HomeGroupsTalkZeitgeist
This site uses cookies to deliver our services, improve performance, for analytics, and (if not signed in) for advertising. By using LibraryThing you acknowledge that you have read and understand our Terms of Service and Privacy Policy. Your use of the site and services is subject to these policies and terms.
  • LibraryThing
  • Book discussions
  • Your LibraryThing
  • Join to start using.

Announcing OverCat

New features

Join LibraryThing to post.

This topic is currently marked as "dormant"—the last message is more than 90 days old. You can revive it by posting a reply.

1timspalding
Edited: Jun 9, 2010, 11:47am Top

2MerryMary
Jun 9, 2010, 11:38am Top

*giggle*

3DaynaRT
Jun 9, 2010, 12:03pm Top

The name goes well with the name of my local library's online catalog, LapCat.

4elenchus
Edited: Jun 9, 2010, 12:16pm Top

From the blog post: "We could make part of the data free, and part closed. But since the free data comes from OpenLibrary it would be duplicative of their efforts. We may explore this avenue in the future, as our primary complaint against OpenLibrary is the lack of exportable library-data formats."

Any possibility of making OverCat available to libraries, or others outside LT, but only including the free OpenLibrary data plus data from individual libraries which opt in? Unless I misunderstand, it appears now the only option is if either OCLC consents to making data open, or that all libraries consent, either of which seems prohibitively restrictive.

I freely admit to dreaming up scenarios without even an inkling of the attendant logistical problems. One advantage of this dream scenario is that it sets up an incentive for individual libraries to consent, and add cumulative pressure on OCLC. Not to mention begin illustrating the value of the open data, a demonstration project.

In any case, I love OverCat as it exists now and look forward to its continued expansion and improvement. Thanks, LT.

5prosfilaes
Jun 9, 2010, 12:24pm Top

Right now, there's no way to make that the default search source. A bug, I presume?

6EveleenM
Jun 9, 2010, 12:29pm Top

It looks great!

But I think it has the same problem as other library sources, that genre books are a bit under-represented. I tried it out on a batch of six books I got today, and it only found records for two. So I entered the other four (one mystery, one fantasy, and two ghost stories) from amazon.co.uk, as usual.

7Heather19
Jun 9, 2010, 12:30pm Top

Yaaay!
I looove the kitty-mascot, too!

8AnnaClaire
Jun 9, 2010, 12:47pm Top

>3 DaynaRT:
What, not NapCat?

9timspalding
Jun 9, 2010, 12:51pm Top

Any possibility of making OverCat available to libraries, or others outside LT, but only including the free OpenLibrary data plus data from individual libraries which opt in? Unless I misunderstand, it appears now the only option is if either OCLC consents to making data open, or that all libraries consent, either of which seems prohibitively restrictive.

Yes, in theory. What stops us most of all is the duplication factor--OpenLibrary already provides access to these records. What OpenLibrary doesn't offer are library-friendly searches, like an open Z39.50 target. But most of this is currently offered by LibLime's ‡Biblios.net. So I'm not sure what we'd add, except more emnity from OCLC.

10caseydurfee
Jun 9, 2010, 12:59pm Top

9 > What we'd be adding would be the search engine and the fact you don't have to parse the MARC records on the other side. Amazon Web Services but with library data. I think there's a niche where Overcat can be very useful for libraries other than as a source of MARC records.

11timspalding
Jun 9, 2010, 1:01pm Top

>10 caseydurfee:

Isn't Biblios searchable? I agree about the XML piece. That would be primarily useful to non libraries, though.

Incidentally, we may allow other cataloging sites to use OverCat. There's an Italian site we may partner with, sharing data either way. But, obviously, personal cataloging is still personal cataloging.

12caseydurfee
Jun 9, 2010, 1:17pm Top

Yes, Biblios is searchable.

I think MARC is a significant hurdle to doing interesting things in libraries as well.

Fundamentally, if there is some way Overcat could prevent anybody else from having to learn all the things I know about MARC (both how it's supposed to work, and how it actually does work) in order to do cool things with library data, that would be nifty. Nifty's not a coherent strategy, though.

13vaneska
Jun 9, 2010, 1:18pm Top

What came over Tim, allowing a cat onto LT?

v

14FicusFan
Jun 9, 2010, 1:33pm Top

Does this do anything to fix the problem that when you press on 'Add this book' or 'Add this book to your Wishlist', it doesn't, but makes you enter it from an outside source ?

If not, then don't really see the point of another outside source.

15elenchus
Jun 9, 2010, 1:39pm Top

I speculate that is part of June: Official Cataloguing Month (tm), but as of yet it does not seem to change the old process of adding from an outside source.

Tim has dropped enough hints about various catalogue-related changes, though, that I open LT each day in the hopes of reading about another major announcement. OverCat is big, but somehow I'm anticipating more. Bigger. Faster.

16jjwilson61
Jun 9, 2010, 1:47pm Top

14> Well, one point is that it's a conglomeration of a lot of library sources so you just one search is like a bunch of searches using the individual sources, and two is that it's faster ta-boot.

17brightcopy
Jun 9, 2010, 2:16pm Top

14> This isn't a "another outside" source. It's effectively like a re-vamp of the entire Add books process. If you never add books any other way than finding an existing work and pressing the Add to your library/Add to wishlist, then yeah, it doesn't really have much of a point for you. Turns out a lot of us add books through the Add books page, completely apart from visiting a pre-existing work page. ;)

18timspalding
Jun 9, 2010, 2:16pm Top

Right. It's significantly more than any one source, and faster, and it combines finding books with choosing which record.

It doesn't solve other problems. It wasn't intended to. But it puts us in a better position to solve other problems.

19timspalding
Jun 9, 2010, 2:18pm Top

>17 brightcopy:

The key missing piece is to mix Amazon in with OverCat. That's probably the next change. As it is, it's the best source for some people, but if you have a lot of paperbacks, it's not necessarily so.

20brightcopy
Jun 9, 2010, 2:30pm Top

Understood. I've actually been rather sad that I haven't had a lot of books to run through it.

Must buy more books.

21Crypto-Willobie
Jun 9, 2010, 2:39pm Top

I been lovin' Overcat since yesterday...
Would it be possible to include Overcat in the "No results found...Try this on" list that is offered after a failed search. As it is you have to go back and click it on the original list. Maybe this is already in the works...

22gwernin
Jun 9, 2010, 2:43pm Top

It's great! Thanks, Tim!

(may have to catalog more this weekend...)

23timspalding
Jun 9, 2010, 2:53pm Top

Would it be possible to include Overcat in the "No results found...Try this on" list that is offered after a failed search. As it is you have to go back and click it on the original list. Maybe this is already in the works...

Yes. I thought it was. Casey? (Or I can do it.)

24Crypto-Willobie
Jun 9, 2010, 2:57pm Top

Great! thanks...

25nathanielcampbell
Jun 9, 2010, 4:29pm Top

Can I make a plea for NOT mixing the Amazon data in? I find that Amazon data are frequently wrong / corrupted (lists the wrong author, gets the publication data wrong, etc.) -- this especially true for academic / foreign language books. When I first joined LibraryThing, I did the first 350 or so books of my library with amazon, then realized how much it screwed up and then spent hours and hours going in and fixing everything. Now, I only use LoC or university catalogues, because they have far fewer bugs.

26TimSharrock
Jun 9, 2010, 4:33pm Top

I would like to use amazon.co.uk as a fallback - try overcat first, have the option of "try amazon.co.uk" if nothing suitable came up from overcat, but leave overcat as the first one to try for the next book

27timspalding
Jun 9, 2010, 4:35pm Top

Yes, it'll be optional—both as to inclusion and to which Amazon. We won't be mixing in stored Amazon results, but live queries. The Amazon queries are at least as fast as the OverCat, so we should be able to "spawn" searches on both and return them at the slower—OverCat—speed.

28caseydurfee
Jun 9, 2010, 4:41pm Top

Yes, I thought it was, too. It'll be there now.

29caseydurfee
Jun 9, 2010, 4:43pm Top

The big idea behind mixing in the Amazon ones is that we'd only show Amazon results for stuff that Overcat doesn't have. We want people using library records if at all possible.

30vaneska
Jun 9, 2010, 4:48pm Top

29: Now that really makes sense. Good.

v

31timspalding
Jun 9, 2010, 4:51pm Top

I wouldn't be quite that extreme, but we should preference library records over Amazon, providing Amazon among the alternatives. We also need to get the manual entries in there.

32Donogh
Jun 9, 2010, 5:17pm Top

I do like this - I have been frequently frustrated by clicking through every other library catalogue source for the odd obscure book.
This may be straying into equine territory, but is there anything in the pipeline to allow us to overwrite our bad catalogue entries (from Amazon for example...) with good library records using this?
You could call the feature OverWriteCat (and the user group OverWroughtCataloguers) *grin*

33rsterling
Jun 9, 2010, 5:25pm Top

We also need to get the manual entries in there.
Cool. That would help a lot with the "someone else has this book; why can't I add it?" queries, especially for obscure books.

34timspalding
Jun 9, 2010, 5:28pm Top

35TimSharrock
Jun 9, 2010, 5:33pm Top

I have added a few books with OverCat today - looking good. One thing I had a bit of trouble with - I tried clicking on some of the "other record" links, but decided I did not want any of those, and tried to get back to where I was before, but failed. (I cannot remember exactly how...)

36infiniteletters
Jun 9, 2010, 6:20pm Top

31: yes.

37jenniebooks
Jun 9, 2010, 6:23pm Top

just when you think LT couldn't get any any better .

38timspalding
Jun 9, 2010, 6:32pm Top

just when you think LT couldn't get any any better .

Ha. Well, on this thread a common feeling is that this was pointless.

http://www.librarything.com/topic/92665

39EveleenM
Jun 9, 2010, 6:36pm Top

I've tried replacing some of my amazon entries with OverCat, and it's worked out very well. I love it!

40lorax
Jun 9, 2010, 7:26pm Top

38>

I don't think that's entirely fair; just that it doesn't have me bouncing with glee the way it would have if I still had hundreds of titles to enter! (That reminds me, though, now that it's public, I need to go see how much it helps for Darwin's LL; not needing to separately search each library could be a godsend there.)

