[Bioperl-l] Workbench3 internals

[David Block]

I am thinking about what I'd like for the next version of Workbench, and
one of the things is a better object/database relationship.  It's not bad
now, just repetitive, and not as easily extensible as I'd like.

So, if I'm going to abstract the database interface layer, I am thinking
of integrating it with bioperl.  Bioperl is very flat-file heavy, but in
this era, with mysql free and open, it makes sense to have easy save
capabilities to a mysql database.  I think we can do it.

So, is it a good idea?  Am I re-inventing the wheel?  And finally, how
much needs to be stored in the database?  Do I need to come up with a
blast report schema, for example?  That might be nice, but it might be
very bloated.  Schema for hsp's, complicated transcript objects, etc...

I know ensembl has gone down this road, but they haven't played for
generality.  I would like this to be a painless addition to bioperl, given
a mysql installation.  

Comments?  Feedback?  Death threats?

-- 
David Block
dblock@gene.pbi.nrc.ca
http://bioinfo.pbi.nrc.ca/dblock/wiki
Plant Biotechnology Institute
National Research Council of Canada
Saskatoon, Saskatchewan