Bio::Seq / Steve's comments

[Steven E. Brenner]

> I'm planning to translate steve's notes into a specific list of changes to
> make to Seq.pm and I'll post the list (along with some questions) to the
> mailing list.

Sounds great.

> Making a separate Readseq.pm module seems to be fairly high up on the
> priority list and would allow me to clean up parts of Seq.pm. I'd be
> willing to write a simple version that uses temporary files instead of
> bidirectional piping. The one part I would need guidance on would be the
> interface -- I'm too new to OO/Module work to trust myself with designing
> the most functional set of methods.

Well, I have to say that I am not terribly familiar with 'dynamic loading'
modules either.   (Basically, we only want to call on ReadSeq if needed;
and we want to fail cleanly if it isn't located.)   I think that looking
at examples will be the best way forward.

I would consider coming up with a good interface for the Bio::Seq::Parse
(see below) and Bio::SeqList (see previous mails; Georg: I'm going with
your suggestion for name, though I still prefer Set) more crucial
and hopefully easier than that for ReadSeq

> On a similar note, Brian Osborne bas noted that he uses the seqio package
> for parsing because ReadSeq doesn't handle large genomic sequences (>
> 100000 bp) that well - If we come up with a good interface for Readseq.pm
> it would be easiy to create a seqio.pm as well.

Exactly!  Good thinking.


> Would those go in Bio::Parse or Bio::Util or something? I really am quite
> new at dealing with modules and such :)

I thought of Bio::Seq::Core or maybe Bio::Seq::Parse.  Do these seem
reasonable?

Bio::Seq::Util suggests end-user utilities, and therefore probably isn't a
name we want to use.

Regards,
 Steve