xISBN and thingISBN compared

William Denton over at the FRBR Blog ran some interesting comparisons between xISBN and thingISBN, two services that allow you to send an ISBN and get back a list of "related" ISBNs.* Denton seems to have confirmed what we found—OCLC's xISBN has more ISBNs, but our thingISBN's paperback coverage is superior. And both can turn out better than the other for no apparent reason.

He concludes:

Upshot: If you have an ISBN in hand and want to find ISBNs of other manifestations of the same work, use both thingISBN and xISBN.

Considering that OCLC has a BILLION records, some 1,200 employees and more left-handed, green-eyed, vegetarian software engineers than LibraryThing has employees, we'll take the tie. And, of course, it's not because our engineers are smarter. It's because social collaboration is a powerful thing.

*Both have APIs; thingISBN is also available as one big take-it-and-run file.
**Someone wrote us to object that LibraryThing should not devalue the work of librarians by calling xISBN "just" an algorithm--it's an algorithm built on the painstaking work of librarians. That's fair enough, but both xISBN and thingISBN rely on that labor. (It's why LibraryThing decided, and I conjecture OCLC decided, to make the service free to libraries.) The question is what happens next, algorithm or crowdsourcing.***
***Also, it's a little known fact, but OCLC's xISBN algorithm requires a constant supply of dead kittens. When you use xISBN, a kitten dies.