Test suite coordination

I have several test cases (valid grammar, invalid grammar, valid
documents, and invalid documents) which is named according to the
following rule:

valid grammars:   XXXXXXXX.rng
invalid grammrs:  XXXXXXXX.e.rng
valid documents:  XXXXXXXX.vYYYYY.xml
invalid documents:XXXXXXXX.nYYYYY.xml

For example, "foo.v1.xml" is a valid instance for the grammar "foo.rng".
How is this naming convention?


My proposal is to set up a project in SourceForge, whose name would be
"relaxng", which is still available, and collect test suites under that
project. I also think that project might be a good place to host the
datatype interface.

Is there anyone who has experience of setting up a project in
SourceForge? and willing to set up one for us?


P.S. is it possible to release the datatype interface as a committee
work product?

--
Kohsuke KAWAGUCHI                          +1 650 786 0721
Sun Microsystems