Work multiples
I've implemented a long-time request, providing a page that shows you what "works" you have more than one of. For a long time, your fun statistics (yours|someone else's) page has provided both "Number of books" and "Number of distinct works. In fact, the calculation was subtly wrong. In the process of fixing it, I went ahead and make a work multiples page (yours/someone else's).



23 Comments:
I see (as I had suspected) that most of these "multiples" were various VOLUMES of multi-volume works. Any suggestions on how to make the system recognize that (for example) Gurdjieff's "Beelzebub's Tales to His Grandson" comes in three books, even if they might have a single ISBN for a particular 3-book edition?
Thank you! I just went and separated two books in a series that had accidentally been combined. My remaing multiples truely are - two audiobooks of books I also own, and one booke where I own two different editions.
And my number of distinct works finally looks correct!
Thank you so much! I don't know why it's such a relief to purge these accidental dupes, but it is.
Okay, so as btripp suggests it is mistakenly treating as a single "work" books which are in fact not identical, in my case this throws up http://www.librarything.com/work/686792&book=1686827 (two separate volumes of a psychology text treated as one) and http://www.librarything.com/work/7167&book=1604423 (again, two separate stories treated as one).
What (if anthing) can I do to split these out correctly?
Many thanks, Tim. Those apparent extra duplicates have been bugging me. Now the numbers are correct and I've got just the duplicates that I knew about.
This has helped me identify true duplicates as well. Thanks!
I'm sure I'll want to use it 3 mths from now, so can we have this tool put on the Joy page etc. so we don't have to source this blog to use it?
er... how exactly does one access this duplicate "thingy" other than via the blog entry?
The work multiples page should be useful - most of my duplicates are intentional, but now it'll make it easier for me to work out which ones go to BookMooch.
One thing I have wondered about for ages, is why the "You and none other" list is called that, as it makes no sense to me. Is it a typo for "You and one other", or am I missing something?
Very nifty! But there are still a few bugs in the system; the following groups obviously are not the same work:
# Carte routière et touristique: Espagne Centre-Est = Central and Eastern Spain by Michelin
Carte routière et touristique: Great Britain, Scotland by Michelin
Carte routière et touristique: Espagne Nord = Northern Spain by Michelin
Carte routière et touristique: Champagne, Ardennes by Michelin
# Christgau's Record Guide: The '80s by Robert Christgau
Christgau's Record Guide: Rock Albums of the '70s by Robert Christgau
# Collected longer poems by Hayden Carruth
Collected shorter poems, 1946-1991 by Hayden Carruth
One thing I have wondered about for ages, is why the "You and none other" list is called that, as it makes no sense to me. Is it a typo for "You and one other", or am I missing something?
It's apparently a French pun that Tim likes but nobody else gets. *sigh*
Note how Tim said multiples, not duplicates. This tool shows you when you have more than one book from the same work. Until the works system is more sophisticated, that can be multiple volumes or multiple editions or duplicates.
"You and None Other" is short for "Me, You, and None Other" so it really does apply to two people, it's just one of the two is implied, and most people don't realize this, since they're not familiar with the reference.
It's a line from something French, with obvious romantic overtones.
Note how Tim said multiples, not duplicates. This tool shows you when you have more than one book from the same work. Until the works system is more sophisticated, that can be multiple volumes or multiple editions or duplicates.
Are you sure? Becuase I have several series of books that should be coming up if this is truly the case.
For example, If this includes multi-volume works, then I should have my box set of Lord of the Rings listed with my individual paperbacks.
gavroche
I do not think that there is anything subtle or magical going on here. As far as I can tell, no new data is added and no new concepts are introduced by this tool. It is a (very useful) new presentation of what was already going on.
When you go to the work page, you see a blue box that says in red, "You have the following copies of this work:". This box will sometimes contain more than book. This new tool gives you a single listing of those cases.
If this includes multi-volume works, then I should have my box set of Lord of the Rings listed with my individual paperbacks.
