[BioRuby] BioRuby passes tests (kinda)!

[Pjotr Prins]

On Sun, Aug 05, 2012 at 04:17:38PM +0900, Naohisa GOTO wrote:
> On Sat, 4 Aug 2012 12:42:21 -0400
> Clayton Wheeler <cswh at umich.edu> wrote:
> 
> > Also, I've seen quite a few failures involving PhyloXML (particularly
> > on JRuby); now that I've repackaged that as a bio-phyloxml gem
> > (https://github.com/csw/bioruby-phyloxml, which is passing its tests
> > on Travis-CI), we could think about removing it from the main BioRuby
> > repository.
> 
> I'm againt to remove PhyloXML parts from the main repository
> only because the tests show many failures on JRuby. It runs
> without errors and errors on MRI (1.8.7, 1.9.2, 1.9.3) and
> Rubinius (1.8 mode).

I agree the failing tests are not a reason for removing phyloxml. One
thing we should discuss, however, is removing code from BioRuby when
functional replacements exist, as it the case with bioruby-phyloxml.

It makes little sense to maintain two source trees (even if they are
phylo trees).

Also, we may strip out parts of BioRuby which are not used anymore, or
not-so-core, and push those out into biogems. That way BioRuby becomes
'lighter', easier to maintain and modify. That discussion we will have
in Japan at the Biohackaton. Perhaps on the way to BioRuby 2.0, and
even 3.0.

Personally I am also developing ideas on what BioRuby should be, and
what it should do. Currently I see the main use of the Bio* projects
as being parsers and outputters of the pletora of Bioinformatics of
data formats. Maybe we should stick to that, support any known input
type, and generalized output, where output should at least support
tabular output, JSON, XML, and RDF.

All other stuff should go into biogems.

Pj.