[Bioperl-l] Bio::Ontology::Term (rollback question)

Hilmar Lapp hlapp at gmx.net
Mon Aug 27 19:07:12 EDT 2007


The B::O::TermI interface actually says that get_dblinks() would  
return scalars. That's why the add_dblink methods accept strings. I  
also agree that this is inconsistent with with the rest of BioPerl.

Oddly enough, Term::add_dblink_context() does ask for DBLink objects,  
though it doesn't seem to be enforced, even though  
Term::get_dblink_context() is advertised as returning scalars.

So it does seem this is messed up design-wise. It seems to me that to  
really fix this would inevitably break the API, and I don't see how  
you would make this backwards compatible w/o creating a lot of messy  
code, the sole purpose of which would be backwards compatibility.

One could only fix Term::add_dblink_context() as it's not in the  
interface but that wouldn't contribute anything to improving  
consistency.

So the alternative to breaking the API in a non-backwards compatible  
fashion would be to add to it, map the existing dblink methods onto  
the added ones, and start deprecating them. For example, you could  
add methods get_dbxrefs() (also on the interface), add_dbxref(),  
etc,   and build in a context argument so we don't need another set  
of methods for that. They would accept and return DBLink objects, and  
the get_dblink() methods could be changed to map those to scalars  
while also getting slated for deprecation.

Does this make sense?

	-hilmar

On Aug 27, 2007, at 6:31 PM, Chris Fields wrote:

> This is related to the ongoing Feature/Annotation rollback.  I have
> found that a few Ontology-related modules are (either directly or
> indirectly) passing strings instead of Bio::Annotation::DBLinks to
> Bio::Ontology::Term::new(), add_dblink(), or add_dblink_context()
> (thelast is where the error occurs).
>
> If needed we could allow strings to be passed but this isn't
> consistent with the API.  Any thoughts on what to do here?
>
> chris
> _______________________________________________
> Bioperl-l mailing list
> Bioperl-l at lists.open-bio.org
> http://lists.open-bio.org/mailman/listinfo/bioperl-l

-- 
===========================================================
: Hilmar Lapp  -:-  Durham, NC  -:-  hlapp at gmx dot net :
===========================================================







More information about the Bioperl-l mailing list