I'm guessing that files labeled as .ann and .con are "converted" to end in ".xml" - and that these are standard UIMA descriptors, is that correct?
The .ann or .con files contain standard uima descriptor files, bute they are not converted to .xml they are accessed with the eclipse resource api and then passed on to uima. The file ending is used to identify the type of the descriptor. The goal was that the Cas Editor can detect annotator and consumer descriptor files. This is necessary to reference them later e.g. in the ui actions to run them against a document.
If there are good reasons to have a different convention for these kinds of things in the CasEditor, it would probably be good to document the rationale - to help motivate the reader to get past any initial confusion about this.
I can add this to the documentation. Jörn
