On Tue, Jun 09, 2009 at 03:08:32PM +0200, Michael Hanke wrote:
> Sure there are psychology people working in health care,
... so we have to catch it.
> as they are working in schools
... I'm not sure whether Debian Edu is interested (or I'm rather
sure they are not) but in a wider sense the packages would be
*also* a target for Debian Edu - just to make clear the idea
of Blends.
> the aerospace industry, or preventing crazy people
> from getting their driver's license back ;-)
Perhaps once a day we might have a Debian Transport Blend featuring
a transport-psychology package ... OK, just dreaming to find
another example how Blends might work.
> Thanks for the pointer. This approach, however, needs thought about what
> is egg and what is hen. a -med task referenced in -science implies that
> the former is a subset of the later.
This is not really a hen - egg - problem. We just *know* that med-bio
and med-bio-dev were there first. And this task is actively maintained.
There was no need for the creator of science-biology to diverge from
the worl of Debian Med but it was useful to add something which was
missing.
> You started it in -med hence you
> would consider the healthcare-related pieces a subset of science?
No. I started with Debian Med because there was no Debian Science
when I was starting.
> Or would you intend another relationship if a -science task would be
> started?
I would just apply common sense. If a science-psychology package
is actively maintained by experts and it *makes* *sense* to just
adopt their work unchanged I might consider just making the dependency
inside med-psychology from science-psychology to keep my effort for
maintenance low. If a science-psychology metapackage would contain
packages I would not like to include into med-psychology (for whatever
reason) the simple inclusion mechanism just would not work and I'm
bound to manual editing - which is in this case not so much work.
In the case of biology it is just a lot of work. So finaly we want
to reach a result which fits the need of our users (=proper dependencies
in tasks packages) and do not make a science of package categorisation.
Kind regards
Andreas.
--
http://fam-tille.de