On 03:31 Wed 08 Feb 2012, Alexey Shvetsov wrote:
> fbissey@... писал 2012-02-07 23:52:
> > Quoting Sébastien Fabbro <bicatali@g.o>:
> >
> >> Hi,
> >>
> >> We are contemplating the idea of switching the official science 
> >> overlay
> >> to an external repository such as github. Here are some of the
> >> advantages we would like to
> >> get:
> >> 1. easier to contribute
> >> 2. specific issue tracker
> >> 3. wiki
> >>
> > Other people have made useful comments already. The only point that I 
> > think
> > is interesting about github is the possibility to clone the overlay 
> > and have
> > pull requests.
> > Effectively that means people can contribute without ever getting 
> > infra
> > involved and an external contributor can send a pull request even if 
> > they are
> > not a member of the github science team.
> >
> > Francois
> 
> Actualy you can send pull request even now =) Its git. See for example 
> linux kernel related work

Sure, the New Yorker publishes short stories.  In my experience this is
a huge barrier for first time contributors.  I'm seeing this with the
offlineimap project which enforces the git-format-patch and mailing list
pull-requests/review.  There are people who just want to fix three lines
in the doc but don't want to get black belts in git-fu.  Then sometimes
the maintainers won't implement the 3 line change to the doc themselves
because they want proper credit for the original contributor, so after
~10 e-mails the original contributor tries git-email and fails to meet
the standards.  Another couple of e-mails are required do explain
sign-off, reply-to headers, ...  I'll stop here you get the point.

IMHO using github or a self-hosted equivalent will make contributing
easier.  Clone, commit, and to some web-thingie for the pull-request.

Cheers,
Thomas


-- 
Thomas Kahle
http://dev.gentoo.org/~tomka/