A Definitive URI for Books?

January 4th, 2003

Books, Conversations and Semantic Web

I like books. Alot. And I think the conversations about books that have arisen on the web are cool. Projects like Bookwatch and All Consuming turn the proliferation of personal reading lists, into a distributed web of collaborative filtering.

But I’ve got a problem. All of these systems require that a book have a uniquely identifying URI (its a sneaky back door application of the semantic web), and the ad-hoc standard that has arisen is Amazon. There are a number of compelling reasons to link to Amazon: decent information, community and proffessional reviews, and of the whole associates thing. And there are reasons not to link to Amazon, primarily I don’t want people buying their books from Amazon but from their local bookstores! Also you are: bombarded with ads for People’s magazine, Epson printers, and clean underwear, tracked and indexed, and have handed over the keys to controlling conversations about books to a single corporation who aspire to be just like Walmart, not the what I look for in a Muse.

Alternatives: In Search of a Definitive URI

Now books do have a unique identifier, their ISBN, so at some level we can be linking them across various URI schemes, and while not ideal it might be a place to start. So what else can we use?

Booksense is a possibility. Booksense is a great idea, its a coalition of independent bookstores that share reccomendations, book certificates, and a website. Unfortunately their website sucks. If I want to link to Dubious Hills by Pamela Dean, a book which is out of print, and no one in my zip code happens to be carrying it, I can’t. Booksense only display information about books in your area in a misbegotten idea to drive people locally. I’ve tried several times to talk to them about this failure of their website, that you don’t get anywhere by slamming shut the door in peoples face, but they are un-interested in listening. That and they expose the implementation of the site in their URLs (booksense.com/product/info.jsp?isbn=$ISBN) which does not reccomend the the longevity of their URIs.

The Library of Congress has potential with a catchy, short URL, besides being logical. But is even worse. Its slow, confusing, and does not, as far as I can tell, I expose any sort of permanent URL.

There have been interesting expirements with setting up book sites based on trackbacks, like the Book Review repository, but I don’t believe these have the potential to be the definitive URIs, more likely they will be consumers of such.

So I’m kind of out of ideas? Anybody else have a reccomendation?

In the short term it would be nice if people writing book crawlers supported Booksense as well, but I understand why they wouldn’t as Booksense is really shooting themselves in the foot. (I mentioned this idea to Bookwatch at one point, not sure what became of it)

I’ve asked Aidan to look into how much it costs to get a digital copy of books in print, but I imagine its prohibitively expensive, and not something that is feasible for a personal hobby project.

Closing Thoughts

The one positive idea I’ve thought about is, if we can come up with something good, then it would be very simple to get great penetration by adding support for it as a macro in Moveable Type, e.g. ISBN:$ISBN and auto-construct the URL.

And a final qualifier, if you see me linking to Amazon in the future, know that it is merely a recognition of their dominant position in the market place as the only decent provider of book meta-data, and not a reccomendation to buy from them.

update: isbn.nu could be a good book namespace, and allcosuming does support booksense, and more. see comments for details.

Tagged: Uncategorized , , ,

3 responses to “A Definitive URI for Books?”

  1. John Brugge says:

    I realize this is an old post, but I have had a similar feeling about book links. I’ve occasionally used isbndb.com as an end source, which isn’t that pretty but is functional.

    There is also WorldCat, which seems to have pretty standard URLs: http://worldcat.org/isbn/nnnnnn. Not necessarily a lot of metadata about the book, but great if you want to give people a non-commercial place to get the book.

    For strictly personal use, have you tried BookBurro? I don’t actually use it as much as I thought I would, but it comes in very handy when I find a book online and want to get it from my library.

    Good luck, John

  2. […] Woot! Something I’ve wanted for forever. […]

  3. George Oates says:

    Hello! I found out today that, although ISBNs are certainly alluring, not every book has one, and sometimes the same ISBN can be shared by two different books. I’ll definitely poke around your references here – it’s all about librarian soul-searching at the moment 😉

    On a personal note, feels great to have the prospect of productivity again :)