But I love Overcat, it's my standard first choice search now, and if it can help wean the site away from Amazon some more, it will be WONDERFUL.

41rsterling
Jun 9, 2010, 8:00pm Top

Yeah, for the record, I really like Overcat. I try to use library sources as much as possible, and Overcat often saves me time, in cases where I couldn't find a book in the first couple of library sources and had to work down the list of sources. (Or had to search WorldCat and then try to match a library found there with one that it's a source here.) I really like that it builds on the searches that users do, so that the next person to try to add my obscure book will benefit from my finding and adding it. I like that it moves towards less dependence on Amazon and towards more reliance on existing data/information within LT. I like that it emphasizes quality data from libraries while simplifying that to what seems like a single "source." Presumably(?), it is also a step that could it make easier to develop other cool features down the road (e.g. other things currently hampered by Amazon's restrictions on the use of their data - like a mobile app, maybe?).

42Littlemissbashful
Edited: Jun 9, 2010, 9:09pm Top

I added a few books to my wishlist just so I could try out OverCat ...

The books I added were all from the 'Shan Tao Yun' series by Eliot Pattison, I used the add to wishlist from existing books and then searched by the title pulled across

I'm not sure if this an 'English' issue but I found that it failed to throw up the normal UK paperback editions and - on a more general note - the records often did not indicate the binding.

Searching by title only on 'Lord of Death' the first search results pulled did not include exact title matches first but those merely including the key words anywhere in the title. By adding an author surname it went straight to the title record but for the hard back only.

Searching against Amazon UK gave more results indicating binding and usually threw up the mass market paperback first (or equivelent if published - i.e. not trade format) as the most recent record.

Eg. http://www.librarything.com/work/329738/book/60951302

On OverCat the top record is for the Century trade edition and does not indicate binding, the Arrow paperback does not appear on the first page

On Amazon.co.uk the top record is for the Arrow paperback

So, maybe the data and hierarchy is not so great for non US users (or at least UK users) who are more likely to have non American editions?

Just an initial impression, I have only added a few books to try it out ...

43Heather19
Jun 9, 2010, 9:21pm Top

I still think there's something Tim's not telling us. Has he finally converted to the dark side? Why is he letting a CAT on LibraryThing??

44caffron
Jun 9, 2010, 10:20pm Top

I think this will be particularly useful for avoiding Amazon data for less popular books...and it avoids the "how do I find a good source" issue which comes up a lot. And adding manual entries would be particularly nice.

My biggest suggestion would be some sort of tweaking of the display order (or perhaps clustering) to rank/group books in the language of the subdomain first (e.g., English books listed first on .com, French on .fr, etc.) It's good to have the other languages listed, but I think most readers collect in only a few languages, and many only in one.

And off-topic...could I get added to BETA? :)

45infiniteletters
Jun 9, 2010, 10:32pm Top

42: Amazon is your best source for ISBN searches, exact edition matches, and binding info. It is also more likely to have typos, duplicate records, and incomplete records.

Libraries have more (durable) hardbacks than paperbacks, more info for the books they do have, and less likely to have records for separate editions.

Overcat will help you find a book in general, particularly if it has been published in hardback by different publishers.

46lquilter
Jun 9, 2010, 11:03pm Top

> 43

Resistance is futile. Mmmrrreeeoow. Purrrrr, purrrrr, purrrr. Knead, knead.

47elenchus
Jun 9, 2010, 11:06pm Top

>43 Heather19:

There is a bit of backstory shared on the BETA group. It is related to an "overarching catalogue" and, to paraphrase, motivated by the desire to no longer be the underdog (i.e. OCLC).

But perhaps I've, er, let the cat out of the bag.

48sarahemmm
Jun 10, 2010, 3:40am Top

Wow! Excellent new Thing! I'm just sorry I don't currently have any books to add.

>43 Heather19: Heather, don't you have a friendly little cat helping you to add your books already? Mine is so helpful, he gets loaned out to friends!

49reading_fox
Jun 10, 2010, 4:33am Top

There will be a few clear divides on who finds this useful I suspect:

a) people who've basically finished cataloguing won't need it much

b) people who primairily read fiction especially paperbacks won't find it at all helpful

c) people who don't care that much about data quality beyond title/author/cover won't care about it

d) hardback non-fiction readers who do care about bibliograghic data will love it.

Is the proportion of d)s sufficently high on LT to justify the time spent to develop this, when so many more critical older features remain broken. Author roles for example.

50sarahemmm
Jun 10, 2010, 5:05am Top

You missed:

e) Members adding their books will love it

I am not a member of d or e, but I think the upgrade will make a huge difference to those people's appreciation of LT, as it will to:

f) LT regulars who work at correcting book data (eg combiners)

51justjim
Edited: Jun 10, 2010, 5:19am Top

>49 reading_fox: I think that's a bit harsh.

a) people who've catalogued most of their library, but who are still buying the odd book here and there, will like it

b) people who primarily read fiction paperbacks will find it gets better as more people search and that data gets added to OverCat, so they'll like it

c) people who don't care that much about data quality will find it quicker and easier (and they'll get better data anyway), so they'll like it

d) i) hardback non-fiction readers will love it

d) ii) anyone who cares about bibliographic data will love it.

Where's the downside?

52reading_fox
Jun 10, 2010, 6:04am Top

#51 how's your b) work? I thought this was a compiliation of library sources. No matter how many people search for a paperback, if it's not in the libraries, it's not in them.

#50 I did forget about Combiners.

53MarthaJeanne
Jun 10, 2010, 6:06am Top

I tried this on four books - two new (to me) US hardbacks, one each German and British paperbacks recently entered by me from library sources.

I will probably continue to use British Library first. For English language books, I might start using Overcat second. For German books it seems to be useless.

One reason I will use British Library first, is that it got knocked down to the bottom of my list (by LT when I tried to use 'add this book') and I don't want to have to rebuild the whole list. So it stays marked on the list.

54justjim
Jun 10, 2010, 6:29am Top

#52 Some library somewhere will have catalogued it. Some Thingamabrarian will search that library for that book. It will get added to Cat Overlord Overcat. Resistance is futile.

55quigui
Jun 10, 2010, 6:53am Top

I re-added a book so I could try this feature. For books is Portuguese will not be much use, because the sources are not very good to start with. If manual entries start to show up in the results, then it will be great.

For books in English I think it is already very good, and I will start using OverCat for everything, and if it can't be found there, then try Amazon.

I liked that in some results it said: "or choose from X alternate records from one library", and there was much more information on the "more" link for those alternate records.

56jjmcgaffey
Jun 10, 2010, 7:00am Top

49> I'm a member of a, b, and d - and I love Overcat. I don't have that many books to enter anymore - not more than 20 a month or so (library book sales are dangerous places. Especially if they do a bag sale on the last day...), but it's very nice to be able to enter (many of) those with library data. I read a lot of fiction, and I buy primarily paperbacks (more of them fit on a shelf!); Overcat is very bad at stuff published this year or last year, but for older things it's pretty good. I used to try LoC, maybe a couple more libraries, then get frustrated and go to Amazon - after one or two unsuccessful tries, I'd just use Amazon. Now I use Overcat and feel confident that I've tried pretty thoroughly - if I don't get anything for an older book I might try LoC anyway (since Overcat doesn't have all the latest LoC records), but falling back on Amazon happens far, far less often (once in ten times instead of 8 out of 10 - something like that). I still have to fix data - I HATE sentence case for titles, and I need to fix Dewey (fic in square brackets is just annoying, and the proper 813/823 ones aren't much better - I use Fic {genre} {author surname}) - but I'm still quite pleased with the results I'm getting.

And BTW, 53> MarthaJeanne, I just entered a bunch of books and I think 4 or 5 of them came from the British Library via Overcat (mostly Teys and Sayers(es?)). It's in there, at least for older books. I don't know about German.

57infiniteletters
Jun 10, 2010, 9:32am Top

This is an alternate easy source of data than Amazon, and Tim has said manual entries will be added in at some point.

This gives LT the option to get rid of Amazon and make a mobile app, so the people who have cataloged everything will be able to see their library in a bookstore sans internet access.

58jjwilson61
Jun 10, 2010, 9:38am Top

My understanding of the impetus behind this is that it is largely because of Amazon's changing their Terms of Service to be much more restrictive.

59catherinestead
Jun 10, 2010, 9:42am Top

I don't have many books to add - just new acquisitions - but I used OverCat to add a couple of books today. I had no trouble finding the exact book I wanted to add, and unlike Amazon it didn't give me three dozen similar books and other editions that I had to wade through to find what I wanted. So far, so good.

60vaneska
Jun 10, 2010, 9:59am Top

I just had the muted pleasure of trying OverCat for the first time in order to enter my 5,000th book. No joy on OverCat despite the fact that once I'd added it, I found 13 other LT-ers with it. It was on amazon, LoC, British Library but not Talis Union though.

v

61SylviaC
Jun 10, 2010, 11:03am Top

I like it. I've been inspired to move on to catalogue the next section of my library. I am having good luck so far with mysteries and picture books.

62calm
Jun 10, 2010, 11:16am Top

I've tried Overcat and like it. Good results are coming up from libraries I would never have dreamt of trying. Saves the frustration of going through the list or adding manually.

63DaynaRT
Jun 10, 2010, 11:21am Top

Maybe I missed something, but OverCat's not available for bulk import?

64richardderus
Jun 10, 2010, 11:25am Top

I've used it. I like it. It's now my default search setting.

65JamesBoswell
Edited: Jun 10, 2010, 1:58pm Top

For those of us in Jeremy's realm, cataloging old and obscure works, this is the greatest thing since haggis.

When you look to extend the data, I'd recommend the Bodleian at Oxford. They have a lot of unique holdings.

66ianreads
Edited: Jun 10, 2010, 2:05pm Top

Tim, three words ;-), 'Choose a source'.1 Very nice!

1 http://www.librarything.com/choosesource.php

67casvelyn
Jun 10, 2010, 2:46pm Top

I like it - a lot. Makes me wish I hadn't already cataloged 99% of my library and that I didn't catalog everything manually.