Yes, you should. But not because this tool should do a better job. But because you should combine them. That is because there is no meaningful social difference within LT between owning each one and owning the set or even just owning a couple for some odd reason. Of course, I am expressing my opinion here. This is a controversial area, fundamentally because the works system is not powerful enough to represent the facts and people argue over compromises. So I would never presume to combine them, if for no other reason than that I don't go out of my way to look for trouble.
The only question to my mind is whether to list the multi-volume set as a single "book" within the work or as multiple. The best answer that came out of earlier discussions on this was single if there is a single ISBN for the box, multiple if there are separate ISBNs, and flip a coin if the books are pre-ISBN. Again speaking for myself, I list Burton's 1001 Nights as a single entry with a comment noting that it's 17 volumes. Others here list 17 separate books. Which is inevitable until the data model is more powerful. They really ought to see them listed as multiples in this new tool, but probably don't because they didn't combine them. But there really is no social difference. (Left open is the question whether there is a social difference betweeen Burton's translation and a more modern one. I would contend that there is. And whether there is a difference between that and a copy in Arabic. I would contend that there is. Or between a modern English and a modern French translation. I would contend that there is not. But I readily concede that others see it differently.)
To summarize, this tool tells you how books relate to works in a new way. You can use it to spot duplicates, to find books that are combined in the same work but ought not to be, and to find (the lack of) books that ought to be combined but are not.
Thanks for the answers, anonymous-es. :)
I notice that although you have an example of Axel's work multiples in your introduction (i.e. someone else's) there doesn't seem to be a way of doing so in practice; each attempt to investigate someone else's multiples ends up with me looking at my own sorry list of expensive duplicates, miscodings, and additional to do's.
Is this a "feature" or a "mini-buglet"?
I'm not sure where I can leave member feedback in LibraryThing so I decided to drop my comment here. Is it in your future plans to develop a database for Chinese books? I have a larger collection of Chinese books than English ones and I'm looking forward to cataloging them online. Thank you.
Re: Chinese. You can leave it here, email us (the addresses are on the blog) or post to the Talk.
We're spending a lot of time thinking about internationalization. How "far" should we go. We've arranged them in a series of steps of difficulty, with better support for Latin alphabet languages as the easiest, and a full-on Hebrew site as the hardest. (We should really do the whole site right-left.) Chinese is up there in the difficulty range. There isn't a lot of good MARC (library) data we can access, or at least we haven't found it. And unlike French or German or even Greek or Russian,nobody at LT can read it on any level at all.
That said, even if we don't make a real Chinese site, we may find a good library to tie into and allow you to catalog your books in Chinese script. (I'm sure you've discovered the LC has Chinese transliterations.)
So, thanks for the comment. I can't tell you the answer yet, but I'll post about it on the blog when we know.
Interesting idea! Is there any way for me to get the data for myself when I'm logged in, even though I'm a private user?
mmcm --
I agree it must be controversial since the first example under What Not to Combine that the instructions give is exactly this instance -- Fellowship of the Ring and Lord of the Rings.
That's why I expressed surprise when someone said that this feature should list multiple volumes -- as it expressly says on the instructions that we are not to combine multiple volumes. That could be argued - and changed, perhaps. But as of now, if someone has combined multiple volumes, they should be reseparated. And they shouldn't appear in our Work Multiples list.
Gavroche
I notice that although you have an example of Axel's work multiples in your introduction
Just click on the Axel link, and replace Axel's name with any other username.
Of course, you're right, clicking on "work multiples" on someone else's statistic page leads to your own multiple list, and not the multiple list of the profile you're looking at. So right now there's no way to go to someone else's multiple list except through the link in this blog entry, and typing over Axel's name. I'm sure that will be fixed.
This is really good, thanks!
I really should not have mixed fact and opinion in one post. Sorry.
There are books and there are works. We, the users, decide which books are combined into a given work.
You will see multiple volumes listed within a single work if someone has combined them. For example, Some Dictionary, Vol. 1 - A-M and Some Dictionary, Vol. 2 - N-Z.
You will see books with identical titles before a colon listed within a single work until someone separates them. LT ignores the rest when forming its initial guess at work combining. For example, Christgau's Record Guide: The 70s and Christgau's Record Guide: The 80s.
Post a Comment
<< Home