Hi Folks,

The University Library at UNC-Chapel Hill has created an OCA API.  We have 
harvested (and continue to harvest) standard bibliographic identifiers and 
link them to OCA identifiers.  The API is deliberately modeled after 
Google's for ease of implementation.

Here is a subjec search in UNC's catalog for "North Carolina" limited to 
the 19th century.

http://search.lib.unc.edu/search?Ntk=Subject&Ne=2+200043+206475+206590+11&N=206596&Ntt=north%20carolina

You will see links to OCA as well as Google.  (The full record has an OCA 
icon if you want to look.)  Right now we are only banging against the API 
with OCLC numbers, but ISSNs, ISBNs and LC numbers are in there.

We are looking for a couple of partners to work with to take use beyond 
our local OPAC.  You would be ideal if: you are interested, you already 
use the Google API, you have a significant corpus of pre-1923 works in 
your catalog.

As the Google API is familiar to many of you, it would be easy to figure 
out how to implement UNC's without working with us.  Please hold off until 
we are ready to open it up all the way? This is why we've not yet put up 
documentation.

Caveats and other notes (feel free to skip):

*We realize that Open Library has an API, but we had already gone a goodly 
distance and we are finding relatively meaningful differences in coverage 
and utility.

*We collect the data from OCA as it comes in (the data should be up to 
date within a half hour or so)...but they occasionally have need to 
correct/remove works.  Right now we are actively working on this issue, 
but do not yet have a great mechanism to pull deletes and update corrected 
identifiers.

*The data is only as good as the data we harvest.  There are a small 
number of bad links.  See above.

*Excerpt from a developer on UNC's holdings (we are an OCA Scribe site):

> ...I decided to run the same script against the [production] database as 
> well to see how much the matching is changing over time with continual 
> updates:
> - 429311 OCLC's tested
> - 72350 matched
> - 2599 of the matches were scanned by UNC
>
> So that's 808 new matches since the end of March, not too bad for one 
> month.

Effectively we are now linking to ~72 K digitized works that we were not 
previously able to provide (though as Google digitized books are being 
added to OCA, there is significant overlap).

*When we do open it up it is the API we are offering, we are not prepared 
to be crawled for data.  If you want the data, get in touch and we will 
see what we can do.

If you are interested in being an early partner, please drop me a line and 
I will be in touch.

Tim

+++++++++++++++++++++++++++++++++++++++++++
Tim Shearer

Web Development Coordinator
The University Library
University of North Carolina at Chapel Hill
[log in to unmask]
919-962-1288
+++++++++++++++++++++++++++++++++++++++++++