68Crypto-Willobie
Jun 10, 2010, 3:56pm Top

> 54

UberPanzer!

69timspalding
Edited: Jun 10, 2010, 4:50pm Top

people who primairily read fiction especially paperbacks won't find it at all helpful

I don't think fiction has anything to do with it. But yes, paperbacks are less often found in libraries.

sufficently high on LT to justify the time spent to develop this

Having invested the time to really nail down a perfect system—on top of an unprecedented data store—is going to make lots of things much easier to do.

Here's a list of improvements that naturally follow from it. That doesn't mean they'll happen tomorrow, but it means they can happen:

1. A fast source that is better than 690 other sources.
2. A platform on which to build combined Amazon/OverCat source which could serve 95% of users, forever.
3. We have already matched up most of Amazon books to parallel MARC data, so we aren't tied down by Amazon restrictions on mobile use. (In parallel we've also acquired another set of commercial data.)
4. The ability to move away from Amazon link policies, and insurance against future goal-post changing by Amazon.
5. A system that we can put manual edits into, so members can more easily bootstrap off each other's stuff.
6. The Subject improvements people want. Separation and reintroduction of other subject systems.
7. A revival and improvement to the current library-data (LCC, DDC, LCSH) recommendations system.
8. Page counts, running times and physical descriptions
9. Dimensions
10. Separating the publisher info line (ie., publisher, publisher place, publisher year)
11. Tables of contents for many books.
12. Other data contained in MARC records we're not using.
13. A system to allow members to upgrade records.
14. Imports that take seconds or minutes, not minutes or hours. (Yes, it's not currently an import source.)
15. Edition-combining and series boost from parsing the MARC records.

Anyway, we feel it's an important cornerstone for future growth--the sina qua non of real cataloging improvement across the site. As LibraryThing is among other things about cataloging, we think it's critical to have the best system we can have.

70TimSharrock
Jun 10, 2010, 4:56pm Top

69> Bravo!

I am looking forward to lots of these, even as it is it is making adding books much easier (1 above). Of the others 5 & 13 are the ones I am most interested in :)

71verenka
Jun 10, 2010, 5:12pm Top

I love it.
So far I've only catalogued the books I read last year and this year. Now I'm finally tempted to tackle everything on my shelves. when the next bad weather spell arrives..

72infiniteletters
Jun 10, 2010, 5:14pm Top

I think I won't be able to touch everything until 5 is working for manual entries...

73tros
Jun 10, 2010, 5:38pm Top

69.
"11. Tables of contents for many books."
This would be greatly appreciated.

74Heather19
Jun 10, 2010, 6:23pm Top

Wait, fiction/paperbacks aren't found very often in libraries? Since when?! My entire life, all my library experiences, I've found tons and tons of paperbacks. Or are you guys talking about University libraries or only certain big, professional libraries? 'Cause otherwise, I think something is seriously wrong with that assumption.

To address 49:

I have catalogued every single book in this house, and every single book I read in my childhood, and Overcat is STILL useful for me. Why? Because I'm STILL a reader and will be adding books to my wishlist and newly-gotten books, hopefully for the rest of my life.

I have added quite a few paperbacks with Overcat already. True, some of them don't turn up, but as has been pointed out repeatedly, this is a work in progress and it will only get more complete with time.

I'm sure that people who don't care about data quality will, at the very least, like the fact that they can search one source to find a book (even if they don't care about edition/correct-ness), rather then searching multiple library sources.

I read very few non-fiction books, and yet I love Overcat. LOVE it.

I'm sorry if I got ranty here, but generic assumptions against a brand-new feature based on absolutely no proof really gets my goat.

75brightcopy
Jun 10, 2010, 6:29pm Top

74> Are you talking about fiction paperbacks as a percentage of total books, or are you talking in absolute terms? It's been my experience in most libraries I've been in (large, small, city, university) that the fiction paperbacks are far outnumbered by the fiction hardbacks.

76timspalding
Jun 10, 2010, 6:32pm Top

In general libraries have fewer paperback books than booksellers. They don't last as long—simple as that.

77Heather19
Jun 10, 2010, 6:40pm Top

... Hmm. Maybe I'm just looking in the wrong sections? I do see mostly hardbacks in the research, true crime, history, etc sections. It's predominately paperbacks in the YA section, though, at least in the libraries around here.

78timspalding
Jun 10, 2010, 6:41pm Top

Yeah, I think paperbacks are more frequent in some sections than in others.

79suitable1
Jun 10, 2010, 7:05pm Top

Is there a way to update info using OverCat without deleting the existing record? (date entered, date read, etc)

80lorax
Jun 10, 2010, 7:06pm Top

I generally can find recent paperbacks in the public library sources on LT. (Much less so in the academic libraries, unsurprisingly, unless it's nonfiction that didn't have a hardcover edition first.) Older paperbacks far less so; they've generally worn out and been discarded, so unless you happen to find a library with a specialist collection in the area (like the UC Riverside for science fiction) it can be tough to find them.

81staffordcastle
Jun 10, 2010, 7:08pm Top

Well, all those people who have been asking for years for a way to search more than one source at once will adore it! It seems to be one of the top questions new people ask.

I'm just waiting for the next acquisition to come along, so I can try it out. Not to mention a block of time to play with it, which is harder to come by, lately :-(

82jjwilson61
Jun 10, 2010, 8:00pm Top

79> It's #13 on Tim's list in message 69.

83Heather19
Jun 10, 2010, 8:18pm Top

81: I deliberately spent an hour looking for books to add to my wishlist just so I could test out Overcat. hehe

84Mr.Durick
Jun 10, 2010, 8:27pm Top

Of two books yesterday, OverCat found one. Of four books today, OverCat found one. In both cases I searched by ISBN. Amazon found the ones that OverCat did not find.

Robert

85anthonywillard
Jun 10, 2010, 9:48pm Top

I am very happy with Over Cat so far. It comes up with the books and it is interesting to see the different libraries it goes to. I do not catalog many books that are very recent, being old myself, so maybe that's why I don't have some of the problems others may have. Keep up the good work! For a new database product, Over Cat is working well! Is it Überschatz in German? Nietszch!

86infiniteletters
Jun 10, 2010, 9:50pm Top

74: If given a choice between a paperback and a (non-cost-prohibitive) hardback edition of the same book, libraries will tend to acquire the hardcover. I was referring specifically to 42, who said that Overcat didn't find a particular paperback edition that Amazon had, but did find other editions. Also, since libraries have mainly hardback editions, they don't tend to show hardback/paperback binding info as readily as Amazon does.

81/83: Just take one of your existing books and search for it on Overcat to see whether it's easily available.

87infiniteletters
Jun 10, 2010, 9:51pm Top

84: Try the same Overcat search with title/author combo and see if you get better results?

88timspalding
Edited: Jun 10, 2010, 10:49pm Top

I've re-cataloging my whole library from scratch. (My library catalog never fully recovered from LibraryThing getting an office and taking a lot of my books with it. I never knew what was cataloged and what wasn't.) It's going to be a great opportunity for me to feel the pain of things that don't work perfectly—and correct them.

I'm finding Overcat invaluable, even for my paperbacks. I'm confident very few are records I found before, and am therefore the cause of.

My big complaint so far is that Overcat doesn't have the Library of Congress Catalog/Control numbers, which are the next best thing to an ISBN for book searches. I'm surprised nobody's complained about this.

Casey, what's the chance we could fold those in?

89brightcopy
Jun 10, 2010, 11:09pm Top

88> Actually, that sounds like an EXCELLENT idea (re-cataloging in order to get the full experience). I suggest that you owe it to the site to do it at least once a year.

>:)

90timspalding
Jun 10, 2010, 11:12pm Top

Well, I'm going to do all my books, all my wife's and my son's. That's going to take a long time, and I suspect it will end up around 4k...

91SylviaC
Jun 10, 2010, 11:14pm Top

>90 timspalding:
Guess you're going to have to upgrade to a paid account, then!

92PhaedraB
Jun 10, 2010, 11:58pm Top

I never knew what was cataloged and what wasn't

Right from the start, I put a little color-coding dot (of no particular color) on on the back cover of every book as I cataloged it. As I usually search by ISBN, if that is on the back cover the dot goes right over it.

It has been invaluable.

93timspalding
Jun 11, 2010, 12:01am Top

I was thinking I'd make a stamp. Which made me think: Would members buy a little "In my LibraryThing" stamp?

94MerryMary
Jun 11, 2010, 12:21am Top

Yep.

95shmjay
Jun 11, 2010, 12:28am Top

>93 timspalding: That would be just what I need to keep track of which physical books I catalogued (as opposed to the books I read).

96timspalding
Jun 11, 2010, 12:40am Top

97klarusu
Jun 11, 2010, 1:46am Top

So, this seems like a good reason to recatalogue my books to a big degree - I didn't worry about edition at all early on and I've been meaning to systematically recheck it all. I have questions though. For me to use Overcat, there have to be benefits because it basically returns titles in sentence case so I'd have to alter every record (just because libraries do it that way doesn't mean it's right or that I have to like it & I really hate it). So, I want to know simply:

1. Is it a positive thing for LT for me to move away from amazon as a data source? If it is, then that's enough for me to do that but I don't *hate* amazon in the way some people do.

2. If something isn't found in Overcat & I manually enter it, does this mean that it will then appear for others who search using Overcat? Or am I right now just servicing my own catalogue with manual entries?

98jjwilson61
Jun 11, 2010, 2:16am Top

Is there any interest in an option to change titles in sentence case to whatever-the-other-case-is-called case when adding a book? It seems that a significant number of people find that a barrier to using library records and if its a goal to move users away from Amazon then it might be worth it.

99SilentInAWay
Jun 11, 2010, 2:57am Top

>98 jjwilson61:

Excellent point. For the last thousand books or so, I've been adding books using library sources (mostly LoC) and then converting the titles from Library case to Bibliography Case. I can understand why some people might consider this a hassle...

100jjmcgaffey
Jun 11, 2010, 3:43am Top

Not-sentence-case is frequently (usually?) called...um, title case. Why libraries decided sentence case was better...

