[BioPython] seq objects etc...

Andrew Dalke dalke@bioreason.com
Mon, 13 Sep 1999 12:12:05 -0600

Brad, replying to Ewan's comments:
> > AnnSeq has-a Seq object.
> I like that model a lot.  How does that effect LSR compliance?

The thing is, it doesn't make much of a difference.  The basic
classes don't have to be the ones given by the LSR; you just need
a way to construct them from what you have.

Thus, we could have a light-weight sequence object and use it to
compose something which meets the LSR requirements.

And BTW, I'm more in favor of even lighter weight objects than
Ewan; that is, I would rather not have id, accession and desc as
part of the core object.  To me that is more related to a sequence
record data type.