go off when Amazon really messes up though. What's the bibliographic equivalent of Spiderman's "spider-sense"? BLongley 21:05, 10 May 2009 (UTC)
Bookmold nose? Pulp Paper Dandruff Detector? Pseudonym Psniffing? Awesome Alliteration Alertness? Title Case Interpretation Sense? Kevin 21:21, 10 May 2009 (UTC)

Sorting logic is live

A limited subset of Fixer's sorting logic is now available. He now knows how to find all ISBNs that exist in Amazon.com or Amazon UK, but not in both databases. This allows him to avoid various nasty problems associated with composite submissions. He then sorts the books by author and can create submissions for complete authors, internally sorted by title. It's not great, but it's a start.

Our first guinea pig was Gene DeWeese. There were 10 matching records in Fixer's database, which resulted in 10 submissions. After approving and massaging a few, I think that our greatest challenge will be the fact that Fixer's data is just a small subset of Amazon's data. There will be times when Fixer creates a submission based on one store's record even though he really ought to be using the other store's record -- except that the other store never sent the record to him to begin with or it didn't have any eligible Browse Node/subject and was auto-suspended. It's not fatal, but it means that we still have to be careful and make sure that the price data comes from the right store. Thankfully, "£4.31" and such are highly suggestive :) Ahasuerus 23:29, 10 May 2009 (UTC)

Ebooks from amazon, that you can't buy at amazon

I noticed two ebooks in the queue and held them (I'll massage and submit them shortly). These items were unavailable at amazon and had no records at worldcat. Only a google search pointed out that they were ISBNs assigned for MobiPocket encrypted versions. Further research indicates that eReader and MSReader encrypted version are also available, though the MSReader version does not have an ISBN assigned in the database at fictionwise.com at this time. I just wanted to point out this new (to me) wrinkle in scrapping unavailable titles in fixer's submissions. I presume Amazon is listing the ISBNs because they own mobipocket now, but since they are not 'Kindle' versions you can't buy them from Amazon. Kevin 01:36, 11 May 2009 (UTC)

I think most of our e-book ISBNs come from Fictionwise.com and similar sources. Amazon may have offered them at one point and then stopped when they released Kindle, but I am not sure. OCLC's coverage of electronic resources is still spotty, but they are making progress as more and more libraries start carrying them. Ahasuerus 01:55, 11 May 2009 (UTC)
Comments welcome on The Final Nexus with link to (UNBOUND) and unavailable title on Amazon.uk preserved. These ebooks are only available under this ISBN legally in the US and Canada (Sorry Bill). EDIT - Though I now notice that the kindle cover has different masthead (text color, etc) and may deserve a separate pub record. Kevin 04:28, 11 May 2009 (UTC)
I have added the kindle specific edition as a separate pub here. Also, note the Powells link on the first pub does work. Kevin 04:37, 11 May 2009 (UTC)

Rie Sheridan

We have one in the can already[8] and it appears this ravishing beauty[9] has more SF and Fantasy. Thanks!Kraang 04:11, 16 May 2009 (UTC)

Good catch, Fixer was aware of 4 other pubs and submitted them a few minutes ago. Thanks! Ahasuerus 05:17, 16 May 2009 (UTC)

Johnathan Rand

Rand[10] has a lot of books(50+) and they appear to be of a spec-fic nature targeted to the 9 to 12 age reader. I think he meets our criteria but the books might only imply spec-fic elements and not actually contain them, I'm not sure, what do you think? Wikipedia[11], authors page[12]. Thanks!Kraang 01:30, 17 May 2009 (UTC)

It looks like some of them are bona fide SF while others are mysteries with fake SF elements a la Nancy Drew, Scooby Doo, etc. They all fall in 4 series -- Adventure Club (3 books, at least one is a collection), American Chillers, Freddie Fernortner, Fearless First Grader and Michigan Chillers -- so I figure we want to capture them all and mark the apparently non-genre ones as such. Fixer has submitted what he had, but there are probably more Strand titles out there in the "wild". Ahasuerus 03:13, 17 May 2009 (UTC)
Look what else I found about him "Christopher Wright"[13]"Wright uses two pseudonyms when writing - Johnathan Rand for the children's books, and Christopher Knight for the books aimed at young adults." I kind of thought that some had fake SF elements myself. Some may be harder than others to identify.Kraang 03:29, 17 May 2009 (UTC)
Hm, I see a number of Web sites about Christopher Wright writing as "Johnathan Rand", but I can't seem to find anything about the "Christopher Knight" YA books except for a Wikipedia page about another author called "Christopher Knight", who writes about Freemasons, the Illuminati, etc. Wonder if the "Christopher Knight" books are SF? Ahasuerus 04:27, 17 May 2009 (UTC)
Fixer has found and submitted more Johnathan Rand books from Amazon UK. They seem to be US editions that are not labeled "SF" by Amazon.com, so the price needs to be blanked out. Ahasuerus 05:27, 17 May 2009 (UTC)