98> That would be _wonderful_. It wouldn't be perfect - what actually gets capitalized in a title varies too much - but if even 50% of sentence case titles could get automatically (via checkbox? button? profile option?) turned into title case it would save me so much time...And even the not-perfect ones would be better - in most titles, either every word is capitalized or most are (should be). So if every word got capitalized, it would either be right or nearly right, and I would only have to fix the nearlys (or ignore them as good enough).

101carport
Jun 11, 2010, 3:55am Top

>98 jjwilson61: and 99 I would love an option to change the case of titles, as I do it manually each time now. Changing title case causes ghost copies to be created, and other problems that plague combiners.

102verenka
Jun 11, 2010, 6:33am Top

> 54
(Sorry, couldn't resist)
I, for one, welcome our new Cat Overlord

103justjim
Jun 11, 2010, 8:45am Top

I, for one, don't believe it took 52 posts!

104felius
Jun 11, 2010, 8:49am Top

I catalogued 92 books the other day - the majority of them paperbacks - and it's a *huge* improvement for collections like mine. The books that I buy might have been published in Australia, or Australian editions of a book originally published overseas, or a book which was originally published elsewhere (usually the US or UK). I try to avoid Amazon data, but previously this meant I had to check the publication info for each book, then work through a list of the most likely libraries for books published in that country.

In the 10% or so of occasions that the book I was looking at wasn't in overcat, I could usually find it in a library source using the old method. Next step is to have that automated for me ;) (auto-selecting a likely library source based on ISBN?)

105DaynaRT
Jun 11, 2010, 8:55am Top

I cataloged my son's wishlist yesterday, about 30 books I pulled off his Amazon list. I searched by title (then ISBN if the title was too ambiguous) and OverCat only choked on a few. Those happened to be Scholastic published YA books that most sources that aren't Amazon seem to have problems with.

Is there a response to message #63?

106brightcopy
Jun 11, 2010, 9:57am Top

Another vote on the case-fixing. Even when title case is wrong, it looks far better than sentence case or all caps (I see that one a lot) until you get around to fixing it.

107jjwilson61
Jun 11, 2010, 10:09am Top

106> I believe that the feature in #63 was one of the items in Tim's list in #69.

108timspalding
Jun 11, 2010, 11:58am Top

I don't actually think you can reliably capitalize book titles automatically. The exceptions would blossom far, far too fast. (French books, for example, are capitalized in the library way, which is presumably what libraries aped when they decided on it.)

109brightcopy
Edited: Jun 11, 2010, 12:07pm Top

108> I don't think you can, either. But you can sure as hell give people the option, so you don't let the perfect be the enemy of the good.

110rsterling
Jun 11, 2010, 12:50pm Top

I like sentence case...

111jjwilson61
Jun 11, 2010, 1:13pm Top

108> While that may be true, the vast majority of Americans, at least, have been taught that title case is the only correct way to list book titles and are shocked to import library data and get it in sentence case. While it may not be perfect, I think it would reduce what resistance there is to using Overcat.

112timspalding
Jun 11, 2010, 1:25pm Top

The thing is, I agree that some people are annoyed by it, but I can't make it the default. And the people who are annoyed by it are, I think, not likely to poke around in the advanced options.

This may draw complaints, but I think I'm right.

113brightcopy
Jun 11, 2010, 1:25pm Top

If I look at my bookshelf, the vast majority of titles are in uppercase. That's simply not going to fly in my catalog, due to the fact that caps take up massive amounts of horizontal space.

The second most common (by an order of magnitude or two less) is title case (first letter of each word capitalized, lowercase for articles, etc.).

I have no books that use sentence case (first letter of first word capitalized, all other lowercase except for proper nounds, etc.).

I'm sure someone in a different country would likely have the last two paragraphs flipped in terms of title case versus sentence case. That's why I think it should be an option. Ideally, you could choose:
1) Title case
2) Sentence case
3) All caps
4) Don't automatically change

This would apply to newly added book records and could also hopefully be a function in power edit.

114brightcopy
Edited: Jun 11, 2010, 1:28pm Top

112> And the people who are annoyed by it are, I think, not likely to poke around in the advanced options.

I think the option should be right there on the Add books screen. No poking around necessary.

ETA: Let me add that I've completely ignored library sources because they had the wrong case for the titles. I stuck with Amazon, because it was more likely to have one with the case I liked, thus minimizing my corrections.

115PhaedraB
Jun 11, 2010, 1:40pm Top

I would like to be a good sport and use non-Amazon data, but the titles drive me bonkers. It's easy enough (for me) to change publication data, but it is tedious in the extreme (to me) to change the capitalization a word at a time.

Way back when I had a word processing program with an edit function that let you toggle between lower case and title case (also sentence and upper case). I'd love that. Something we could highlight and maybe right click would be amazing. Of course, I'm no programmer (I don't even play one on the Internet) so I don't know if that is pony or a Percheron.

116SylviaC
Jun 11, 2010, 1:42pm Top

>112 timspalding:

I think the people who are annoyed by it are the ones who pay attention to the details, and want to make it right. Those are the people who are more likely to poke around in the advanced options.

I would rather use OverCat and edit every title than use Amazon. But I would be happier if I didn't have to edit every title.

117timspalding
Jun 11, 2010, 1:46pm Top

I think the option should be right there on the Add books screen. No poking around necessary.

Space on a screen like that is precious. I don't think this rates anywhere but in the "Advanced options" area.

118MikeBriggs
Edited: Jun 11, 2010, 1:54pm Top

I agree with the first paragrah in 116.

I'm going to finally break down and ask: Is there a good thread or the like I can be directed to that will explain the deep feeling of dislike towards using Amazon as a data source?

119brightcopy
Edited: Jun 11, 2010, 2:11pm Top

117> Space on a screen like that is precious. I don't think this rates anywhere but in the "Advanced options" area.

Vertical space is virtually unlimited. There's nothing wrong with scrolling down. Even with Advanced options expanded, when I search for a title I wind up with the left side taking up only 2/3rds of the vertical space that the right side takes up.

120brightcopy
Jun 11, 2010, 1:57pm Top

118>

I'll attempt a summary:
1) Amazon puts a lot of restrictions on re-use of its data, which is a problem for LT re-using it for things like mobile
2) Amazon lists books before they're published, and often the data is incorrect due to this compared to the final data
3) Amazon COVERS change when amazon changes them. So if you use an amazon cover, it may change out from under you.
4) Amazon includes listings from 3rd-party sellers of used/new books, and often that data has errors in various places.

There are a lot of other points, but I hope I've hit the highlights.

121vaneska
Jun 11, 2010, 1:59pm Top

119: While for me the reverse is true. Just to make the point that this is not necessarily so.

v

122keristars
Jun 11, 2010, 2:02pm Top

118> I'd like the same regarding OCLC. I think I get the gist of why LT/LTers don't care for it, but all I see is praise in my library school textbooks, and there's a bit of a disconnect.

But the dislike for Amazon is because of faulty data (which we get from library sources anyway) and also using their data prohibits LT from developing a mobile app, because of their terms & conditions. As far as I can tell.

123brightcopy
Edited: Jun 11, 2010, 2:11pm Top

121> That seems a bit odd. Is it because you added a bunch of sources and/or collections? I'm not really claiming it's true if you do that. I'm just trying to hit more of the "normal" case. When you start getting people who add tons of sources and collections, I'm going to postulate they're also going to be a lot more capable of scrolling down a little anyway.

Just for reference, here's a search for "grapes of wrath" using Amazon, then Overcat, just to show what I'm talking about on vertical space:





ETA: These are fulls screenshots of the page, not screenshots of what I can view in the window at one time.

124vaneska
Jun 11, 2010, 2:22pm Top

123: The reasons why are irrelevant. It isn't odd - it is just different from you.

v

125nathanielcampbell
Edited: Jun 11, 2010, 3:08pm Top

I do have a few books that use sentence case for the titles: all of them are either 19th-century, editions of medieval texts, or modern French or Italian.

In German books, English-style sentence case is, in fact, wrong; nouns are always capitalized (at least in modern German; medieval and early modern German orthography eschewed what we know as capital letters). If the person who entered the library catalogue entry was smart, they entered the German correctly; if not, they used English-style sentence case and failed to capitalize the nouns. For example, a book I entered just a few days ago (before OverCat), Sacrum Imperium, came from LOC with the subtitle in all lowercase letters, which I then had to edit. Of course, I still preferred to add it from LOC rather than Amazon, mainly because of the LOC call number.

Nevertheless, a toggle would be a nice feature to have in "Advanced Options". Of course, there are other nice features that I'd like to see more, so....

Edited because touchstones hate me. Even after editing...

126klarusu
Jun 11, 2010, 3:09pm Top

I'm not too sure that an automatic capitalisation thingummybumpkin would do the trick. To be honest, I'm quite happy putting the time in to alter them if I use Overcat as a source but my questions in #97 still stand:

1. Is it good for LT if I move away from using amazon as a source? In which case, it's worth the effort even though I don't have big issues with amazon myself.

2. What about my manual adds? Do they benefit everyone now we have Overcat or still just me?

127nathanielcampbell
Jun 11, 2010, 3:16pm Top

>126 klarusu:, question 2:
A quick check just now with Overcat and a work I've had to enter manually shows that it hasn't shown up in Overcat...at least not yet.

128timspalding
Jun 11, 2010, 3:25pm Top

Right. At present, manual ads are not being added to OverCat.

129jjwilson61
Edited: Jun 11, 2010, 3:40pm Top

Even if it were in advanced options, at least if someone complained about it in Talk we'd have somewhere to point them to.

130brightcopy
Jun 11, 2010, 3:44pm Top

124 The reasons why are irrelevant. It isn't odd - it is just different from you.

That makes no sense. If it's because you cranked your fonts way up but left image sizes the same, or vice versa, it's both odd and the reasons are relevant. If it's because you're using IE4, again, it's both. Etc. etc.

I'm going off the definition of odd along the lines of "not usual or ordinary; singular; peculiar; strange". If some aspect of your setup is common to only you (or you and less than 0.1% of LT population), then it is quite possible that the fact that it's different might in itself be irrelevant.

