Back at the start of this year, IBM decided to open source its Unstructured Information Management Architecture (UIMA) and put the project on the SourceForge.net open source software repository.
Ten months later, IBM is saying goodbye to all that. It is now taking the UIMA project, which is a framework for for building search and content analytics software, to the world of open source Web server group Apache What gives? “We believe Apache is a better venue to open UIMA to collaborative development,” said David Ferrucci, PhD, a senior manager and architect for UIMA in IBM’s T.J. Watson Research Center. Technically speaking the UIMA project is now being “incubated” at Apache, which means it is not yet a full project, though it is on the path to becoming one. Ferrucci said open sourcing UIMA has been a good move. But it has other paths to tread since work began on the framework in 2004. For one, IBM is now also pursuing OASIS (Organization for the Advancement of Structure Information Standards) standardization for the UIMA specification. “Open-source has significantly increased willingness for universities, research institutions, government and life-sciences to adapt UIMA,” Ferrucci told internetnews.com. When IBM originally chose Sourceforge.net to host UIMA, Marc Andrews, IBM director of strategy and business development for content discovery told internetnews.com that though SourceForge.net is the first open source home for UIMA; it may not be the only one. The pursuit of OASIS standardization is the next step in making UIMA more widespread and helping to ensure collaborative development beyond just having it available as open source. With OASIS standardization the scope of UIMA can become broader according to Ferrucci. “We have released UIMA as a software framework that implies an architectural specification,” Ferrucci explained. “But this specification is bound-up in a specific platform implementation.” “We have now generalized and distinguished a platform-independent architecture specification.”
