[Bioc-devel] RFC - annotation packages and ExpressionSets
mcarlson at fhcrc.org
Tue Jun 17 19:06:27 CEST 2008
Christian Kohler wrote:
> Robert Gentleman wrote:
>> Hi all,
>> As most of you are probably aware in the 2.1 release we will be
>> introducing new variants of (almost) all annotation packages. For the
>> 2.1 release, at least the naming scheme will be
>> hgu95av2, for environment based packages
>> hgu95av2.db, for database variants.
>> But, we have no intention of maintaining the environment based
>> variants (and they may indeed go away by 2.2).
>> So this suggests that in the 2.2 release we would then move to
>> hgu95av2, for database variants
>> hgu95av2.env for environment based ones.
>> An alternative is to simply retain the hgu95av2.db scheme (but that
>> gets a bit ugly).
>> Where things might become problematic is the interaction with
>> ExpressionSets, where we have (and often use) the annotation slot.
>> I propose that we modify the accessor function to return the .db name,
>> for this release and that we try and track down all direct access via @
>> in the code base and replace it with use of the accessor function.
> Dear list,
> sorry for reviving this rather old post, but I am interested in how the
> future plan concerning the notation of annotation packages will look like.
> At the moment, a 'Deprecation warning' appears while loading the
> environment based annotation packages and the use of the *.db*
> equivalent is recommended.
> Will the *.db* nomenclature endure or will the *.env* notation be
> introduced in one of the upcoming releases?
> I mean, it is a bit cumbersome to change your analysis-scripts over and
> over again :-)
> Thanks a lot for any advice.
Wow that is an old post.
For the foreseeable future, I believe that the plan is to keep the new
I think it could be pretty confusing for people if we started calling
the new packages by the old names right after we just got rid of the old
ones. Whenever possible, I think that we want to simply phase out the
older environment style packages as their existence is largely redundant
at this time. It also creates a lot of overhead for us to generate two
styles of each package produced here. The newer packages emulate the
old ones so the changes should be transparent for most people. It also
seems that having a *.db * extension conveniently "labels" a package as
being an "annotation" package so I don't think that it is really so
terrible to have that extension on there.
More information about the Bioc-devel