[Bioperl-l] Bio::FeatureHolderI interface confusion

Peter Wilkinson pwilk at videotron.ca
Tue Jun 17 22:13:25 EDT 2003

> > Otherwise I'm having difficulty seeing the point.
>Probably not many people do

Well, for fun I was looking into making the fasta_ncbi object someone 
suggested. As I got into perusing the interfaces , I got a little lost. I 
though it would be good to put the class hierarchy to memory.

I appreciate Chris' point, and I think the interfaces are at least edging 
towards spaghetti with continued bioperl growth.

The problem with OO is the interpretation of what OO metaphor means. And I 
think strict adherence to OO programming can be its own undoing.  1 
particular OO model is not adequate for all life's problems.

A moderate position would be to have a multi-paradigm approach where code 
is constructed appropriately to get a specific job done, built for 
extending, or built with strict rules in mind.

In many cases code is a lot faster and efficient top-down. In many cases 
the OO model is not needed, in other cases its an elegant solution.

Is it possible to refactor bioperl with sensible multi-paradigm divisions? 
and by multi I mean only a few, say 3 or 4?  Refactoring could then be 
limited to what is appropriate.

Peter W.

More information about the Bioperl-l mailing list