From: Jeff Squyres (jsquyres_at_[hidden])
Date: 2007-09-14 14:30:44
I'm moving a thread that Richard and I were having to the list so
that it can get archived / be public / etc.
On Sep 14, 2007, at 2:17 PM, Richard Friedman wrote:
>> Do we want a separate SVN for the OMPI docs themselves, or should
>> this stuff be a subdirectory off the main OMPI SVN repository?
>>
> Well, if we want to open this up to project members, we should
> follow whatever procedures you already have created for projects. I
> haven't gotten that far into the OpenMPI community to determine
> what's the best approach, but if you have already a way of
> differentiating projects, then we should look at this as just
> another project.
This probably makes sense. I think there's one minor drawback to
having a separate SVN repo for docs and several benefits:
1. Drawback: If we make it a separate project (i.e., different SVN
repository), we have to do a little integration work when we make
Open MPI tarballs. But this is probably not a huge deal since the
process to make an Open MPI tarball is already fully automated.
2. Benefit: Making a separate SVN repository would make a clean
separation between code developer committers and documentation
committers.
3. Benefit: With a separate SVN repository, the docs group can have
their own Trac (bug tracking system) and wiki, vs. sharing the Trac/
wiki of the main OMPI developer group.
The icky thing is that the OMPI group already has a heavily-populated-
but-not-public SVN repository named "ompi-docs". It's full of
academic papers written by the OMPI members (submitted to conferences
and journals and the like), etc. So we can't really use that name
for a SVN repository. ompi-documentation, perhaps?
Thoughts?
-- Jeff Squyres Cisco Systems