Lets see how this could look one day in the future:
products
products/opennlp
products/opennlp/trunk
products/opennlp/tags
products/opennlp/branches
products/corpus
products/annotation-tool
sandboxes
sandboxes/opennlp-clojure
sandboxes/dependency-parser
sandboxes/entity-disambiguation
sandboxes/stemmer
sandboxes/spellchecker
training
training/conll06-data (GPL license (?), only the trained model can be
released)
training/muc-data (should be private, only trained model can be released)
sites
sites/opennlp
We have to figure out how we can handle the training data, but that is a
different
discussion.
Jörn
On 12/2/10 10:53 PM, Benson Margulies wrote:
I have a somewhat more expansive proposal, but I'm teetering on the
mentor/contributor boundary here:
products
products/opennlp
products/opennlp/trunk
products/opennlp/tag
products/opennlp/branches
sandboxes
sites
sites/opennlp
The idea here is to allow for the possible of multiple releasable
items some day, leave a space for sandboxes, and keep the TTB
structure(s) in their own directory, not mixed with other things.
On Thu, Dec 2, 2010 at 4:47 PM, Jörn Kottmann<[email protected]> wrote:
Hello,
we have to decide how to make the initial
subversion layout.
Here is my proposal how it could be:
+-opennlp/
+-site/
+-trunk/
| +-opennlp-tools/
| +-opennlp-uima/
| +-opennlp-maxent/
+- branch
+- tags
Do we need trunk, branch and a tag dir
for the site also ? The stanbol people have it
but I cannot see why, because the checked in
version should always match the deployed version, right ?
Any opinions ?
Thanks,
Jörn
