[Bioperl-l] 0.7 release: Blast

J.C. Diggans jdiggans@genelogic.com
Tue, 14 Nov 2000 17:35:26 -0500


> What do we do -
> 
>         - battle on with blast modules?
> 
>         - fork bplite?
> 
>         - rewrite?

I just took a look at the code for Blast.pm (for the first time) and
you're right, 'ouch' was all that came to mind. The Blast modules are
quite useful and have by themselves saved me many hours over the past
year or so. I cannot imagine a good sequence manipulation package that
didn't involve some sort of interface to Blast but avoiding the
ever-dreaded feature bloat is a problem here. Everyone Blasts in his or
her own way and wants Blast.pm to handle that peculiar way. Perhaps the
goal would be to standardize to a common, easy-to-handle-and-extend
Bioperl Way (e.g. sticking to XML as mentioned earlier, which I think is
an excellent suggestion and keeps bioperl open to interaction with the
other Bio*s) and then put a little more burden on individual users to
create translations between their widely diverse methods and a root
bioperl method. If the module tries to fit everyone's preference, it'll
end up in the same feature-bloated unmaintable place.

Considering its innards are antiquated, I'd vote for a re-write. Forking
bplite would just lead to the feature bloat from which Blast.pm now
suffers and battling on w/ the current Blast.pm would just lead to
severe migraines (w/ apologies to Steve C.). I'd volunteer to help out
where I can, but it's definitely not a one-man job.

- jc

-------------------------------------------------
James Diggans              Phone:    301.987.1756
Gene Logic, Inc.           FAX:      301.987.1701
jdiggans@genelogic.com     Cell:     301.908.2477
-------------------------------------------------