But you brought it up, so I asked.

131carport
Jun 11, 2010, 3:53pm Top

116> I would rather use OverCat and edit every title than use Amazon. But I would be happier if I didn't have to edit every title.

I agree ... and would highly appreciate the advanced option that brightcopy proposes in message 113.

132nathanielcampbell
Jun 11, 2010, 3:53pm Top

133jjwilson61
Jun 11, 2010, 4:04pm Top

131> I think you'll find that jjwilson61 first proposed it in msg #98.

134timspalding
Edited: Jun 11, 2010, 5:12pm Top

The problems with Amazon are as follows:

1. Reliance on Amazon hurts LibraryThing's ability to use the data, to provide the data by API, and requires us to promote Amazon above all other retailers.

2. Amazon data is often from "sellers," often used-book sellers, whose data is of very low quality--titles in all caps, erroneous dates, misspellings, etc. When you're selling one copy of something, getting the data right just isn't that important.

3. Amazon data lacks Subject headings, LCC classifications and the DDC (Dewey) classifications are truncated at the top level. Author names aren't separated first vs. last so LibraryThing has to guess, producing some very odd results. You know, stuff like II, Pope John Paul :)

4. Library data is just better. The parts we now show are richer and checked better, and there are parts we're not showing—physical description, some TOCs, etc.—that we could show which would improve things further.

135klarusu
Jun 11, 2010, 5:06pm Top

Well, that sounds like a good enough list of reasons to re-catalogue my amazon entries.

136jmnlman
Jun 11, 2010, 5:25pm Top

As mentioned up thread this is awesome for legacy libraries. After stalling around for months I finished off entering Basil Henry Liddell Hart last evening.

Two quick questions for Tim.

1.There was talk on the beta group about making OverCat the default for new accounts is that still the idea?

2.Since it was introduced do we know what percent of added books were added with OverCat? Just curious.

137timspalding
Jun 11, 2010, 5:45pm Top

We'll make it the default when we integrate Amazon into the results.

I don't know what percent are OverCat-added. Actually, I need to make sure we know. The source is actually listed as the library that made it. Then again, I can watch the library percents and see if they creep up.

138carport
Jun 11, 2010, 7:41pm Top

>133 jjwilson61: indeed you did, jjwilson61. sorry I overlooked it!

139caffron
Jun 11, 2010, 8:01pm Top

It might be sad for the capitalization-picky soon. RDA provides an "option" for a library to accept whatever capitalization is found in a digital record, no matter how rough. LC has chosen this option, which means that ONIX/publisher data, often awful and in all-caps, will be accepted as-is rather than corrected by LC. I find it ironic that LT is trying to improve metadata quality while the leading cataloging agency is deciding to let it slide. Anybody who reads large chunks of ALLCAP text knows that this will make catalogs look horrible.

http://www.loc.gov/catdir/cpso/RDAtest/RDA_alternatives.doc

140SilentInAWay
Edited: Jun 11, 2010, 8:28pm Top

>137 timspalding: I don't know what percent are OverCat-added. Actually, I need to make sure we know. The source is actually listed as the library that made it. Then again, I can watch the library percents and see if they creep up.

Wow, really? Last night, I tried to determine how many of my recent book loggings had been entered using Overcat, and was surprised to find that only the source library was listed on the book page. However, I'm really surprised to learn that there's no trace at all within the database. I would think that, after spending months developing a truly excellent feature such as this, you would want to monitor the rate at which it is adopted by members.

Although you could extrapolate some general conclusions from the changes in percentages for libraries contributing to Overcat, it would be hard to make definitive arguments one way or another. Also, recent discussions about data quality could result in some users abandoning Amazon and moving to, say, Library of Congress. Would you interpret a marked increase in LoC loggings to the adoption of Overcat or rejection of Amazon?

IMHO, it would be worth your while to somehow mark in the database which loggings used Overcat -- even if you lose the first couple of days worth of data...

141Katya0133
Jun 11, 2010, 8:43pm Top

I'll 3rd (or 4th?) the comment that this is a godsend for legacy libraries. I just got a chance to use it with Alexander Pushkin. Up until now, Yale had been my largest source, but they'd accounted for only about 15% of my matches. Today, I'd say that I've been able to get about 70% of the records I need from OverCat. It's going to make a huge difference in productivity if I don't have to pre-search everything in WorldCat.

142Heather19
Jun 11, 2010, 9:16pm Top

..... I have over 1.7k books in my catalogue. I've spent countless hours tweaking/reviewing/cover-uploading/etc. ..... Am I *really* thinking about re-cataloguing everything because of Overcat? *faints*

Seriously though, I'm itching to REALLY test out this thing. I might re-catalogue everything from the first month, at least, back before I grasped the importance of data.

143timspalding
Jun 11, 2010, 10:14pm Top

>139 caffron:

I'd like LT to start using the alternate title info—spine title, cover title, inside title, etc.

144SilentInAWay
Jun 11, 2010, 10:38pm Top

>143 timspalding:

But first we'd need the ability to distinguish between (combine/separate) editions, right?

And then--gasp--Common Knowledge for editions!!

145timspalding
Edited: Jun 11, 2010, 11:01pm Top

I think editions per se is virtually impossible to pin down. We just don't have enough info about what people have. There's way to much guessing involved, and weird things that aren't any definable, existing edition. It's got to be by ISBN only, I think. That would allow a lot anyway.

146SilentInAWay
Jun 11, 2010, 11:14pm Top

Agreed. For the purposes of alternative titles (and most other things), that's all that would be needed -- and you've already got most of the mechanism in place, no? I mean, when we click on one of the Quick Links to Amazon or Abebooks we're presented with a set of editions by ISBN, right? Also the Cover "game" was hooked up to ISBNs, right?

Actually, I may be falsely assuming that alternative title info would be Common Knowledge. Do you imagine ISBN-specific stuff like this being Common Knowledge or just private data in the member's book record?

147timspalding
Jun 11, 2010, 11:43pm Top

No, there are two links there. Amazon and a quick-link. They're different.

148SilentInAWay
Edited: Jun 12, 2010, 3:53am Top

I understand. The header over that entire section is Quick Links. The link labeled "direct" is, of course, even quicker.

But...back to Overcat.

I added an art book using Overcat last night and was presented with two primary choices. I decided that the data from the Metropolitan Museum of Art was slightly better than that from the Library of Congress. Just to see how things work, I entered the same ISBN this evening and was presented with one primary choice--the one from The Met (several others, of course, being only a click away).

My question is this: Did my selecting the data from The Met last night actually influence the way in which Overcat presents the results for future searches?

149timspalding
Jun 12, 2010, 12:52am Top

Casey?

150jjmcgaffey
Jun 12, 2010, 3:03am Top

117> Sure, put it in Advanced Options. Those that care (me!) will go find it, those that don't won't see it. Hmmm...more opinions will go in the Toggle thread (msg 132).

151klarusu
Jun 12, 2010, 6:25am Top

It's taking me about, oh I don't know, forever to add books using Overcat in comparison to using amazon as a source (even taking into account some of the fixes I have to do on amazon titles - which is pretty much balanced by the fact that I have to alter anything in sentence case meaning that I actually edit title data far more often than if I'm using amazon as a source). That said, the quality of the data I'm putting in is much better and it's worth the effort. I may have missed this but Overcat doesn't seem to be particularly good at search relevance order - is this right or am I imagining it?

I like Overcat as a source and, having played with it, I'm certainly going to work on a long term plan to eradicate amazon sourced entries from my library and make sure I've got my editions right. That said, I still think that amazon is an easier source for people new to LT to add from. Whilst a lot of people are very specific about editions and back-end library data like classification numbers, I would imagine a large number of users attracted to LT as a social cataloguing site probably just want to get their libraries up and interacting as soon and as easily as they can. Certainly, I was like that way back when I started using LT. I think amazon is a better source for that. I'd change my view on this if I got better relevance returns from Overcat, I think. Also if you could eventually add from generic editions via the work page without going to Add Books.

I'm just pleased that I can recatalogue my stuff now. It's like joining all over again with the hindsight of a few years LT-knowledge (I am aware I could be quite damaged for looking forward to that!).

152skittles
Jun 12, 2010, 9:05am Top

a cautionary note on ISBN's in book/work records:

when someone chooses an amazon book cover, the ISBN will change to the ISBN of the cover image. There will be no record of the original ISBN number for the member's book... andthe member will probably not be aware that the ISBN has been changed.

Until this is changed, we cannot assume that the ISBN listed is the ISBN of the edition that the LTer has.

153DaynaRT
Jun 12, 2010, 9:14am Top

To add to what skittles said, the pop-up warning used to state that the ISBN would change when choosing an Amazon cover. Since this wording was removed it's now even more likely that people have no idea that their ISBNs are being edited without their consent.

154klarusu
Jun 12, 2010, 9:26am Top

Yeah, I remember that message. Didn't realise it had gone away 'til you mentioned it.

155EveleenM
Edited: Jun 12, 2010, 10:53am Top

I've been experimenting with replacing amazon records using OverCat, and one thing I've noticed is that it often seems to miss the 'Member-uploaded cover, best-guess for ISBN' step in choosing covers. I've re-entered a number of books where I know there's a member-uploaded cover for that specific ISBN, because I uploaded it myself, and it's sitting there in another open tab, but when I enter the ISBN using OverCat it gives me a blank cover.

eta:
It's also annoying that adding from library sources is bringing in Amazon covers when there are good member-uploaded covers for those ISBNs.

156brightcopy
Jun 12, 2010, 12:23pm Top

152/153> Oh dear lord. Why oh why isn't there a message? I was never around when there was one, and therefore had no idea. While I endeavor to replace all my amazon covers, at times I replaced empty covers with amazon ones until I could get around to getting the proper one. You're telling me it munged up my data without bothering to say a word about it?

Et tu, LT?

157rsterling
Jun 12, 2010, 12:24pm Top

There used to be a warning message about this, which probably fell through the cracks at some point during a coding update.

