Re: RE : Kudos to High Energy Physics packagers (Re: r3479 - /projects/science/trunk/debian-science/tasks/highenergy-physics)
Hi Yaroslav,
On Sat, Jul 14, 2012 at 10:53:57AM -0400, Yaroslav Halchenko wrote:
> my 1 c
1 more c :-)
> On Sat, 14 Jul 2012, PICCA Frédéric-Emmanuel wrote:
> > Hello, yes we are preparing a PAN Blend (Photon and Neutron Blend) dedicated to the synchrotron and neutron facilities.
>
> Although I might sound hypocritical, I am not sure if separating out
> into a new blend would be worthwhile unless there is something radically
> new to be added to it which would not suite the whole Debian Science
> blend.
I do not think so. From my perspective Debian Science was rather an
umbrella for several potentially separate Blends of different scientific
scopes until somebody (read: some team) might start caring for it.
For instance we do have DebiChem and the biological part in Debian Med
which actually are historically before Debian Science but I do not see
any reason for not separating other sciences as well ... as long as
somebody will care for it. Considering the quite packed tasks of
mathematical, physical and engineering flavours this could be easily
done.
Moreover I think that some smaller more focussed teams could establish
a stronger connection between the team members (think for instance of
NeuroDebian team ;-) ... actually I consider NeuroDebian as a perfect
candidate for an own Blend - just talked with Michael about this).
While a Debian PAN Blend is even more focussed than a Debian Physics
Blend as I had personally in mind - it just depends from people who are
running the Blend and if there are people who are volunteering to do so
I can not find a reason to stop them.
> If that happens, IMHO, it would still be great if corresponding
> task page in Debian Science blend would stay maintained/updated as
> well.
Above I was talking the organisational matters - this one is a technical
fact. The technical solution for your problem is possible and was even
discussed on an Extremadura sprint back in 2007. We just need to define
some "Meta-Depends" which could resolve packages defined in a task of
some Blend into another Blend like Debian Science. If you actually look
into the biology task of Debian Science[1] you see that there are even
such Meta-Depends/Suggests fields defined. However, they are not yet
regarded for the moment because of technical details I do not want to
explain inside this context. I just want to say: A technical means to
cross-link Blends could be implemented (but needs a certain amount of
work, thought).
> And a separate blend issue imho is almost orthogonal to a having
> of dedicated (sub)teams, alioth project, mailing lists etc.
>From my point of view a separate Blend has the following advantages:
1. Forming a more dedicated team and by doing so attract people
who might be afraid of the bad signal - noise - ratio regarding
their specific workfield in the general Debian Science scope
(and so a dedicated list + alioth project should be created for
sure)
2. The design of the tasks could be made more fine grained
3. QA team from time to time criticises that Debian Science team
is not working properly because the size creates some kind of
annonymity - this could be solved by smaller teams.
4. Debian could become more attractive to very specific workfields
because they find a direct contact point (again: think of
NeuroDebian)
Having a separate Blend does not necessarily mean that people will
leave the contact to Debian Science (we have several examples of
members of specific Blends who are working also here - you and me
are at least two of those).
Kind regards
Andreas.
[1] http://anonscm.debian.org/viewvc/blends/projects/science/trunk/debian-science/tasks/biology?revision=2786&view=markup
--
http://fam-tille.de
Reply to: