[Bioperl-l] Annotation proposal

Hilmar Lapp hilmarl@yahoo.com
Tue, 10 Apr 2001 11:04:30 -0700

Matthew Pocock wrote:
> Something along the lines of a perl object with a data hash and a
> functions hash may work out - you use the AUTOLOAD method to pull out
> the apropreate function and execute it, and the function grabs the data
> hash for it's stoorage (implemented as a blessed 2-element list?). ISA
> can be over-ridden easily enough (in the instance-method case), as can
> CAN. One benefit of this is that the subs that make up an object can be
> closures so that you can create the equivalent of inner classes.

Even though I don't doubt that you can produce working code along
that route, I have to admit that I don't particularly like it. I
don't think it's the language that limits us expressing a good
model, neither in Perl nor in Java; i.e., I'm convinced that
strong typing is well compatible with a good annotation model. Of
course new types of annotation might come along that don't fit
into an existing scheme, but I'd rather extend or redesign then.

This is a bad comment in terms of BioPerl policy, because I don't
provide working code :) But I think Ewan's proposal is already a
good start.


Hilmar Lapp                              email: hilmarl@yahoo.com
GNF, San Diego, Ca. 92122                phone: +1 858 812 1757