Re: New Debian Science metapackages (Was: debian-science_0.16_amd64.changes ACCEPTED into unstable)
On Tue, Apr 17, 2012 at 10:03:19PM -0400, Yaroslav Halchenko wrote:
> So, here are few activities I think worth pursuing
> - catchy release notes statement highlighting well-covered fields of
> science, computing platforms (distributed, MPI, ...),...
We did it last time (squeeze) for the neuroscience-related bits and got
substantial press echo -- ranging from: "WTF...who needs that" to "This
is simply amazing".
> - some short high-level tech-review paper to be submitted to some
> relatively well known journal (not sure if we could aim at Science ;-)
> but who knows...) which provides a tasty introduction into such an
> ideal platform for a scientist, and then brief glance overview how it
> addresses those topics everyone in scientific communities
> drooling about: reproducibility (e.g. snapshoting with cpack, VMs),
> sharing (obvious for anyone in FOSS, but might be worth making it
> explicit), methodologies rapid dissemination and adoption,
> community-driven (once again -- obvious, but making explicit lack of
> commercial company baby sitting etc, would imho be beneficial), etc.
I agree -- this is what we should aim for eventually. It is relatively
easy to convince tech-folk that Debian has its benefits, it is a totally
different thing to get the message out to non-geek scientists. A
publication in a journal with good visibility would be very beneficial.
It should not be difficult to collect enough superlatives to justify a
publication. However, I think, we also have to present a couple of
actual scientific endeavours that rely on Debian where we can get some
stats on how many lab/people use it for what. Two things I can think
of immediately are:
- LIGO (http://www.ligo.org; gravitational waves observatory consortium)
- and due to personal involvement: NeuroDebian (http://neuro.debian.net)
There where numerous others that popped up on this list over the last
couple of years. I think that the focus of a manuscript should be on the
breadth of scientific fields that Debian supports. We should identify
the ones were it excels and emphasize those, but mention all of them.
Maybe the respective maintainers could give an opinion how many
non-Debian bits are necessary to do research in medicine, physics,
Main theme of the paper could be that only via collaborative work we can
support such a universal tool. The more specialized a particular
application is, the more it relies on software that is written by other
and maintained by other from other fields of research (or even outside
science). No individual field of research has enough manpower to support
such an endeavour on its own. It would not be hard to argue that
interdisciplinary collaborations are the major source/driver of innovation
in pretty much any context. Debian is just that -- an interdisciplinary
collaboration we vast practical benefits that are readily available to
anyone who wants them (and knows about them).
This paper should have a looong list of authors, with a list of
affiliations that the world hasn't seen before!
I'd be happy to help writing it. If we start soonish, it should give us
enough time to get it published with wheezy's release.