News

ISBN number is structured and contains four parts of information: Group or country identifier, Publisher identifier, Title identifier and check digits. Because ISBN are allocated by blocks to publishers, all ISBNs in same block should have same publisher information. So it's possible to guess ISBN's publisher information from its structure, however this will only work well when there is a big database of ISBN-prefix and corresponding publisher name. Incidentally, there are over 19 million ISBNs in xISBN database, from this database, we can actually create a rather comprehensive ISBN-prefix database. Inspired by Publisher Name Server project in OCLC Research, in last data release we count 551,528 ISBN publisher-prefixes, and it is not very far away from ISBN agency's directory ( 880,000 prefixes) . From our ISBN prefix table, we add "guessing" service for xISBN, which tries to guess publisher information of any ISBN, even they are not covered by Worldcat, such as following:

http://xisbn.worldcat.org/webservices/xid/isbn/7806281622.js?method=hyphen

returns

{"stat":"ok",
 "list":[{
	"isbn":["7-80628-162-2"],
	"area":"China, People's Republic",
	"publisher":"San Qin chu ban she",
	"city":"Xi'an Shi"}]}

It will be interesting to see how this information can be used. On the other side, the "guessing" service is still approximate, we don't know all ISBN prefixes, and sometimes one ISBN prefix doesn't neatly identify one publisher, so use with caution.

xOCLCNUM has a less-used, but I think very useful feature of limiting search scope to a collection, we call it search in library feature. The goal is to limit FRBR expansion to a smaller scope, such as a library or special collection. Thanks for wonderful help from Jeremy York, we recently added hathitrust.org as a collection. This feature is implemented in following way: a request can put an additional parameter "library=hathi" in xOCLCNUM request, the service will only return records which marked as free access in hathitrust.org, when I gather hathitrust data last time, we collected 189,723 free access records with OCLCNUM. Similarly, we have 129,239 free access records from Open Content Alliance, and overall there are 322,629 records about ebooks (mostly free content) in xOCLCNUM service. Besides that, because of the FRBR expansion of xOCLCNUM, when a user requests an OCLCNUM, the service can lookup other OCLCNUM with free access content in same work group, for example, OCLCNUM: 51848364 is a book in copyright, by using the search in a library request, we can tell there are multiple versions of free ebooks about this book, including copies in archive.org and hathitrust.org. I just did a quick calculation, by using the FRBR expansion, 2,446,005 OCLCNUMs can link to a version of ebook. it is still a small percentage of the worldcat database, but I suspect it might be large enough for real world usage, so it would be nice to see this feature get better usage. We have a similar feature in xISBN service, but because most free access books don't have ISBNs, the xOCLCNUM's version might be better for real world usage.

I want to thank all of you who came to the first OCLC Developer's Network Hackathon at NYPL's Science, Industry and Business Library. We had good discussions about SRU, OCLC services like xID, Identities, and WorldCat Search API, and how all these services can play in Drupal, VuFind and other open source systems. See this video of the Hackathon. Also, some attenders put some pictures and info on the web including Eric Lease Morgan's summary, and our own site. And.. I want to thank Ann and Josh from NYPL for hosting - great facilities and you can't beat the location in NYC. Bruce Washburn and I had a nice long walkabout Saturday night up to Central Park and back down Park Avenue. Bruce took some pictures of his NYC trip. And I took a trip with a group of Hackathon'rs to the Empire State building on a beautiful misty night. Josh gave Alice and I a tour of the 42nd St NYPL HSSL - containing the main reading room, one of the great public spaces in the country. So my summary of our two days together -- fun, great code fragments and sample apps, great food, and opportunities to connect with library developers across North America (and two from Europe). We are planning the next Hackathon - maybe in Europe? Don

From the beginning, the promise of web services generally -- and OCLC Grid Services specifically -- has been that as more services are deployed, the more options there will be to weave them together in interesting, effective, and imaginative ways. Here at OCLC we're already beginning to benefit from that in our own services. For example, we've long known that a key piece of infrastructure for many of our web services would be a registry of all kinds of information related to institutions -- their catalog web address, the address of their OpenURL resolver, plus a pile of other things that would be needed by a variety of services we hoped to build down the road. We implemented this a while back as the WorldCat Registry. Now, with the release of the WorldCat Search API, we are starting to see how various Grid Services can be knitted together to enrich the whole. When the WorldCat Search API returns search results, depending on the response type requested, it can be sending back information pulled from the WorldCat Registry -- such as the link to a library's catalog. Therefore, keeping your institution's information accurate and up-to-date in the WorldCat Registry will populate that information out into a growing set of services in a pain-free, effective, and efficient way. Now isn't that what this should all be about?