jason at bioperl.org
Tue Nov 7 13:27:27 EST 2006
i think it is important - we just have not had a dedicated developer
who enforced these requirements by auditing the code. I had tried to
establish a standard through WrapperBase but some contributions
ignored this and some modules had grown up before the wrapperbase was
in place. I also think the ontology of applications is a bit screwed
up so people were putting modules in directories willy-nilly. I
think bioperl-run is even more cat-herding than the other aspects of
the project because applications change and there are very many
combinations of parameters and version of applications. There was a
flurry of contributions and no good way to marshall them initially.
So what to do - I guess make do with how it is but not rename
anything at this point but try and standardize things - it will
depend on your desire to really take this on.
The PISE modules may need to be re-evaluated to insure they are still
working - there was hope that some of them could be simplified
through some automated class generation from the XML they are based
on, but that was never developed.
We will try and a have an open-bio linux server available as a
platform where we can have a consistent testing environment but I am
not sure if we want this to represent every possible version of apps
or just the most recent.
On Nov 7, 2006, at 9:48 AM, Nathan Haigh wrote:
> Chris Fields wrote:
>> On Nov 7, 2006, at 10:07 AM, Sendu Bala wrote:
>>>> Is it possible to have the interface base class contain an
>>>> abstract program_version() (or similar) method? One that could be
>>>> implemented to just grab the current version of the program? I
>>>> was thinking about something like this for the RNA-based wrappers
>>>> I want to add but it seems to be something that most Run modules
>>>> would benefit from. Just haven't had time to look into it with
>>>> much detail.
>>> What's wrong with version() ?
>> Didn't realize there was a version() already.
>>>> Almost wonder if it would be easier to have all bioperl-run
>>>> modules have a Run-specific Root object for common methods,
>>>> regardless of the other interfaces used. Maybe something
>>>> inheriting Bio::Root::Root...
>>> What's wrong with Bio::Tools::Run::WrapperBase ?
>> I agree that WrapperBase fulfills most of this functionality.
>> However, don't a few bioperl-run wrappers 'roll their own', i.e. not
>> implement WrapperBase? Requiring all bioperl-run module inherit the
>> same Run-specific base object would integrate them a bit more and
>> distinguish them from the core modules (which would inherit
>> Bio::Root::Root). Anyway, it's just a suggestion.
>> Christopher Fields
>> Postdoctoral Researcher
>> Lab of Dr. Robert Switzer
>> Dept of Biochemistry
>> University of Illinois Urbana-Champaign
>> Bioperl-l mailing list
>> Bioperl-l at lists.open-bio.org
>> From what i've seen of Bioperl-run there is not a whole lot of
> consistency between modules. I think it would be a good idea to
> have all
> run modules inherit WrapperBase - would this be a big change??
> Bioperl-l mailing list
> Bioperl-l at lists.open-bio.org
More information about the Bioperl-l