[Biophp-dev] The humunguous IO class etc

S Clark biophp-dev@bioinformatics.org
Mon, 28 Apr 2003 16:47:07 -0600

On Monday 28 April 2003 01:01 pm, Nico Stuurman wrote:
> I thought that Serge hoped to develop a bioinformatics library, not an
> application.  Seems to be the same goal Sean had in mind with Biophp.
> I for one hope that the two of you can merge the projects sooner rather
> than later (also see my post concerning cvs), it would be a pity to
> dilute the efforts (and confuse outsiders) by splitting it in two.

I'm certain a "middle ground" can be worked out, but there isn't
much point in immediately "throwing together" both sets of code as they
are now - the two sets of code are CURRENTLY approaching from different
(somewhat incompatible) philosophies.

GenePHP is "a library", and currently seems to be tightly connected to itself
- for example, the way the GenePHP::Parsers all return only GenePHP::seq

BioPHP is "modules", designed to be independent as possible but interoperable
(the parsers stand alone and can be used outside the context of other BioPHP
modules, for example, and they return strings or arrays than can either be
imported into a sequence object or used independently).

So, if we throw everything together as it is, pretty much everything I've
done disappears anyway, since it doesn't tightly couple into the GenePHP 

The solution/"middle ground" would mean GenePHP being somewhat redesigned
to allow for more "independent" modules to be a part of it (e.g. to pass
data more often between modules/classes as standard arrays rather than
GenePHP objects, so as to make it easier to "drop in" new components) but
I'm reluctant to ask that.  Not because I don't think it'd be a good idea
(obviously I do) but because I hate the idea of making demands that other
independent volunteer programmers rewrite their code to accomodate me. Maybe
I'm just insecure that way :-)

I DO think a less-tight coupling would make it easier to contribute modules
to the framework, but that's me.