158SilentInAWay
Edited: Jun 12, 2010, 1:06pm Top

Ok, I think its time to give due attention to the elephant in the room:

LT members like Overcat. In fact, many members want to recatalog(ue) their original amazon entries in order to get the better data available through library sources. However, the process of moving over one's own customizations to the new book record -- adjusting the author, title & publishing data to one's preference; copying tags, start/stop dates, & comments (public and private); reselecting covers, etc. -- is time-consuming. Especially for those who wish to recatalog(ue) their entire library!!

It would be supremely useful if there were an "update book data" feature that allowed you to first use Overcat (or some other source) to select an edition, and then choose which fields to update in your existing book record. The two versions of each field could be listed next to each other with the option to keep either one. Personal data (tags, comments, user-selected covers, start/stop dates, etc.) could be retained automatically. A quick and dirty dialog would be all that's needed --containing only those fields that can be lifted from a library source -- and with only minimal attention to aesthetics.

Is this is an elephant or a pony?

159skittles
Jun 12, 2010, 1:10pm Top

#158 & others:

please see this thread:

http://www.librarything.com/topic/92774

160klarusu
Jun 12, 2010, 1:13pm Top

I don't care what it is silent, I wants one. It's taken me ages to copy across each record's data for the entires I've done so far and they didn't even have reviews attached. I haven't touched the reviews and they'll probably always be amazon sourced as I'll lose all my thumbs if I transfer them across. I don't know if it's feasible to do this but it's certainly desirable. If it's a good thing for LT to get us away from amazon source data, then maybe Tim could help it to be as painless a process as possible for us.

161timspalding
Jun 12, 2010, 1:20pm Top

We're definitely going to have an upgrade feature--part of a more general feature. It won't be very soon, though.

Klarusu: What was the cause of the slowness? Did you find OverCat slow to add records?

162SilentInAWay
Edited: Jun 12, 2010, 1:43pm Top

>159 skittles: Ah, thanks skittles -- I was surprised that noone seemed to have mentioned it here (hence the elephant) -- elenchus has got it right in the other thread, I think: it's definitely a pony with wings!!

163klarusu
Edited: Jun 12, 2010, 1:53pm Top

Tim in #161> No, rather I meant that found the process slower rather than Overcat as a source itself. Partially, it may be because I'm still negotiating it as a source but the real issue for me was the relevance of source returns. I did a search for 'Mort' looking for the Pratchett book, for example. If I use amazon, the first page of returns was mostly populated by that title, whereas Overcat returns one entry on the first page, two libraries and very little information on the records from those. Ultimately, I found a good record by refining my search through Overcat for 'Mort Pratchett' which return 64 hits but I would have expected a first page more like the amazon one for my first search term as it was the exact match rather than some of the other obscure results. Overcat added the records fine though.

I like Overcat, don't get me wrong, but I just get the impression that it requires more dedication to find the correct work, edition and get the best data than simply searching amazon if you're dealing with common paperback editions.

164timspalding
Jun 12, 2010, 2:47pm Top

>163 klarusu:

It's definitely got worse relevance than Amazon, especially for short stuff. It also has no spelling correction at all. Once thing we plan to do in integrating the Amazon stuff is to have the Amazon results inform the OverCat results on ordering issues, even if we don't end up using the record itself as the primary one.

165klarusu
Jun 12, 2010, 3:29pm Top

Thing is, now Overcat has made me realise how ropey a lot of my catalogue data are, there's no going back. The only issue I have is with the ordering.

166jjmcgaffey
Jun 12, 2010, 5:45pm Top

I've been searching it entirely by 'title, author' - not the full title, just some of the words, and whichever of the author's names is least common (for Pohl Anderson I'd definitely search Pohl!). And my relevance has been excellent. I haven't tried pure title searching - well, Add to Wishlist has and it's been pretty poor, but I thought that was because it's new to not-yet-published books, which Overcat is poor at anyway.

167Bookmarque
Jun 12, 2010, 6:07pm Top

earlier -

pure title search + overcat = no results
pure title search + amazon = book found

YMMV

168timspalding
Jun 12, 2010, 6:33pm Top

OverCat works with titles, if they're in there. OverCat is clearly better than any library source we have, but it's not perfect.

169Heather19
Jun 12, 2010, 8:07pm Top

168: As I mentioned in the beta thread before it's release, I'm still getting much better results on Nancy Drew books in LoC then in Overcat, even though Overcat supposidly includes LoC.

170timspalding
Jun 12, 2010, 8:10pm Top

Casey: Is the problem that the LC only sold LC-derived records to OL?

171vaneska
Jun 13, 2010, 4:52am Top

169-70 - that might help explain my experience in post 60. The book in question published in 1995 so not very recent.

v

172MarthaJeanne
Jun 14, 2010, 2:11am Top

This was interesting. I am entering a few of the Donna Leon books - paperbacks. Of the four I have entered so far, two have showed up properly in British Library on an ISBN (10) search.

One did not turn up on an ISBN 13 search, but Overcat found the right British Library entry - I then edited the ISBN 10 to ISBN 13.

One did not show up in BL or Overcat on ISBN 13 search. I then searched BL on Girl Dreams, Leon. No luck. Overcat found a BL book on that search, but with a different ISBN.

173klarusu
Jun 14, 2010, 5:06am Top

Glitch with tags/collection assignment when using Overcat

Found an annoying glitch with adding from Overcat. I normally add books as follows:

1. Fill in title field
2. Fill in tags field
3. Fill in collections info
4. Search
5. Click entry to add (with above info 1-3 added too).

When using Overcat, it works fine if I just select the first entry on the add books page but if I click on 'More' and view the alternative library sources, then click on one of them to add, none of the tags and collections info comes across. This is slowing me down quite a lot because I have to go back and re-enter this every time I look further into the source. I think that it's already counter intuitive to have to add the tags before the you search rather than before you add (as has been discussed elsewhere on various occasions when people add things and think that it's a bug that the tags don't come across, not realising they should have added them before searching) but once you've trained yourself to add them before the search, it should work the same with Overcat.

174skittles
Jun 14, 2010, 9:53am Top

#173: klarusu, I've always had that happen, even before overcat. Irritating, but common.

175infiniteletters
Jun 14, 2010, 10:08am Top

174: But she's adding the tags/collections info before searching. The problem is that it's not transferring if she clicks on more to view an alternate library source.

(And the more for alt library source wasn't available before Overcat).

176skittles
Jun 14, 2010, 10:12am Top

#175 & 173: I add tags before I post/enter a book... before collections, if I clicked 'more' (not one of my additional sources) then I lost the tags. I added more sources & that helped the problem.

177klarusu
Jun 14, 2010, 10:14am Top

Yep, what infiniteletters said ;-)

It's just clicking on 'more' for Overcat that loses the tags. As long as I search after listing tags, they are always added when not clicking on 'More'.

178klarusu
Edited: Jun 14, 2010, 10:23am Top

Sorry skittles, cross post. Then I think this needs to be fixed with the advent of Overcat as actually navigating within the add books right-hand panel shouldn't behave differently depending on which level you navigate to. I might cut some slack if it was simply adding more sources using the list below the search box but clicking on anything in the Add Books panel should behave consistently.

179skittles
Jun 14, 2010, 10:28am Top

oh, I agree, but just wanted you (& anyone else) to know that it wasn't a 'new' problem. I'm sure it has been mentioned before, too.... but I'm too lazy right now to find the talk posts about it.

180klarusu
Jun 14, 2010, 10:43am Top

I probably didn't notice because I haven't used that many sources to add my books and those I have used, I added to my list.

181vaneska
Jun 14, 2010, 11:30am Top

I discovered a pile of books right in front of my nose that somehow hadn't got catalogued - a series I'd put aside to find appropriate shelving then kind of forgotten. They are coming up beautifully in Overcat. One thing I am noticing though: adding via Overcat seems to add a ghost copy, not just to my book but to the copies owned by other LT members.

v

182timspalding
Jun 14, 2010, 1:20pm Top

seems to add a ghost copy

You mean the member count increments twice?

183vaneska
Jun 14, 2010, 1:28pm Top

No. If you go to the editions page, it always shows 2 copies, even when there is only one member with one copy and with no editing having taken place. I noticed this applied to my copies then realised it was also applying to other people's copies that were already there.

v

184timspalding
Jun 14, 2010, 1:40pm Top

Interesting. I'll look at it when the current process finishes--it's slowly finding and isolating all no-copy editions. It's not deleting them, but it will.

185ty1997
Jun 14, 2010, 1:40pm Top

I removed a couple of books from my library (three Dickens novels, in this case) to re-add them via Overcat. I searched Overcat by title. I was surprised (at first, but then it made sense) that some of the Overcat records that came back were for DVDs/VHS of mini-series and movie adaptations. I presume it would be difficult (if not impossible) to filter these records out of results.

186timspalding
Jun 14, 2010, 1:41pm Top

Casey: Don't you have this magic algorithm for identifying them?

187vaneska
Jun 14, 2010, 2:01pm Top

184: Tim, actually this has nothing to do with Overcat. I started looking around and it seems to me that every single book now shows 2 copies on the editions page. Something awry with the processes you are running?

v

188vaneska
Jun 14, 2010, 2:05pm Top

Also, I know this is off-topic but perhaps associated with the above: something has started going funny with combining/separating with books staying on 2 author pages instead of moving over to the majority one. I seem to remember a similar problem being reported recently which you sorted quickly?

It was fine earlier today.

v

189caseydurfee
Jun 14, 2010, 3:11pm Top

186 > Yes, we have an algorithm for guessing the physical format, though it isn't foolproof by any means. We could definitely be using it to restrict to book results. The format is already in the search index, even. It's just not exposed through the search interface on the addbooks page.

Should it be doing it automatically, or should we have a way for the user to manually add that requirement? Behind the scenes it is just a matter of adding " AND format:Book" to the query before it goes to Solr. So the question is how to best expose that capability for users.

190caseydurfee
Jun 14, 2010, 3:17pm Top

Oh, forgot to announce this. At Tim's recommendation, I added a way to search on LCCNs through overcat. You just have to add "LC" at the beginning of the number, for instance to find LCC# 327017587 in overcat, you simply search on:

LC27017587

I understand it would be nice if you didn't have to add the "LC" bit at the beginning but there are technical reasons behind it (the LCCNs aren't in the main keyword search index.) Those could be overcome but not without more work than I can devote right now.

