[Bioperl-l] Packaging bioperl for Fedora

Alex Lancaster alexl at users.sourceforge.net
Sat Mar 31 00:55:31 EDT 2007


>>>>> "AD" == Allen Day  writes:

AD> The majority of the Bioperl classes are file parsers, or
AD> manipulate data that comes from the file parsers.  Yes there are
AD> exceptions like the Eutils and Ensembl-intefacing classes, but
AD> they are the minority.  The types of files that are worked with
AD> are generally either A) primary data sets such as genome data, or
AD> B) derivative data, such as sequence alignments that are derived
AD> from primary data using an algorithm.

AD> If we're in agreement that the primary data sets and
AD> libraries/applications for producing derivative data should not be
AD> present in Fedora Extras, then it follows that the Bioperl classes
AD> for manipulating these primary and derivative data should also not
AD> be present in Fedora Extras as they are of little use without data
AD> to manipulate.

That's not entirely true: I'm using Bioperl (just the "live" package
alone) to do some data analysis since I'm not using any of these
genomes.  Some data sets won't come from these large genome databases,
but maybe local population datasets produced inhouse, for these kinds
of analyses it's sufficient to just have bioperl-live.

I think it's still workable if a subset of bioperl lives in Fedora
space ("Extras" wil be no more as of F7).  As an example of a split
that works: look at how Fedora/Livna handles the media player xine:
xine-lib lives in Fedora but various non-free plugins live in Livna.
I agree that this may not work in our case, but it's worth thinking
about.

My sense is that if the license of a package is OK with Fedora and it
doesn't otherwise break up a group of packages too much (e.g. breaking
up Bioperl may or may not be a good idea), then it could go into
Fedora.  (For example I believe that the NCBI C++ toolkit's license
should be just fine being a work of the US government is in the public
domain[1], so it could go in Fedora, same for BLAST[2], which is has
an implementation within the toolkit).

Is anybody in Biopackages.net interested in packaging biopython there?
Given that whether or not Bioperl lives (partially) in Fedora is still
under discussion, it might make sense for me to suspend work on the
Bioperl stuff (although not the CPAN Perl deps) and work on a
biopython package for Fedora for the moment until we decide how to
proceed with Bioperl.

Alex

[1] http://www.ncbi.nlm.nih.gov/IEB/ToolBox/C_DOC/lxr/source/corelib/README
[2] http://www.ncbi.nlm.nih.gov/blast/developer.shtml
-- 
Alex Lancaster, Ph.D. | Ecology & Evolutionary Biology, University of Arizona




More information about the Bioperl-l mailing list