[Bioperl-l] BioPerl CPAN

Jarkko Hietaniemi jhi at iki.fi
Tue Jul 1 21:14:20 EDT 2003


> One important further question:
> 
> If individual CPAN modules are owned by its author, 
> how can others go and change them?
> 
> Most core bioperl modules are continually fixed by several people.
> The "ownership" is not a very relevant concept there.

The easiest (as in "the smallest number of moving parts") solution
would be one person to be selected as the "lead developer" for a module
and he/she would be the only one uploading new releases.

It is possible for a CPAN module owner to grant admin rights (upload)
for a module of his/hers to other CPAN authors.  But that is perhaps
not a good solution of there are really several (>3) authors working
on the same module.  (This is usually used only in the case where the
original authour is temporarily/until further notice busy, but agrees
to someone else maintaing the module for a while.) I suggest falling
back to the "one lead developer" model in the case there are really
*many* developers.

How is BioPerl developed?  Is there are a central repository (like
CVS) or is it by many independent developers and developer groups?
I'm certain if you feel like a central place would be a good place
e.g. perl.org can arrange for a CVS repository.

One Good Thing coming out of using CPAN (instead of creating a BPAN)
would also be the rt.perl.org-- a central bug database for all the
CPAN modules.

For the packaging-- the Bundle:: concept would probably be really,
really useful for you.  Take a look at
http://www.cpan.org/misc/cpan-faq.html#How_make_bundle

Also, since you will be making sweeping changes to the existing
bioperl structure, the modules 'if' and 'only' could maybe also
going to be useful:
http://search.cpan.org/author/ILYAZ/if-0.01000001/if.pm
http://search.cpan.org/author/INGY/only-0.26/only.pm

> For us it is important that anyone, at least in the core group,
> can go and make sweeping changes that affect large number of modules.

I think before you start uploading anything you really need to sit
down and hash thoroughly the new structures and dependencies.

> 	-Heikki
> 

-- 
Jarkko Hietaniemi <jhi at iki.fi> http://www.iki.fi/jhi/ "There is this special
biologist word we use for 'stable'.  It is 'dead'." -- Jack Cohen


More information about the Bioperl-l mailing list