Here's the detail, from the home page announcing the
The latest release of LingPipe is LingPipe 2.3.1. This is a
maintenance release replacing 2.3.0. It patches all bugs that have
been reported to us, and includes a few new features.
New Classes 2.3.1
There's an exact dictionary chunker in the <code>dict</code> package
that uses Aho-Corasick to find all chunks in linear time independent
of dictionary size.
There's an XHTML output formatter in the <code>xml</code> package. This
builds in support for most of the elements in XHTML and most of their
There's a new object to (mutable) double map in <code>util</code> that
is very similar to our object to counter map.
New Sandbox Projects
There are several new projects in the sandbox, including our
Biocreative 2 entity recognition entry, word sense disambiguation for
Senseval 3, XHTML document object model, and UIMA adapters.
This release patched a bug in keyword list extraction in the MEDLINE
parser. It also patched a major issue with thread safety in the
bounded priority queues. You'll also find some minor changes in APIs
driven from some of our applications if you look hard enough.
Xerces 2.8.1 and Lucene 2.0
This release updates dependent libraries. The major change is
the move to Lucene 2.0 in the demos. The upgrade to Xerces 2.8.1 from
2.7.1 is minor.
- Bob Carpenter