191caseydurfee
Jun 14, 2010, 3:18pm Top

177 > "It's just clicking on 'more' for Overcat that loses the tags. As long as I search after listing tags, they are always added when not clicking on 'More'."

Crap. I'll look at it. I thought I got that part fixed, but clearly I did not.

192klarusu
Jun 14, 2010, 3:22pm Top

#191, Cheers! It'd be great if you could because I'm a heavy tag user & migrating records to ditch the amazon source is a bit of a pain when it loses tags.

193timspalding
Jun 14, 2010, 3:23pm Top

>190 caseydurfee:

There's code in there that picks out the LCC requests. I'll find it.

194SilentInAWay
Jun 14, 2010, 3:40pm Top

>190 caseydurfee: I added a way to search on LCCNs through overcat.

Excellent!! I was looking for this yesterday. (Although the brief keyword search that I ended up using found the book easily enough)

195caseydurfee
Jun 14, 2010, 5:39pm Top

193 > Yeah, I looked at that. That's only for the older-style LCCNs that have a dash in them. Stuff that matches this regular expression:

#(^0-9{2})\-(0-9{3,6})#

Those will get treated as an LCCN by overcat without having to preface them with "LC" (this is inc_remote_search if you want to change the logic.)

The newer ones are 10 digit (sometimes more) numbers w/o any kind of punctuation, and no check digit. Not that there would be anything wrong with treating *any* 10 digit number as an LCCN if it's not an ISBN, I guess.

196prosfilaes
Jun 14, 2010, 7:01pm Top

#195: Don't all the new ones start with the four digit year? 20(01)(0-9)* cuts it down quite a bit.

197PhaedraB
Jun 14, 2010, 9:19pm Top

> 188

re: ghost copies, it's happening all over. I've been doing a lot of separating and combining, and the copy count on the edition page has been very consistently x2.

198carport
Jun 14, 2010, 9:34pm Top

>188 vaneska: and >197 PhaedraB:

same for me. When combining or separating, I consistently see double the number of actual copies. This has happened since the appearance of the "no current copies" entries.

199wyvernfriend
Jun 15, 2010, 5:37am Top

>93 timspalding: a stamp I would buy, I keep threatening to get a Wyvernfriend's Library Stamp!

200prosfilaes
Jun 21, 2010, 4:35pm Top

It looks like that Overcat, when you search with an ISBN, will sometimes return a book with a different ISBN? Why? I've started using Amazon because of that; if I'm searching by ISBN, I'd rather nothing turn up and I have to search by the title rather than have a book with the wrong ISBN show up.

201timspalding
Jun 21, 2010, 6:45pm Top

Casey, can you reply to that. I think that it's a case where the Overcat record has *multiple* ISBNs. In such cases, I think we should lie, and turn the secondary ISBN into the top one. Doable?

202jcbrunner
Jun 21, 2010, 7:43pm Top

>200 prosfilaes:+201 The secret switching of ISBN caused me quite a bit of trouble recently.

Search overcat for 1848325479.
It returns a record for 9781848325104/184832510X - which is mean, as you add vol. II instead of the new vol. III. The titles differ only in the second part of their subtitles, so the switch isn't immediately obvious (Amazon returns the correct volume).

203timspalding
Jun 21, 2010, 7:46pm Top

The problem there is actually a little different. Libraries consider many multi-volume works to be one record.

204jcbrunner
Jun 21, 2010, 7:49pm Top

>203 timspalding: Shouldn't it then logically return the first volume as the main entry (instead of volume II)?

205timspalding
Jun 21, 2010, 7:54pm Top

Not sure. What library did yours come from? Let's look at the rcord.

206jcbrunner
Jun 21, 2010, 7:58pm Top

A. Overcat says:
1809 : thunder on the Danube : Napoleon's defeat of the Habsburgs by John H. Gill (2008).
from British Library (powered by Talis); retrieved 2010-04-12
ISBN: 9781848325104
Publication: Barnsley : Frontline, 2008.
Dewey Call: 940.27
Subjects:
Napoleonic Wars, 1800-1815 > Austria. > Campaigns
Wagram, Battle of, Deutsch-Wagram, Austria, 1809

B. As no library had a copy at the time I bought the book, I chose data source: amazon.co.uk

207timspalding
Jun 21, 2010, 8:25pm Top

Casey?

208caseydurfee
Jun 21, 2010, 11:40pm Top

Yes, the library record in question is for all three volumes, and has ISBNs for all three on the record.

It's odd that it was taking the 2nd volume one as the 'primary' ISBN. I changed what was causing that (using an unordered set object to deduplicate results -- now I do it in a way that preserves the original order.) So it'll give you the 1st one first, now (though MARC doesn't really make that distinction -- the order shouldn't matter. From MARC's standpoint, they're all equally valid ISBNs for the record.)

I can definitely make it 'promote' the one you searched on to be the primary ISBN as well, if that isn't too much magic.

209timspalding
Jun 22, 2010, 12:40am Top

>208 caseydurfee:

Yes, I think you should to that.

210MarthaJeanne
Jun 22, 2010, 5:57am Top

>208 caseydurfee:-9

For all library records, whether or not through Overcat, please. Pretty, please with sprinkles and a cherry on top. This is particularly a problem with LoC.

211caseydurfee
Jun 22, 2010, 11:23am Top

OK, I've made the change, for both library sources and Overcat. If your search term is an ISBN, it'll use that as the ISBN for the book when it saves it, even if there are other ones.

212elenchus
Jun 22, 2010, 11:37am Top

I like this change, too -- nice attention to detail, LT.

213MarthaJeanne
Jun 22, 2010, 11:41am Top

Wonderful!

214TimSharrock
Jun 22, 2010, 11:42am Top

great :)

215lquilter
Edited: Jun 22, 2010, 1:32pm Top

Umm, did this possibly mess something up?

Because now when I search-to-add from any source, I get this error message:

Warning: Invalid argument supplied for foreach() in /var/www/html/inc_remote_search.php on line 1369

216caseydurfee
Jun 22, 2010, 2:27pm Top

Yes, it did, sorry. Fixed now.

217MarthaJeanne
Jun 23, 2010, 11:29am Top

211>

This is not working right. Searched on 9780571221837 found by Overcat with ISBN 0571221807 / 9780571221806

218infiniteletters
Jun 23, 2010, 4:34pm Top

217: Uhoh, it's leaving off the initial 978 then...

219JGKC
Jun 23, 2010, 4:44pm Top

Clicking on "more" isn't always working for me but there's no discernible pattern to which results will work and which won't.

220MarthaJeanne
Jun 24, 2010, 2:13am Top

218> No, there are differences at the end. It has found the right book, but put the wrong ISBN in the record, just what was not supposed to happen any more.

221jasbro
Jun 24, 2010, 4:14pm Top

It seems I’m late to the party, as per usual.

17>Turns out a lot of us add books through the Add books page, completely apart from visiting a pre-existing work page.

Is it really possible to add books any other way?!?

I’m with FicusFan (14); whenever I press on “Add this book” or “Add this book to your Wishlist,” it doesn’t, but makes me enter it from an outside source. If I’m “dune it rong (noob!),” I wish somebody would tell me how to “du it rite.”

25,27,29>Amen!

98>Absolute interest in Initial Caps over sentence case. WordPerfect calls it “Initial Caps,” but recognizes definite articles, prepositions, and other stuff that shouldn’t be so; Microsoft Word calls it “Title Case,” but every initial letter gets capped (Ugh!). e.e.cummings give ‘em both fits. ;-D

173>I often find it handy to do extended text in a word processing application, then cut & paste it into the applicable LT data field; makes quick work of author names with accent / other non-standard (U.S.) characters, elaborate subtitles (otherwise, try entering E.P. Thompson’s, The Sykaos Papers with its full subtitle: “Being an Account of the Voyages of the Poet Oi Paz to the System of Strim in the Seventeenth Galaxy; of His Mission to the Planet Sykaos; of His First Cruel Captivity; of His Travels About Its Surface; of the Manners and Customs of Its Beastly People; of His Second Captivity; and of His Return to Oitar. To Which Are Added Many Passages from the Poet’s Journal, Documents in Sykotic Script, and Other Curious Matters. Selected and Edited by Q, Vice-Provost of the College of Adjusters. Transmitted by Timewarp to E.P. Thompson”!)

182>Tim: I’ve also noticed recently that, in separating / combining to clean up editions, several editions have “wavered” between 1 and 2 copies. I don’t know enough about how it works to speculate why, only that it caused me to wonder about the accuracy / stability of some of our counts.

222lorax
Jun 24, 2010, 5:08pm Top

221>

Turns out a lot of us add books through the Add books page, completely apart from visiting a pre-existing work page.

Well, of course.

I've never understood people who do otherwise. Occasionally, sure, especially for wishlist -- I often add wishlist books from the work page myself. But people whose routine method of adding a book is to search on LT first, then click on the green plus, rather than just going directly to Add Books? That's just bizarre.

223TLCrawford
Jun 24, 2010, 5:26pm Top

Ok, I am bizarre.

Maybe I am just getting more disciplined about buying only books that are on my wish list.

No, that can't be it.

224PaulFoley
Jun 24, 2010, 8:44pm Top

>211 caseydurfee:: great...now how about making it go through and fix all the ISBNs on my books that it changed behind my back :(

225PaulFoley
Jun 24, 2010, 8:47pm Top

>221 jasbro:: STM the "Initial Caps"/"Title Case" things in your word processors get it right...E. E. Cummings should be written with capitals, like everybody else. http://www.gvsu.edu/english/cummings/caps.htm

226jasbro
Jun 25, 2010, 8:28pm Top

"E.E. Cummings"? No; that must be a same name, different author ... . Sure wish we could separate them into distinct pages of their own!

227brightcopy
Jun 26, 2010, 12:31am Top

Finally built up enough of a stack to use OverCat for the first item for anything "real."

A couple of notes. First off, I found that the capitalization thing did rather bug me. Tired of having to go back and fix them. Really wish something was in place and considered just switching back to amazon instead to avoid the hassle.

