[Open-bio-l] Re: [Bioperl-l] seq namespace method

[Ewan Birney]

Ok. Both Lincoln and Steve prefer I think the pattern


  Bio::IdentifierCollectionI


which Bio::PrimarySeqI inheriets from


gives out something like ->identifiers() -   # a set of Bio::Identifier
                         ->identifier('sp'); # identifier(s?) for this namespace


Lincoln is suggesting this handles cardinality/split/merges



I am wary of this because of two things:


   (a) Most objects come from one authority/namespace - how do we indicate 
the "main" authority/namespace of this object. There are *very few* 
objects which honestly have two equally valid Identifiers - 


    --- my view is that we are giving ourselves too much rope to hang 
ourselves with and we will endlessly have things about "assummed 
namespace" or - even worse - assumming that the first identifier in the 
list of identifiers pulled out of the colleciton is the "main" one.


   (b) We are making life harder for non-standard implementations who have 
to implement this - certainly putting in split/merge capibility as part of 
the *interface* is a bad idea - split/merge is nasty to implement.



Does anyone else want to weigh in. Steve/Lincoln - can you show me that my 
fears are unfounded?






-----------------------------------------------------------------
Ewan Birney. Mobile: +44 (0)7970 151230, Work: +44 1223 494420
<birney@ebi.ac.uk>. 
-----------------------------------------------------------------