Daniel Cohen

It appears that Daniel Cohen has written a good many books, many aimed at young (pre-teen) readers, about ghosts, ghost stories, ESP, magic, and the like. All I have encountered so far are non-fiction, at least in intent, and IMO not suitable for inclusion in the ISFDB. I have rejected a bunch of fixer submissions for this author. Whether these indicate a good way to detect nonfiction I can't say, but all were listed as "Novel" and if they had been includable at all, would have been anthologies or collections. This is just FYI, not a complaint. -DES Talk 16:02, 20 May 2009 (UTC)

Update, one valid work found to date, The Monsters of Star Trek. -DES Talk 18:47, 20 May 2009 (UTC)
They are clearly mostly non-fiction, although we also have two of his anthologies on file. Fixer's data currently comes from Amazon.com and Amazon UK, which provide virtually no information that would help determine whether a particular book is a Novel, Collection, Anthology or Non-fiction. Eventually (hopefully soon) Fixer will start submitting data from libraries, which typically have separate codes for single author collections, non-fiction, novels, etc, so things will improve at that point. For now the only thing that Fixer can do is check whether the ISFDB already has a Title with the same author(s) and title and, if it does, add a new Publication record for that Title, re-using the Title's "title type". If this information is not available, Fixer defaults to Novel -- not a great solution, but the best one I have been able to come up with so far.
The next question is whether these Non-fiction titles are germane in the ISFDB context. The reason that Fixer submitted these books was that they were labeled "science fiction", "fantasy" or something similar by Amazon -- Fixer has a long list of eligible subjects and "Browse Nodes" as well as an exclusion list for comic books, RPGs and such. Unfortunately, Amazon's data is not particularly reliable, e.g. they often put westerns under "SF", so it's entirely possible that they marked books about ESP and magic as "fantasy". Worse, some titles may be borderline eligible, e.g. The Magic Art of Foreseeing the Future (which you deleted earlier today) "[d]escribes the various methods used by man throughout history to predict the future" (according to OCLC), so it may conceivably contain a chapter or two on science fiction, but there is no way of telling. Oh well, it's not a big deal in the grand scheme of things, just one of those irritating side issues :) Ahasuerus 19:29, 20 May 2009 (UTC)

Foreign language books

Bill has requested that Fixer stop submitting German language books. I am sure Fixer would be happy to comply and auto-suspend them until such time when we have volunteers who would be willing to handle them. However, let's take a look at the larger picture first since there are 63 different language codes used in Fixer's database.

Amazon can have the following (optional) language types associated with every record:

  • Original Language
  • Published
  • Unknown [?]
  • Translation (only a few dozen records use this code)
  • Parallel (about 20 bilingual English/Spanish books)
  • Subtitled (20ish subtitled multi-media books)

Also, the following "Edition Statements" may be relevant:

  • SPANISH LANGUAGE
  • FRENCH LANGUAGE
  • FRENCH LANGUAGE EDITION
  • ITALIAN LANGUAGE EDITION

At this point Fixer adds the "Edition Statement" information to Notes, but ignores the Language fields. The question then is how do we want to alter Fixer's behavior?

  • Should Fixer add the "Original Language" to Notes as long as the language is not English?
  • Should Fixer automatically suspend all records whose "Edition Statement" field includes the words "Spanish", "French", "Italian" (or "German", although there are none in Fixer's database at this time)?
  • Should Fixer auto-suspend all records whose "Published Language" codes (if defined) do not include one of the following:
    • ENGLISH
    • MULTILINGUAL
    • UNKNOWN
    • AMERICAN SIGN LANGUAGE
    • MIDDLE ENGLISH
    • OLD ENGLISH

Anything else?

