[Bioperl-l] Pushing towards 1.0
Ewan Birney
birney@ebi.ac.uk
Sun, 16 Dec 2001 21:32:25 +0000 (GMT)
Hmmm. Hilmar - I like your style and all seems very sensible written.
One minor point - I don't think even that separating out the Factories
help real extensibility. You have to get the Factories somehow and put in
the right things for the constructors - different implementations need
different constructors eg, the DB::* "Factories" need just an id or
possibly a query whereas the in memory ones in a string - the
sequence-on-disk ones need serious collusion between the thing making it
and the object for it to make sense.
All and all - my experience is that Factories can only be unified in
classes of Factories - which is useful, but less useful than people expect
- the main thing is to make the interfaces of the resulting objects
inter-operate.
I realise that you can argue the other way.
I agree about Expression and also declaring whatever 0.9 series release as
the 1.0 alpha candidate sounds like a neat finesse for the end-of-Jan
deadline....
ewan