[Bioperl-l] Auto-method caller proposal

Sendu Bala bix at sendu.me.uk
Wed Jan 10 17:00:58 UTC 2007

Sendu Bala wrote:
> Nathan (Nat) Goodman wrote:
>> Just to be clear.  Based on my understanding of your requirements, I appears
>> you don't really want the full-blown auto-generation and auto-initialization
>> capabilities of AutoClass.  My suggestion is that you look at the code in
>> AutoClass and its companion Args handler, and extract any pieces that are
>> relevant to what you're trying to do.  I'd be happy to help with this. 
> As far as I know, my proposed method already does everything I want it 
> to do, so there wouldn't be any need to 'extract any pieces' from 
> AutoClass...
>> If I've mis-understood your requirements and you do want the full AutoClass
>> capabilities, I'd be happy to assist with adapting it to fit the BioPerl
>> framework. 
>> A lot depends on whether you need the initialization to include things like
>> default values and to work correctly in the presence of inheritance.  These
>> requirements really complicate the picture.
> ...So yes, if AutoClass were to be used it would be to take advantage of 
> the whole thing.

Having looked at it a little more, you're quite right in saying I don't 
want the actual auto-classing (generation and initialization) stuff. Am 
I right in thinking the method generation is done using AUTOLOAD? I 
think we'd prefer it not be done with AUTOLOAD, so I don't think there's 
much I can borrow from Class::AutoClass. (Also, fix_keyword() doesn't 
convert --multi-word to multi_word afaict.)

I think on balance (though David Messina made a good point, I don't 
think AutoClass is the tool for this job) I'll stick with the method I 
proposed, unless there are any further comments from people?

More information about the Bioperl-l mailing list