[Bioperl-l] Problems with Bioperl-ext package on WinVista?

[Chris Fields]

On Aug 17, 2009, at 10:45 AM, Robert Buels wrote:

> Yee Man Chan wrote:
>>    As to the release, my thinking is that I do understand that your  
>> desire to maintain a high level of quality in BioPerl code base. So  
>> if the HMM doesn't meet that standard, I am ok with it being  
>> spinned off.
>
> We're not pushing to spin it off because of code quality, we're  
> pushing to spin it off because we're spinning everything off.  The  
> plan is to break BioPerl up into many discrete distributions on CPAN  
> with the dependencies between them well-known and codified.  This  
> will make maintenance of BioPerl *much* easier in the long run.
>
> So this means that the plan of action should be
>  1.) get the code so that it's working on all platforms,
>  2.) create a CPAN distribution for it and put it on CPAN,
>  3.) remove it from bioperl-ext
>
> Also, doing a search for bioperl-ext on CPAN brings to light a  
> couple of issues that probably need to be dealt with.  To wit:
>
>
> 1.) there is an ancient version of bioperl-ext that probably needs  
> to be removed, it's under ~birney's account.  Thoughts on this?

This subject just recently popped up on perl.module.authors, more in  
relation to abandonware, but a similar thing.  Andreas has indicate  
there is an abandoned flag that can be set so it's worth looking into,  
but using it requires another release.

I have been in contact with that group on ideas for the split;  
libwin32 did the same thing, so I'll contact Jan Dubois on the matter  
for some pointers.

> 2.) Bio::Tools::( dpAlign | HMM | pSW ) all state that they depend  
> on bioperl-ext, which suggests that these really need to be split  
> off, each with the Bio::Ext::Modules they depend on.   
> Bio::Tools::HMM could be the first case of this:
>   * make a dir in the repos called Bio-Tools-HMM alongside bioperl- 
> live, having trunk/, and branches/ subdirs
>   * move Bio::Tools::HMM out of bioperl-live into that
>   * move Bio::Ext::HMM stuff out of bioperl-ext into that
>   * repeat with Bio::Tools::dpAlign and pSW, which would probably
>     go together into a Bio-Tools-Align distro, I think
>
> Sounds like this is moving along nicely.
>
> Rob

Yes, that's essentially the idea.  The more significant impact of this  
(both here and in core) is allowing updates to be made as needed, and  
not be blocked due to issues in unrelated modules.  We have been  
waiting years for fixes to pSW, Staden::read, Align w/o progress,  
which has hindered overall releases of bioperl-ext.  Similar problems  
exist in bp-core.

Re: bioperl-ext, BioLib has rendered some of those implementations  
obsolete.  I would rather do that incrementally (individual  
implementations) vs. wait for a full-blown bioperl-ext release, so  
splitting these up makes that possible.

chris