[Bioperl-l] Splitting Bioperl and Test related Suggestions
cjfields at uiuc.edu
Thu Jul 5 12:13:30 EDT 2007
On Jul 5, 2007, at 11:02 AM, David Messina wrote:
> I know this has been discussed before, but while we're talking
> about future release plans, it might be worth revisiting the
> BioPerl policy of designating only even-numbered releases as
> 'stable'. It's taking so long to get from 1.4 to 1.6. While the
> principle of keeping a stable API between 'stable' releases is
> valid in the ideal case, I think that continuing to label 1.5.2 (or
> whatever the latest 'dev' release is) as a developer release (which
> implies potentially unstable or bleeding-edge code) is highly
> misleading since we would never ever tell anyone to get 1.4 instead.
> Alternatively, if we adopt a more aggressive release schedule as
> Chris proposed a couple days ago, then perhaps we could agree to
> push out an even-numbered release once a year or so, so that there
> is a 'stable' release we could recommend.
I think the idea of 'stable' is best summarized back in Hilmar's post
(i.e. we support a particular API for that release). The 1.5
releases I believe break some aspects of 1.4 API (some of the Feature/
Annotation stuff introduced before the official 1.5 release). We
still need to address some of those issues before a 1.6 which seems
to be the only real stumbling block, but they are unfortunately not
well-documented and are somewhat interwoven with GMOD code.
> To make it easier for people to obtain the latest tarball, we could
> put the 'download tarball' link directly on the 'Getting_BioPerl'
> wiki page instead of only a link to the viewcvs interface. That way
> they wouldn't have to navigate the source tree to figure out which
> tarball they want (which is almost always going to be the bioperl-
> live tarball).
> I think the actual URL underlying the 'Download tarball' link on
> viewcvs is stable:
Sounds reasonable enough. Do you want to do the honors?
More information about the Bioperl-l