I plan to create a new branch to do some experiments in support of https://cwiki.apache.org/confluence/display/UIMA/UIMAJ+v3+PEAR+Support, in particular migration tooling, since the current top idea separates JCas customization from the generated files. I'll call it uimaj/branches/experiment-v3-jcas.
Some thoughts on migration tooling support. It would be good to have a tool that could be turned on in an existing pipeline, and have it analyze and dump out the current JCas cover class source code, categorized as follows: a) standard (no customization) classes b) For customized classes, just the customized part. c) (eventually) suggested source code for the customized part This would be driven by a "decompiler" (to generate the pseudo source code) and some simple (probably regex expressions) pattern matching that scans the generated sources. The decompiler I'm thinking of using is an Apache V2 licensed thing: Procyon (available via maven mechanism). -Marshall
