[Biopython-dev] Bio-contrib project

[Sean Davis]

On Mon, Jun 18, 2012 at 6:48 AM, Tiago Antão <tiagoantao at gmail.com> wrote:
> On Mon, Jun 18, 2012 at 1:18 AM, Brad Chapman <chapmanb at 50mail.com> wrote:
>> For organizing additional contributions, what do you think about the
>> BioGems approach of organizing associated packages?
>
> My £0.02: I think it is a great idea. It is a way to maintain a stable
> core with a framework for additional contributions.
>
> I am probably not the only person that has code sitting for years but
> is afraid to suggest submission because I think the quality is not
> good enough. The biogems approach would be a vehicle both for new
> developers or old people with less-stable stuff. This without
> compromising the quality of the core.

At Bioconductor, we have taken a slightly different mindset, but the
end result is similar.  For a submission to make it into bioconductor,
the code undergoes a light code review (use OO practices, incorporates
standard classes and data structures rather than recreating new ones).
 Documentation MUST be present.  The contribution should add something
(may be incremental) to the codebase (though this is NOT a review of
the scientific merit or integrity of the code).  There are also some
resources (people) who help with comments on improving code, etc., so
that these loose requirements can be met by first-time contributors.
The code base still grows organically, but with some constraints.
Other details to consider are a build (and test) system, doc and test
coverage, social coding approach (or lack thereof), and maintenance
suggestions/requirements for submitters (what happens if the code is
broken--does it remain indefinitely in a contrib location?); these
details help with long-term maintenance and project integrity.

I just add these as food for thought and not as "suggestions" about
how to go forward.  I do think that formalizing the contribution
mechanism (not simply contributing to pypi, for example) is a
desirable feature of biopython so that someone can at least say "I
contributed to biopython" and folks know what that means.

Sean