Second, I was adding this book using my cuecat and it kept returning this:


search elsewhere | library searching tips

CueCat entry: 7830029512.

Search took 0.0034 seconds
Page rendered in 0.009 seconds


I finally typed in the ISBN by hand to add it.

228jjmcgaffey
Jun 26, 2010, 3:42am Top

I've found an oddity when using the green plus. It may just be that I have something other than Overcat as my default search for that? But I entered 5 or 6 books yesterday via green plus, and each one came up with nothing or 2-3 entries (mostly wrong) when I clicked the green plus and it went to the Add Books page. That was great, since I wanted to add tags and collections - did so, clicked Search, and got great Overcat results each time.

Yeah, that's it - my default is Library of Congress. And as far as I can see, there's no Overcat option on the external search. Will it be added there at some point?

229FicusFan
Jun 26, 2010, 6:23am Top

I mostly use the add page to add books like everyone else. Though in the past I Imported all my books with ISBNs.

I use Amazon quite happily and easily (with ISBN). My old non-ISBN books are already entered, and there won't be anymore, so there is no problem.

Library data is of no interest to me at all.

If my skimming of the thread is correct: there are no stats being kept about using Overcat ?
If so, I find it interesting after investing a year's worth of work.

I understand that others are excited about this but none of the proposed benefits matter to me (other than the manual additions, which could be accessed by making the plus sign actually work).

I would rather have time spent on fixing bugs and finishing other new features that were never completed.

230brightcopy
Jun 26, 2010, 11:48am Top

229> Please read #69 and #134. I understand you simply skimmed, but this really has been covered already (and re-covered).

231brightcopy
Jun 28, 2010, 10:42am Top

200/201/208/211> (about promoting secondary ISBNs to show as the primary when that is what was searched for)

Casey said this change is in, but there were some questions below that never got answered. On a related topic, I searched for 9780345484086 (the book only has the ISBN-13 listed anywhere on or in it. In amazon, the ISBN-10 is listed as 0345484088) and the overcat results were for:

9780345484109
and
9781841494784

Neither of these is the correct one for this edition. When I clicked on the "or choose from 5 alternate records from 3 libraries." link for the first result, then I got results listed in this order:
0345484088
0345484088
9780345484109
0345484088
0345484088

This just all seems a bit counter-intuitive. First off, I would have expected one of the first set of results to be the 0345484088 one. And I also would have expected it to show the ISBN-13 that I searched for, 9780345484086. I'm sure part of this is constraints on the return values the libraries are passing back. But if no, can you do something to clean that up a bit?

232AFArchuleta
Jul 5, 2010, 10:06am Top

"This will come as a disappointment to many, including us. We have long argued for library-data openness and against OCLC’s bid to privatize and monopolize library data. But we also made it clear to the libraries we search that their data will not be made available outside of the context of personal cataloging without their permission. This will not change, now or in the future."
As a future librarian, I would hope that my future profession would embrace the idea of new technologies that allow cataloging such as yours in cyberspace. I have to wonder if you have written the American Librarian Association or submitted an article on your argument and perhaps with visibility you may have the ability to change the process. I am on this website because of the course I am in and I found it quite enlightening. AFArchuleta

233kristenn
Jul 5, 2010, 10:21am Top

>232 AFArchuleta: Tim is very active on several cataloging listservs and regularly appears in ALA newsletters. (Apparently he doesn't sleep.) If cataloging is one interest of yours, the discussion lists are well worth subscribing too. The largest, Autocat, is a huge fuzzy time sink if you're not involved in the day to day nuts and bolts, but you'll get mostly discussion and new ideas on NGC4Lib. There's also RADCAT, which is pretty quiet these days. Visibility is not the issue though. Your future profession just moves very very slowly.

234caseydurfee
Jul 5, 2010, 7:18pm Top

231 >

So, what is happening here is that overcat gets back all the books that have that ISBN, and then figures out the EIDs (quasi-editions) the results fall into. Then, for each EID, it picks what it thinks is the best record for that EID and that's the primary result that you see.

What happened here was that the best record picked for the EID didn't have the ISBN that originally matched your result. Even though, it appears, it was the only book that didn't.

I think what this suggests to me is that if the term is an ISBN, it should be a straight search -- no edition collapsing.

235reading_fox
Jul 6, 2010, 9:01am Top

OK, I've been using this for a month or so now, and I take back my earlier reservations. it works as well as amazon does fro me, without amazon's issues.

I'm still not convinced it was worth the development time, but at least it has proved useful.

236EveleenM
Jul 6, 2010, 9:43am Top

I decided to re-enter all my amazon records, as I felt that in the long term I'd rather have library data for my books. I've a huge block of genre paperbacks, which are mostly UK editions, and clusters of books about natural history/botany/gardening/cookery/opera/travel guides/Ireland, plus a batch of Folio Society editions. I originally used Amazon because I was getting a low hit rate with individual library sources.

With OverCat, my top 15 sources (ignoring the earlier manual entries) were
543 Talis Union Catalog
338 British Library (powered by Talis)
321 Library of Congress
229 National Library of Scotland
83 Vlaamse Centrale Catalogus
78 Trinity College, Dublin
47 Boston College
35 University of California, CDL
27 Helsinki Metropolitan Libraries
23 ILCSO (Illinois Libraries)
18 Oxford University
15 Christchurch City Libraries
12 Washington State University
11 ACCESS Pennsylvania
11 National Library of Australia

Faced with the 600+ list of sources, most of these are ones I never would have chosen for myself. The Talis Union catalogue came out tops because of the sheer number of UK paperbacks, but the National Library of Scotland, and, surprisingly, the Flemish Vlaamse Centrale Catalogus were also very good for UK genre paperbacks. Trinity College Dublin was the obvious source for Irish publications, but Boston College did nearly as well as a specialist source for books of Irish interest. I'd heard that the University of California, Riverside, has a good SF collection, but I got more hits from the University of California, CDL.

Where OverCat really scored, though, was in the long tail: after those top 15 sources, I now have records from 56 other sources (I won't list them all!) There were only a handful of books for which I had Amazon records that didn't show up with OverCat (though in quite a few cases I was getting editions with different ISBNs which needed editing). So overall, it was a great success.

Three small irritations about library sources, compared to Amazon:
-I don't like the lower-case titles.
-I miss the hardback/paperback information that amazon includes in the publication data. Do the library sources have this in some other part of the record, which isn't imported into LibraryThing?
-I think it's a pity to add amazon covers by default, when there are good member-uploaded covers available.

237staffordcastle
Jul 6, 2010, 6:51pm Top

>236 EveleenM: I'd heard that the University of California, Riverside, has a good SF collection, but I got more hits from the University of California, CDL.

Hi, EveleenM -
University of California, CDL is the on-line catalog for ALL the UC campuses, so it includes Riverside plus all the others. UCLA also has a good SF collection, because it has the library of Bruce Pelz, a major SF fan.

238markbarnes
Jul 7, 2010, 4:12am Top

On the whole ISBN discussion:

If we search for ISBNs, then results should be returned for those ISBNs, not a best match. At the very least exact matches should be showed first.

When we search for ISBNs, LT should search for both the ISBN-10 and ISBN-13 variants. You should be able to convert from one to the other simply by calculating the correct final digit, without doing any lookups. This should happen with any search, not just overcat. I've had to give up searching by ISBN for this reason.

239MikeBriggs
Jul 7, 2010, 9:15am Top

I agree with 238. It is annoying to go to the trouble to find and type in ISBN and get "best match" through OverCat instead of exact matching ISBN. Has happened most times I've tried ISBN's, though it has been a while since I last added a book with a ISBN so that might have been fixed.

240TLCrawford
Jul 7, 2010, 9:30am Top

Maybe I am a Luddite but I do not believe that a search engine will ever get the exact book I want every time. I prefer the search engine to cast a net wide enough to catch something but selective enough not to bring back more garbage than I have time to sift through. OverCat has been doing just that. Searching LoC to often comes up empty and Amazon spews out a flood of offal.

241MikeBriggs
Jul 7, 2010, 10:00am Top

240> Sure, but if the ISBN is in there, and you put in the ISBN, you should get the exact ISBN match, not some random "best match" . . .match.

242jjwilson61
Jul 7, 2010, 11:04am Top

240> I'm not even sure what a "best" match would be for an ISBN other than an exact match. Does it return editions that match the ISBN except for one number? Anything but exact match for ISBN just doesn't make sense.

243EveleenM
Jul 7, 2010, 11:54am Top

#237 staffordcastle: thanks for the info! That certainly explains things.

#242
I think the cases where OverCat returns a different ISBN are where the library has linked the ISBNs. I noticed this was a lot more common with some libraries than others - the British Library, pretty often; the Vlaamse Catalogus, hardly ever. I've actually quite a few UK paperbacks with a different ISBN which are otherwise identical to the more common edition, cover and all (some UK publishers produce airport/export editions which differ only in release date and ISBN from the standard UK edition).

I'd like a red warning note (like the present duplicate work/duplicte ISBN ones) to say that the record entered has a different ISBN from the one searched for.

244caseydurfee
Jul 7, 2010, 1:59pm Top

242 > The best record is the one with the most number of fields we can parse out of it. Best denotes the best record for that edition, not the record that most matches your search term.

It's to correct for the fact that, in general, the most relevant record from the search engine's perspective is often the worst rather than the best record we have for that edition (the more fields a record has, the less relevant relative to the whole record your search term is. If you search on "Harry Potter", the most 'relevant' record is the short record that just has the title/author/ISBN, not also subject headings, series info, etc.)

I think that's fine for KW searching, and I think it should re-rank the results of an ISBN search based on the quality of the records. It just shouldn't combine them at the edition level.

I agree about searching for both versions of the ISBN for any data source. The problem is there isn't one central place to patch that in. I'll take a stab at it next time I dive into the adding books code, though.

Group: New features

45,201 messages

This group does not accept members.

About

This topic is not marked as primarily about any work, author or other topic.

About | Contact | Privacy/Terms | Help/FAQs | Blog | Store | APIs | TinyCat | Legacy Libraries | Early Reviewers | Common Knowledge | 134,226,741 books! | Top bar: Always visible