[Bioperl-l] GenBankParser comparison to bioperl parser

[Jason Stajich]

On Thu, 12 Sep 2002, Ian Korf wrote:

> I'll de-lurk for this.
>
> I, for one, think that a bio-util-perl project would be a good
> idea. Not everyone wants or needs the whole of bioperl.
> Sometimes, a small utility outside the framework is useful.
> Analogously, I use a full-featured text editor that counts
> letters, words, paragraphs etc, but I still find myself using wc
> from time to time.
>

This seems like a fine idea by me.  Once we're sure servers are back
online I can create the necessary CVS repository.  I'd like to spread it
to a separate repository mainly for the sake of keeping permissions
somewhat protected in the bioperl core code.  This way we can fairly
liberal about who we might give contrib access to while restricting
bioperl core code a little more.  This is of course only to aid the
ever-present battle to clean up code and keep code/docs/etc consistent.

> I think that if something is going to be part of bioperl, it
> ought to have the same consistent interface. But I like what
> John has done, both at the interface and implementation level.
> Is it possible to leave John's parser the way it is and offer a
> bioperl interface to it?
>

I have no problem with this - I think it might take some work to figure
out how we might do this efficiently and still hiding things under the
guise of the same Bio::SeqI interface, but it seems possible.  I would
certainly support someone working on this.

John don't think all our defending of the way things are now means we are
uncomfortable chucking some of the current code for a better
implementation.  Things have just built up in a way that makes sense for
the data model we've been building for the systems that interact on top of
this code so we're loathe to change the interfaces very much or introduce
changes that break things.  But we're not opposed to making things faster!

Later this week or this weekend we'll get the repository setup and look
into getting people access.  I'd be excited to see Boulder in there as
well as any other assorted projects which people might want to poke with
and improve.

-jason


> -Ian
>
> On Thursday, September 12, 2002, at 09:08 AM, Elia Stupka wrote:
>
> >> As Jason says we are both aware of the performance problems in
> >> the bioperl
> >> parsers, and have some suspicision of where they are. In fact,
> >> I am not
> >> 100% sure if my speed fixes went into the main trunk or the branch.
> >
> > Hello guys, I've been following this discussion and I am not
> > sure what the
> > conclusion is, but it looks to me like you are saying "yes your
> > parser is
> > faster but we know why ours is slow so when we can we will try
> > to get your
> > speeds", without taking the code in as part of bioperl.
> >
> > Doesn't that go against "who codes it wins it"? I mean if somebody is
> > interested in a non-bio-object-centric way of parsing that is
> > much faster,
> > doesn't it make sense to take it in as a worthy contribution to
> > bioperl?
> >
> > What it seems to suggest in the long run is the setting up of a
> > separate
> > bio-util-perl project where useful but not wonderfully integrated stuff
> > gets developed, and surely that's not what we want to see happening?
> >
> > Basically I am suggesting that the parser should be commited in
> > bioperl. This will make it easier for all of us to assess it
> > and also to
> > reuse concepts to improve our own parsers, and then once our
> > parsers are
> > so good and if nobody shouts they love the other one, we can
> > always remove
> > it.
> >
> > My two stirrin-the-waters cents,
> >
> > Elia
> >
> > ********************************
> > * http://www.fugu-sg.org/~elia *
> > * tel:    +65 6874 1467        *
> > * mobile: +65 9030 7613        *
> > * fax:    +65 6779 1117        *
> > ********************************
> >
> >
> >
> > _______________________________________________
> > Bioperl-l mailing list
> > Bioperl-l@bioperl.org
> > http://bioperl.org/mailman/listinfo/bioperl-l
> >
>
> _______________________________________________
> Bioperl-l mailing list
> Bioperl-l@bioperl.org
> http://bioperl.org/mailman/listinfo/bioperl-l
>

-- 
Jason Stajich
Duke University
jason at cgt.mc.duke.edu