Also, please keep in mind that Amazon's data is, as we all know, quite dirty, and language codes are no exception. For example, Amazon UK lists El Comando Tribulacion as an "Original Language - Spanish; Published Language - English" book, i.e. the exact opposite of the actual situation. If we ask Fixer to stop submitting what Amazon lists as foreign language books, we will likely lose some legitimate submissions, but it may be the price that we have to pay. Ahasuerus 02:27, 30 May 2009 (UTC)

I don't object to Fixer submitting foreign language editions, but I don't have the language skills to work on them. I think the point about possibly losing data to false positives in the language code, given the state of Amazon data, is well taken. -DES Talk 19:17, 31 May 2009 (UTC)
When Fixer "suspends" a record in his database, he notes why the record was suspended, so if we ask him to suspend non-English records, it would be easy to resubmit them later if and when we have more editors proficient in other languages. False positives will be irritating, though -- I just scanned all "Polish" records and there is at least one false positive. Ahasuerus 19:43, 31 May 2009 (UTC)

Fixer - Off the Leash?

I see that Fixer got loose and has been digging in Mrs. Grundy's flower bed again (Chuckle) Kevin 03:48, 21 June 2009 (UTC)

Fixer is very sorry -- I have posted his apology on the Moderator Noticeboard. Ahasuerus 04:25, 21 June 2009 (UTC)
Bah - Mrs. Grundy needed new flowers anyway. Kevin 04:32, 21 June 2009 (UTC)

Unneeded Comma in Name

I saw two of these (so far) tonight. FIRSTNAME, MI LASTNAME and FIRSTNAME, MIDDLENAME LASTNAME - Not sure if Fixer is checking things like that. Just FYI. Kevin 04:42, 21 June 2009 (UTC)

I asked Fixer about it at one point. He said that there were so many permutations, e.g. "(Wells), H.G", that he'd rather leave the task of sorting them out to humans. Probably a good thing too... Ahasuerus 14:18, 21 June 2009 (UTC)
Could he flag them better? For all NAME like '%,%' and not like '%, Ph.D.%' and not like '%, M.D.%' and not like '%, II%' NAME = NAME + " Check Name" - Since any name with a comma and without one of the top three reasons for having a comma is VERY unlikely to match and merge with an existing title since the author is very unlikely to match, putting a flag in the author name field as a clear text item, isn't going to prevent matches and merges in well formed names. Kevin 14:59, 21 June 2009 (UTC)
If you are going to do this, better include '%, Jr.%" as an exclusion. -DES Talk 15:22, 21 June 2009 (UTC)
Sure, it can be done quite easily. Do we want it in the Author Name field so that if the moderator misses it, we could find it later by searching on "Check Name"? Ahasuerus 15:37, 21 June 2009 (UTC)
That was my thinking. I fixed a couple last night in the Title block, and forgot to fix it in the contents title block so I had to go back and hunt for them. An easy 'fix me' search would have been nice. Kevin 15:46, 21 June 2009 (UTC)
OK, changes made and a couple of examples submitted. The pubs need to be changed from pounds to dollars since Mundania is a US publisher, but the Amazon.com records for these ISBNs were marked "comics" for some reason, so Fixer used Amazon.UK's data instead. Ahasuerus 16:07, 21 June 2009 (UTC)
Oh, and the "exclusion list is currently set to "Ph.D.", "M.D.", "Jr.", "Sr.", "II", "III", "IV", and "Esq." Ahasuerus 16:08, 21 June 2009 (UTC)
", II" will also exclude ", III" (saving you a step of useless code?), and I recommend against ", IV" as Fourths are rare enough, and ", Ivan", Iverson, etc will be wrongly excluded. I would rather have Fourths be flagged. The example looks fine to me. Kevin 17:08, 21 June 2009 (UTC)
Oh no, Fixer only accepts names that end with ", II", ", III", ", IV", etc, the rest are tagged. (That's why ", II" and ", III" have to be done separately.) As far as ", IV" goes, it's surprisingly common - just search on "," from the main page and you'll see. Ahasuerus 17:19, 21 June 2009 (UTC)

UK prices on US editions

Consider Black Wind, which resulted from my massaging of a fixer submisison. As Submitted this had a price of £5.37. I understand that the data was derived from Amazon UK, but that site shows "US List Price: $7.99". Is there any way for Fixer to pick up this field? -DES Talk 14:42, 2 July 2009 (UTC)

Unfortunately, Fixer was unaware of this field in the Amazon UK database, so he didn't grab it. Let me see if I can find it in their record description and update Fixer's ~250,000 records. At 1 secons per record, Fixer should be done in about 3 days... Ahasuerus 01:25, 3 July 2009 (UTC)
On further investigation, it looks like we are out of luck -- this field doesn't seem to be returned by Amazon's publicly available API :-( Ahasuerus 02:35, 3 July 2009 (UTC)
It figures. Well, that's why we have human mods. -DES Talk 04:13, 3 July 2009 (UTC)

Sprechen Sie Deutsch?

As I don't, and you're giving me headaches again! :-( BLongley 19:55, 6 July 2009 (UTC)

"Sie"? No need to be so formal, Bill! :) [And by the way, the way the use of du/sie has evolved over the last century+ in countries like Germany, Sweden and Russia is a fascinating subject, but that's a whole different story.] And yes, that was a particularly bad run and I will be adding a level of human supervision to Fixer's submissions tonight -- see my last post on the Moderator Noticeboard. Ahasuerus 21:07, 6 July 2009 (UTC)
I have no idea what the correct usage of "Sie" is, hence my admission (yet again) that I cannot cope with German, and I haven't attempted these, and my colleagues that can aren't actually interested in ISFDB. I've no idea why - they mention words like "wife", "kids", "sport" and other such words as if they're time-suckers that prevent them from being here. I think those might be English words, but I haven't got any of those problems. BLongley 22:09, 6 July 2009 (UTC)
I actually studied Latin rather than German at school, after giving up French, but only to fail at that with a slightly higher mark, so don't ask me to deal with "Harrius Potter et Philosophi Lapis" either. And if I return to Wales, it's for the Hay-on-Wye bookshops, not to check on whether we got "Harri Potter a Maen yr Athronydd" right. BLongley
Unfortunately, even if you knew German, it still wouldn't help in quite a few cases since German/French/Italian/US/UK/etc publishers can slap a seemingly unrelated title on a translated book. It's that inability to match titles one-to-one that was my biggest problem when I revamped our Russian titles a few months ago. I know the language reasonably well, but when the publisher changes the title from Predatory Things of Our Times to The Final Circle of Paradise, it's impossible to link the two works unless you have access to a pre-existing bibliography or are in a position to compare the two texts side by side. Ahasuerus 02:06, 7 July 2009 (UTC)
Test. Ahasuerus 22:20, 19 July 2009 (UTC)
Test what? BLongley 22:46, 19 July 2009 (UTC)
Whether "My Messages (new)" appears correctly when your Talk page is updated :-) Ahasuerus 22:52, 19 July 2009 (UTC)
Ok, you've tested yours, how about testing mine? (I should probably go check "Data Thief" at some point too.) BLongley 23:04, 19 July 2009 (UTC)

I hope you're not loose on your own...

As an Abridged Audiobook Box set that is Illustrated and Large Print is beyond my abilities to moderate. Is the box big with notable pictures, or is the reading SHOUTED and of such quality that people can visualise certain scenes, even when some words are missing? BLongley 00:29, 26 November 2009 (UTC)

Ah yes, that's the one that, upon closer examination of the Amazon record, turned out to also claim to be written in 6 languages! I wonder what they added to the penguins' feed that day...
P.S. To make Keri Arthur's bibliography even more interesting, Amazon.com lists a number of UK reprints, but claims that they were published by Dell. Ugh. Ahasuerus 02:38, 26 November 2009 (UTC)

Not chasing these myself....

But Karen Whiddon seems to have another 8 Werewolf books to add ("The Pack" series). Something to consider later when we go back to filling in Paranormal Romance authors. BLongley 20:30, 15 December 2009 (UTC)

Lori Devoti has an "Unbound" series that appears to be Paranormal Romance. BLongley 20:38, 15 December 2009 (UTC)

Vivi Anna should have seven Vampire books, eight paranormals and four SF/Future. Seems to be especially "adult". BLongley 20:46, 15 December 2009 (UTC)

Bonnie Vanak has some werewolf stuff. Not sure whether the Egyptian stuff has any spec-fic element. BLongley

Larissa Ione has a few more Demonica books to add. BLongley 22:41, 15 December 2009 (UTC)

