I’m on LibraryThing

I’ve finally got around to putting my books on LibraryThing, the online book cataloguing service. I, too, have a hard time reading “online book cataloguing service” without snickering, but as they go, this one gets things right. I was able to import my Amazon “I own this” list by pasting in Amazon’s HTML, and the tagging and rating interface is all ajaxy and thus fast. Here’s a random chunk of books, because I can (pardon the cut, but I don’t think this included Javascript is going to work on LiveJournal):

And if that’s not Web 2.0 enough for you, you can experience my tag cloud or author cloud instead! (Boy, I sure read a lot of dead white men. Much of that is thanks to a sociology degree, though.)

One nice thing about LibraryThing is that it has a librarian on staff, so the organizing concepts have all been thought through well. For instance, acknowledging that a book will have multiple editions and multiple countries of publication and multiple formats — and thus dozens of ISBNs — but still be the “same book”, it tracks your collection both at the book level (this particular edition and ISBN) and the “work” level (all of the published copies of the same book). And all of the metadata about the books is already there, imported from Amazon, LC, and other library catalogues. So once you enter your books (by ISBN, by CueCat scanner, by importing lists from elsewhere, or by searching titles one at a time).

Ages and ages ago, I decided that I needed book recommendations, and decided to find all of the books I own on Amazon and add them to my “I own this” list which it uses to come up with new recommendations. The results were uninspiring — its recommendations rarely strayed from “You have five books by author X, so we recommend this book by the same author” or “I see you have a few books on motorcycle safety, here are two more.” (That’s why I had a big Amazon “I own this” list to begin with.)

But that’s one thing that looks promising on LibraryThing: you can get recommendations based on your entire library from four heuristics: books owned by people who own lots of books that you own; books that have tags similar to those that you used; the most popular books on the site that you don’t have; and “special-sauce” recommendations using LibraryThing’s proprietary recommendation algorithm. All of them have their flaws (and the geek-heavy userbase doesn’t help things, either), but together I think they’ll be useful in tracking down good reads. The rating and review system won’t hurt either, and if I want to bother there’s a big community on there in its forums — sort of a “FaceBook for bibliophiles”.

The only real problem with the recommendations is that they’re based on your entire library. I have a lot of computer reference books that I don’t care much about anymore, but I have a lot of computer-related recommendations. Yes, I have half an O’Reilly Perl library; no, I don’t program in Perl beyond quick hacks anymore; no, I don’t want the other half of that library.

It’s no killer app, but for $10/year or $25/forever to list over 200 books (and free for under 200), it’s a neat toy.

Comments 8