Amazon/Google/WorldCat

The switched-on librarians have been having a bit of a hackfest today. Jessamyn West and Andrea Mercado got to wondering how to make use of the integration between Google and OCLC WorldCat ("a worldwide union catalog created and maintained collectively by more than 9,000 member institutions.") They discovered that if you construct a Google query using site:worldcatlibraries.org plus an ISBN, you'll find WorldCat records for the book. For example, this query returns this link, which is the WorldCat record for Bruce Schneier's Secrets and Lies. If you follow that link and fill in a ZIP code (WorldCat will remember it for you), you'll see all the nearby libraries that have the book, and you can check its availability at each library.

So far so good. The next problem: how to skip the intermediate Google page and go straight to the WorldCat record? They realized that Google's "I'm Feeling Lucky" option was the ticket. Then Michael Fagan chipped in with a syntax tweak that made the idea work. Here is Andrea Mercado's post-mortem:

The coolest part of all of this is the thought and collaboration process. Jessamyn and I tooled at it for a while, and released it to the ether. It was impressive to read about everyone's problem solving processes in emails and in blog posts, and how accomplished they felt about it. Too cool.
Indeed. Before I even had a chance to receive and respond to the email that was sent my way, an ad-hoc group had worked the whole thing out.

Although I was late to this party, I found a way to add a bit more value to their Google/WorldCat combo. As noted in comments attached to this item, their solution shares a problem with the original LibraryLookup: multiple ISBNs per book. So for example, this hardcover edition (0471253111) is found, but this paperback edition (0471453803) isn't. OCLC addresses this with its xISBN service. If, for example, you feed it 0471253111 or 0471453803, you get back both ISBNs in either case.

I've had some back-and-forth with the OCLC guys about incorporating xISBN into LibraryLookup, but never pulled the trigger. The problem is that most OPACs can't query for a set of ISBNs, so you have to run a set of single-ISBN queries, and that gets clumsy.

Of course, people don't have that problem when searching interactively, because they don't tend to use ISBNs. Instead they use title or author. From a Google/WorldCat perspective, that would work like this. If you could get the the title and author automatically from a book-related Web page, you could automate this. As a matter of fact, in the case of Amazon, you can, as I showed how almost two years ago. Amazon book pages tuck the author and title info an HTML META tag; the JavaScript DOM can extract them.

I could have switched LibraryLookup over to this technique, but never did because it only works with Amazon. I've always liked the idea that LibraryLookup can also work with isbn.nu and All Consuming and other book sites. (It'd be cool if they all emulated Amazon's metadata pattern, but they don't.) Still, in the context of this excellent new WorldCat hack, I thought the non-ISBN-dependent solution might be useful. So here it is:

Amazon/Google/WorldCat bookmarklet: A/G/W (drag to linkbar)

The main lesson I'll take away from this, as always, is the explosion of creativity you can get when Web services are built in a RESTful manner, so that people get to see and tinker with the URLs.

That doesn't necessarily imply a REST-versus-SOAP death struggle, by the way. More and more people seem to be taking sides, but like Phil Wainewright, I see these as complementary aspects of the same thing.


Former URL: http://weblog.infoworld.com/udell/2004/09/23.html#a1082