[BioSQL-l] RE: [Bioperl-l] Bio::BioEntry

Hilmar Lapp hlapp@gnf.org
Mon, 9 Sep 2002 10:35:38 -0700


I wouldn't think so. If a piece of code expects a Bio::BioEntry, it shouldn't try to second-guess what it really is. This may be different for an adaptor that you want to cover the whole range from bare-bones to somewhere down the class hierarchy. But generally speaking, code expecting a Bio::BioEntry would just deal with it as a describable identifiable (sounds funny :) which is BTW (almost) exactly what bioentry in biosql is, not less, not more.

As a somewhat aside, to open up the old debate, if a bare-bones Bio::BioEntry is really not useful and/or not relevant anywhere, why is then there a table that reflects exactly that bare-bones object? Personally I lean towards thinking that it is useful, but there are good arguments for the opposite ...

	-hilmar

> -----Original Message-----
> From: Lincoln Stein [mailto:lstein@cshl.org]
> Sent: Monday, September 09, 2002 6:14 AM
> To: Hilmar Lapp; Ewan Birney
> Cc: Bioperl; Biosql
> Subject: Re: [Bioperl-l] Bio::BioEntry
> 
> 
> When a piece of code is handed a random Bio::BioEntry, what 
> is the approved 
> mechanism for determining what it is?  Will the approved 
> method be to ask if 
> it supports Bio::Seql vs Bio::BiblioI interfaces?  (h'mmm, I 
> don't see a 
> Bio::BiblioI....)
> 
> Lincoln
> 
> On Monday 09 September 2002 03:35 am, Hilmar Lapp wrote:
> > On Sunday, September 8, 2002, at 06:38  AM, Ewan Birney wrote:
> > > On Fri, 6 Sep 2002, Hilmar Lapp wrote:
> > >> BioEntry would be different from PrimarySeq in not having a
> > >> sequence associated with it.
> > >>
> > >> 	Bio::BioEntry is-a Bio::IdentifiableI
> > >> 	Bio::BioEntry is-a Bio::DescribableI
> > >>
> > >> and that's about it.
> > >
> > > [...]
> > > Is there any better solution. What do you have as 
> BioEntries and not as
> > > Bio::Seq's?
> >
> > All objects that are to end up as Bioentries in biosql, but aren't
> > sequences. E.g. OMIM entries, or clusters (Unigene). Of course,
> > those would have additional annotation that needed to be stored too.
> >
> > The advantage could rather be that you can inherit off a base class
> > that gives you IdentifiableI and DescribableI implemented via @ISA.
> > Not a big deal OTOH.
> >
> > I needed it already in one place in bioperl-db to store a namespace
> > in some object, so I did implement it. Right now it's in
> > Bio::BioEntry in bioperl-db. It can stay there as it may indeed only
> > be relevant in bioperl-db, or I can move it over to the bioperl-live
> > trunk. Whatever people vote (also, I'm open to renaming
> > suggestions) ... :-)
> >
> > 	-hilmar
> 
> -- 
> Lincoln Stein
> lstein@cshl.org
>