Today I'm introducing a new feature called The Screening Room. It will be a monthly series of screencasts exploring new software that strikes me as both interesting and important. I expect these screencasts to run 15 or 20 minutes -- enough time to go fairly deep into each subject, but highly compressed relative to the several hours of raw interview and demonstration material that I'll have captured for each one.
The topic for today's episode is IBM's UIMA (Unstructured Information Management Architecture) SDK. I first mentioned UIMA back in August, when Irving Wladawsky-Berger announced that IBM would be donating the software to the open source community.
More recently, I met with IBM Software Chief Technologist Alfred Spector who, among other interests, is very jazzed about UIMA. In our discussion I noted that in addition to the Java and C++ bindings and Eclipse integration that's currently available, it would be very cool to repackage this stuff for the Web 2.0 crowd. AJAX, scripting, and lightweight service orchestration aren't just buzzwords du jour. They're also ways to spark the kind of collaborative ferment that IBM hopes will surround UIMA. He liked the idea, so we'll see what happens.
Meanwhile the Java kit is available now, is nicely adapted for Eclipse, and can be used to build on other open source components like OpenNLP, the natural language processing library that's used with UIMA in the screencast.
The presenters for today's episode are Dave Ferrucci, UIMA's lead architect, and Eric Brown, manager for Unstructured Information Management Systems, both from IBM Research.
Former URL: http://weblog.infoworld.com/udell/2006/01/31.html#a1379