How does one go about it?

I have a bit of time before I embark on the last half of my library and I thought it might be well spent tidying things up in my catalogue.

How does one go about cleaning ones catalogue, and if necessary, combining/uncombining and removing confusing or erroneous data from the entries? I have several hundred "unique books" and I suspect some may be duplicate titles in disguise. Obviously I know how to edit my data (pencil thingy), but I sometimes find my books confused with other books once I venture out of my own catalogue and I'm not sure how to go about resolving that. Does it matter if my data is correct for my copy as I look at it in hand, even if it's lost or miscatalogued elsewhere?

How have other people embarked on a clean-up campaign?
TIA,
cobb

My clean-up campaign
When I had my books entered (both at the 200 limit for unpaid users and again at about 500 after paying) I sorted by the Shared column at the far right and went to the author pages for each book that had no other books in the work.
On the author page I clicked on the 'combine/separate works' and searched for other titles that could be the same book as the one I had.

If your copy of a book has an ISBN and Author listed there's a good chance that clicking on the 'debris' link at the left of the book/work page will bring up other works that could be the same thing.

Nothing anyone else does is going to change your data. So if what you have entered is what you want it won't matter if it is 'correct'
e.g. My Star Trek books have all had their title's adjusted to 'ST: TOS...' ST: TNG...' etc. depending on series so they sort 'correctly'. It has required me to go in and recombine them all after I change the title but it's worth it to me to have them show up in the order I want.

Thanks for your description of how you do it, that's where I thought I might start.

I was just whiling away some time doing exactly what you proposed, but I'm still stumped to know if my book is the same work as someone else's. How would I know that with enough certainty to perform a combination? (I see how to do the combination, if I were sure it was correct.)

For example, I own two books by Cynthia Westcott, one of which is listed in my catalogue as unique (the bug book) : it may be an early - or middle - version of a work that has a couple of other possible combinations. I believe my info is correct, and my data for this item came from the LOC. But without all the books in hand, how could I know it should properly be combined with the other similar sounding titles?

Seems like there's a built-in potential for the creation of additional erorrs if anyone other perhaps the author, or a specialist in the field, attempts to combine assorted listings into one work.

cobb

I tend to click on the 'go to work' link at the end of the title portion of each work and look at book covers. The tags other people have used for the book and 'common knowledge' are also good places to look for corroboration.

If you aren't sure it is strongly recommended that you don't combine the books. It's always easier to combine than it is to uncombine cleanly.
Many people post here in Combiners! with the book title and a request for help/clarification.

As to not having the book in hand, I haven't personally bothered to try with any books I wasn't quite sure of. My collection has not yet reached the point that I can't easily get to individual books, even with what seems like half of them in boxes.

Personally, I wouldn't sort by shared users--as you combine, books will get moved in this order, which may cause you to miss books as you change pages, and some books will reappear.

Whenever I'm cleaning my library, I sort by author. That way, duplicates of the same book are listed next to each other, even if they are currently uncombined (and therefore have different #s of shared users).

After checking the debris page, I then go to that author's page, and see if there is any necessary combining there.

I sort by shared users as the first step in my cleanup process if I haven't done it in a while -- but only as a quick way to catch the really obvious, surely I can't be the only person here with Frankenstein!! problems. As collsers pointed out, it's not a stable sort if you're trying to clean the whole library.

After that first sweep, I sort by date added -- this is one of the few sorts that will remain the same no matter what I change in the records; plus, it lets me leave new acquisitions until the end.

For each book, I start with correcting any misspellings in the imported record, which is often enough to make the strays autocombine. Then, check both author and debris for possible combinations. While I'm there, combine/separate whatever else needs fixing (this is usually a matter of scanning the list and picking up a few stray copies, unless there's a case of Infernal Idiocy going on; and if there is, I'm probably working on that than catalog cleaning). As always, go carefully.

When the combining is done, I make sure every book has tags, and a cover if one is available. If I have to scan a cover, that's a separate project.

I tend to do a few dozen at a time in front of the TV, and to wander off into combining side projects like getting an author record sorted out -- I see my library cleanup as an ongoing process rather than a one-time thing. I just tag the last book I checked, and since sorting on entry date means the list won't shift around, I can pick up where I left off.

if you have the ISBN listed, then it is fairly easy to go to worldcat.org (link on right side of page) and check the editions tab to see if there are other editions of the book. I go there frequently to check on foreign language editions to find the "English" &/or original title (especially if the "non English title is quite a bit different from the English title)

That's one way to check on editions & possible revisions of a book.

My way to check & combine my books
I check author pages & see if my book/edition is listed separately from other copies.... and combine/separate as necessary.

I will also tag my series books with the series name & sometimes will put the series name in parentheses at the end of the title. Ishmael (Star Trek OS) is one example... especially when the authors in a series are different from book to book.

If when you go to worldcat, you get no results there, you can simply type in the title and author and you'll get a bunch of listings that way. You might have to wade through lots of stuff, but it is usually worth it. If it's not on Worldcat, it is really obscure.