Ok, done with the Vampire-Shaggers for now. Please don't rush to submit the missing titles, just keep them in mind. (As I won't.) BLongley 23:20, 15 December 2009 (UTC)
Thanks! (also Kathleen Nance) Ahasuerus 00:46, 16 December 2009 (UTC)

Arcana Studio

You may want to put "Arcana Studio" on the avoid list. After some investigation I believe it's only comics/graphic novels.Kraang 05:03, 26 December 2009 (UTC)

Good point, thanks! Of the 41 "Arcana" books in Fixer's internal database, 14 were manually rejected before a submission was created, 24 were auto-rejected by Fixer, 2 resulted in (erroneous) submissions and the remaining few will be rejected shortly. Many comic/manga publishers like Dark Horse and TokyoPop dabble in written SF these days, but Arcana apparently remains comics-only. Ahasuerus 01:45, 27 December 2009 (UTC)

Repo Men / The Repossession Mambo - Eric Garcia

Hi, Just to let you know I've put this on hold as I can't resolve what the title should be for this item allegedly to be published Feb 2010. Even HarperCollins (US) doesn't seem sure (though Fixer seems to have found the item through Amazon UK as price is £). Amazon US & UK have Repo Men as the text title, but the cover image provided has The Repossession Mambo. Google search on the ISBN gives a HarperCollins.com hit saying it's The Repossession Mambo; but when you go to the Harper web page, it has Repo Men & no image available. (I can't find it on the HarperCollins UK site.)

The Repossession Mambo by Garcia was (1st?) published 2009 it seems. The question is, is the item Fixer found another edition? or the next in the series?

So ... Reject? Leave on hold 'til February? --clarkmci / j_clark 11:52, 26 December 2009 (UTC)

I have confirmed the 2009-02-00 edition of The Repossession Mambo at OCLC, which lists a number of libraries which own copies. OCLC also links to various online reviews, so the 2009 version is clearly legit.
The 2010-02-23 version is iffier. On the one hand, it comes from Amazon UK and it's entirely possible that HarperCollins UK waited a year before releasing a UK-specific version, perhaps to time it closer to the upcoming movie. They may have also come up with a different title, but used a scan of the US version when submitting pre-publication data to Amazon UK. On the other hand, when I find incomplete/spotty data with wrong cover scans etc from a major publisher, it's at least a 75% chance that the book has been canceled/postponed, but the publisher never bothered to inform Amazon, so the ghost record can linger for years.
If you want to keep it on hold until March and then check Amazon UK and OCLC to see if the ISBN shows up, it will probably be the best way to handle the situation. Thanks! Ahasuerus 02:10, 27 December 2009 (UTC)
As of end May '10, there was sufficient data to create a proper record for this new title. New cover image available. Entered the [pub] for the US only as there seems to be just one edition, did a variant on the title [it is a movie tie-in] so the hold can be removed and the original fixer submission scrapped. ~Bill, --Bluesman 02:53, 4 June 2010 (UTC)
Thank you for dealing with this. ... --clarkmci / j_clark 03:48, 21 June 2010 (UTC)

Houghton Mifflin Harcourt <--> Clarion

Anther US "publisher" possibility - Clarion ...

An item I've just looked at was derived from Amazon UK (with a likely UK-published price!) & listed as having publisher Houghton Mifflin Harcourt. However, Amazon US has that ISBN as published by Clarion. Upon investigation ("Look Inside"), it seems the book is US published by the Clarion imprint of Houghton Mifflin Harcourt. The book is a picture book - [here]. --clarkmci / j_clark 13:49, 26 December 2009 (UTC)

Thanks, I have added it to the list of US publishers/imprints. It's an iterative process, so all feedback is welcome! Ahasuerus 02:45, 27 December 2009 (UTC)

Aside: Is it possible (i.e. a simple change) to get Fixer to create an "Interiorart" record too, when Amazon has an illustrator listed? --clarkmci / j_clark 13:49, 26 December 2009 (UTC)

Amazon doesn't include illustrators in the "response groups" that it sends back Fixer. I may be able to find a "response group" which contains this data, but it will take some poking around and likely re-programming since not all response groups are supported for the purposes of the query type that Fixer uses. I'll add it to Fixer's to-do list, thanks! Ahasuerus 02:45, 27 December 2009 (UTC)