[Bioperl-l] RootI detachment proposal.

Hilmar Lapp lapp@gnf.org
Fri, 12 Jan 2001 14:32:11 -0800

Ewan Birney wrote:
> [Ewan recovers from rereading the Bio::Root:: stuff...]
> This is *mainly* for Jason and Hilmar, but in case there are other
> people who want to chip in:
> I want to completely detach RootI from the other Root::Objects (in
> particular Err). This means a heavy refactoring of RootI - mainly in
> removing the code.
> I will keep ->throw and ->warn but not ->verbose as a real method. (jason
> - do you mind this?) (I will have a "deprecation warning" on verbose)
> I am planning to do this on my local copy now and see how it pans out...
> Bio::Root::Object in it's full glory will still be there for modules we
> have not migrated to Bio::Root::RootI
> thoughts anyone?

verbose() is being made use of heavily as far as I saw some code and code
migrations from Jason. I do think that it is beneficial and desirable to
have a central mechanism for regulating 'verbosity' (e.g., what happens
upon a warning being issued). I also don't see yet why having verbose() in
RootI hampers disentangling RootI from the other objects, or where this
should interfere. (People who don't want that feature simply override it
with a stub.)

Maybe I'm missing something. Ideally I don't have to come up with a
SeqIO-specific mechanism concerning client-side regulation of the severity
of warnings.


Hilmar Lapp                            email: lapp@gnf.org
GNF, San Diego, Ca. 92121              phone: +1-858-812-1757