Thank you all for taking the time to help me understand how you go about cleaning your catalogues.

I'm sort of doing that by starting with the works which I appear to be the only member to have catalogued.

On one hand I can see a modest communal gain to having the books accurately organized into what LT calls "works", but it seems the risks of doing so incorrectly are significant, too.

For now I think I'll just focus my efforts on tidying my own catalogue and trying to figure out an exportable way to capture (consistently) the extent info.

Cheers!

cobb

choose Library of Congress as your first choice... amazon "appears" to have the greatest number of errors in their data.

Enjoy!

To export your own books, look under the tools tab.

>10 skittles: I know everyone touts the LOC as the best source, but I find it often error-filled, especially for older books. Not as bad as Amazon, but still shockingly bad. The best source is manual entry, I think and I'm surprised that it seems to get short shrift from so many users. I'm contemplating redoing my entire cat. by hand but waiting for long-promised changes in user-controlled field possibilities which would allow me to record the info I want and then take it off to my own hard drive.

>11 mvrdrk: I know how to use the export tools, but apparently not all the info visible on my screen is exportable (the green v black type thing). I think this may be unresolvable short of some kind of screen save/scrape thing. And even if I had the technical capacity to do that, there may be copyright or proprietary data issues.

Thanks for your suggestions and help!

cobb

>12 bcobb:

I've been wanting to reenter my whole library too--make certain that all of the data is exactly correct, and using a more systematic tag system than I did the first time around. It's just a question of finding the time...

>12 bcobb:

I agree about the LOC! I gave up on it when I saw that practically ALL the foreign editions had wrong languages entered in the primary and original language fields. Pitiful.

Language data is often wrong... that doesn't mean that the rest of it is bad... :/

True enough, but I can't consider that as a respectable source, nor am I motivated to trust any of its data. Even if some poor LOC librarian from the antediluvian era couldn't, say, identify the original language by any means, that's no reason to put "English" in the field.

>16 LolaWalser:

Library of Congress has a
form to report catalog and authority record errors.

>16 LolaWalser: Thanks for the link. If I do some serious catalogue cleaning or re-entering I'll keep that page running in the background and report what I see.

I am a bit relieved that others see the LOC data as less than perfect, too. For long time it seemed that here on LT it was being held up as a paragon, but I was shocked to discovered how error-filled it was. I agree that relatively speaking it is better than Amazon.

(Amazon, however, has more of my junky, anodyne, supermarket novels.)

But I'm still hesitant to do any combining other than discovering, and correcting if neccessary, where my own books are falsely singletons because of errors in my data. What I don't get is how anybody could feel confident enough about other people's data to go about combing without everybody's books piled on their work tables.

In one of my examples above (e.g. Cynthia Westcott) above, how could you be certain that the various editions of the garden bug book were substantially enough the same to be one "work" as we define it on LT? I have two copies of her garden bug book, published in separate editions more than 20 years apart and I don't feel at all tempted to consider them the same work, just books on the same topic by the same author. Maybe this is more of a dilemma with non-fiction, than with fiction?

Thanks again for your answers to my queries.

cobb

#12: I know how to use the export tools, but apparently not all the info visible on my screen is exportable (the green v black type thing).

Anything in green isn't actually in your catalog, which is why it isn't exported. It's not your data. (And it's quite possible for what's displayed in green to change over time.)

If you want to, for non-subject fields you can add the data to your catalog by editing the field (and if necessary, reverting your edits after saving—this is the only way to get exactly the same data as the green autofill, by supplying something different, even with just a "_" added to the end, then changing back to the desired data).

It's slightly messy and complicated, I'm afraid, but unfortunately if you don't edit the data there's no way for the site to know that you do indeed want that field stored in your catalog. It thinks that you left the field unedited.

#18: Maybe this is more of a dilemma with non-fiction, than with fiction?

Definitely. Although fiction books often have multiple editions, very rarely is there any change in content beyond minor typo fixes. And for those cases where there are significant alterations to the text (like The Stand vs. The stand: the complete edition) it's usually a fairly well-known text.

this is the only way to get exactly the same data as the green autofill, by supplying something different, even with just a "_" added to the end, then changing back to the desired data

If you do it from your catalog view, rather than the edit page, you can just double-click the appropriate cell and then click save.

#20: Ah. That would, uh, be much easier. ^^;;

#18

Cynthia Westcott's four editions (1946-56-64-73) of The Gardener's Bug Book and six editions (1950-60-71-79-90-01) of the Plant Disease handbook are lumped together in Worldcat, so I would tend to follow suit, as it gets too complicated to have everything extensively split apart. I went into Common Knowledge and put in this list of editions.

The solution is to regard Westcott's books as something like a serial (yearbooks, almanacs, etc), and then to have a subcategory under the basic title, where editions can be noted and notated. LT, although it handles series, does not handle serials as yet.

>16 LolaWalser:

The reason the data about language is wrong is because of the way it had to be coded until a few years ago: as one big long string of language codes. Nowadays the data can be coded in separate little categories to say "this is the language the book was originally in", etc.