dear sac'cers,
I asked a short questionnaire on the list at http://flossfoundations.org/
a kind of "OMG we started a foundation" support group, about
what online services people are supporting for their projects and
the volunteer/paid support tradeoff.
I got a small pile of responses which in the spirit of a shared
curiosity, I am forwarding here (the list has closed archives though it
is quiet and anyone is I think welcome to join)
Jo Walsh wrote:
- what are foundations offering (svn, trac, mail, wiki, etc)
The MoFo offers:
- CVS (Repository too large and crufty to convert to SVN, and people
don't think the incremental gain is great enough; other systems are
being looked at)
- Bugzilla
- Tinderbox for continuous builds
- Bonsai for looking at the CVS tree
- LXR
- Various MediaWiki installations (developer.mozilla.org,
wiki.mozilla.org)
- Mailing lists using mailman
- Newsgroups, in partnership with Giganews
We don't offer email services except to Foundation/Corporation employees
and Mozilla staff.
- what are projects/members actually finding useful
All of the above 
- are there real legal advantages to having one centralised repository
for different projects' source code and timestamped mail archives
What legal situation are you trying to create? Are you trying to avoid
concern about contribution provenance?
- is volunteer sysadmin and hosting 'enough', or does it make a lot of
difference if people are paid; how does that effect volunteer energy?
We have a number of paid sysadmins working for the Corporation; we
couldn't do without them.
The reasons I am asking might provide context for commentary; when the
Open Source Geospatial Foundation started back in February, We inherited
a contract with CollabNet that MapGuide, one of the charter projects,
was already using and paying for. The other OSGeo projects are all
mature and have long maintained their own cvs/svn, lists, etc, and
have been resistant to moving to a non-free platform.
Quite right too 
Gerv
Here is the data from Eclipse
I wonder if the Borg here has any commentary on "best
practise" for running hosted services for projects in their
flossfoundation;- what are foundations offering (svn, trac, mail, wiki, etc)
CVS, bugzilla, wiki, mail, news
- what are projects/members actually finding useful
Really, all of the above. Bugzilla and wiki are very heavily used.
- are there real legal advantages to having one centralised repository
for different projects' source code and timestamped mail archives
I responded to that one in a separate note. Legal stuff is not necessarily
the only advantage.
- is volunteer sysadmin and hosting 'enough', or does it make a lot of
difference if people are paid; how does that effect
volunteer energy?
Eclipse does have full-time sysadmin staff, but I realize that's not an
option for everyone. It definitely does have a negative impact on volunteer
energy.
The reasons I am asking might provide context for commentary;
when the Open Source Geospatial Foundation started back in
February, We inherited a contract with CollabNet that
MapGuide, one of the charter projects, was already using and
paying for. The other OSGeo projects are all mature and have
long maintained their own cvs/svn, lists, etc, and have been
resistant to moving to a non-free platform. The OSGeo board
is keen to take online service maintenance "in-house" next year.
In management-speak this is a "build versus buy" decision. There is lots of
literature on how to do these analyses. Obviously, community dynamics have a
role to play in the decision but doing a decent cost/benefit analysis of the
options wouldn't really be all that hard. And then you have some real data
to use as part of the decision process. If it turns out that you're going to
spend extra money to roll your own (or vice versa) it's nice to at least
know that when you're deciding, right?
Hope that helps.
Mike Milinkovich
Executive Director,
Eclipse Foundation
Hi,
Jo Walsh wrote:
- what are foundations offering (svn, trac, mail, wiki, etc)
All that (s/svn/cvs for another couple of weeks, and s/trac/bugzilla).
Plus SSH shell accounts, tinderbox environments, and hosted websites in
the gnome.org domain. And of course uploads of tarballs, and bandwidth
for downloads. And an IRC network, irc.gimp.org or irc.gnome.org
- what are projects/members actually finding useful
Bugzilla, source control, mailing lists, irc, wiki - for different
projects, different things are more or less important. It goes without
saying that regular web servers and databases are also provided by the
foundation.
- are there real legal advantages to having one centralised repository
for different projects' source code and timestamped mail archives
Not really. Copyright assignment would give a legal advantage in the
case of a court case, but having copyright spread around is a good
protection against a licence change (and also a pain in the ass if a
licence change is desired).
- is volunteer sysadmin and hosting 'enough', or does it make a lot of
difference if people are paid; how does that effect volunteer energy?
Tough question - no comment. Our sysadmins are all voluntary, and have
been great, but we've had a few things slip through the cracks for a efw
months, and a paid guy would have been nice then.
The reasons I am asking might provide context for commentary; when the
Open Source Geospatial Foundation started back in February, We inherited
a contract with CollabNet that MapGuide, one of the charter projects,
was already using and paying for. The other OSGeo projects are all
mature and have long maintained their own cvs/svn, lists, etc, and
have been resistant to moving to a non-free platform. The OSGeo board
is keen to take online service maintenance "in-house" next year.
I agree with the other OSGeo projects
Good luck with your migration.
Cheers,
Dave.
--
Dave Neary
bolsh@gimp.org
Lyon, France
Jo,
I wonder if the Borg here has any commentary on "best practise" for
running hosted services for projects in their flossfoundation;- what are foundations offering (svn, trac, mail, wiki, etc)
PostgreSQL is using Gforge and Trac in different places. My fantasy is a
merger of the feature sets of the two tools.
- what are projects/members actually finding useful
Mailing lists (more than anything), CVS/SVN, wikis, hosted homepages, file
release system, doc sharing, news. Less useful: issue tracking, forums,
surveys.
- are there real legal advantages to having one centralised repository
for different projects' source code and timestamped mail archives
No, but there are considerable new-user orientation advantages to having
one place to look.
The only legal advantage would be if you were trying to enforce contributor
agreements, in which case it would be a requirement.
- is volunteer sysadmin and hosting 'enough', or does it make a lot of
difference if people are paid; how does that effect volunteer energy?
Well, volunteers haven't been enough for us, so far ... our big "projects
merger" is now a year behind. That being said, the budget for paid staff
is likely far too high for most OSS-NPOs.
My take on this is that there are a couple of core services for
projects- code repository and mailing lists - which it may make a lot of
sense (for legal reasons) for a foundation to manage in one place, and
pay someone a retainer to support very high availability of. Then the
foundation needs to run some services for itself, as much as for its
projects - wiki, CMS for PR and tutorial documentation, homepages in
different languages for local user groups. If nesc, We'll sponsor some
open source development to get this together, and groupwise there's
quite a lot of experience with Plone and connections to plone
developers, which could lead into some neat member/group mapping
features using http://plone.org/products/primagis , etc etc. This is all
some way off, but perhaps other foundations would be interested in
sharing the love?
Well, it seems counter-productive to try to build this on Plone when you
could focus on improving Gforge or Trac instead.
--
--Josh
Josh Berkus
PostgreSQL @ Sun
San Francisco