Bioperl-guts: Re: "just sequence"

Ewan Birney
Sat, 31 Jul 1999 16:40:28 +0100 (BST)

On Fri, 30 Jul 1999, Steven E. Brenner wrote:

> Ewan,
> 1) I think I was the one who originally suggested getseq().  I now think
> it is "too clever by half" and probably best removed from the interface

;) But now we have code relying on it. This is where we need to have a
'meta' discussion about how we help do the migration to the new object.
For example, I would be strongly in favour of providing this in the new
object but throwing a warning on this method saying that in 0.07 this will
no longer be supported.

> 2) I strongly feel that the Bio::Seq should contain the identifier &
> descrption information.  95% of my sequence work requires "identifier,
> sequence, & description."  If my only options were a "just sequence" and a
> "big annotated object," I would have end up writing my own object for
> storing sequences -- or using the old (current) Bio::Seq

I tend to agree - thanks for your vote. We need to persuade ian korf about
it ...

> 3) The description should probably required to have no newlines

Nice point.

> 4) Your description should be explict about the functions doing set/get
> via number of parameters.  I interpreted your page as to suggest that
> setting was being elimiated.

The bioseqI should only guarentee gets, as some objects (eg, wrapped over
a database) cannot support sets. but different implementations might want
to support sets.

> 5) You should consider methods for "in-place" truncation & modification of
> sequences

yup. But I prefer that we did the object constructor tyoes first.

> See you at ISMB!  (Sorry I can't make it to BioPerl-99 -- it conflicts
> with my talk. :(  )

I'll guess we will see you around. Looking forward to it...

> Steven

Ewan Birney

=========== Bioperl Project Mailing List Message Footer =======
Project URL:
For info about how to (un)subscribe, where messages are archived, etc: