Backports of scientific packages (Was: Re: GROMACS version)
Am Freitag, den 25.09.2009, 12:23 +0200 schrieb Steffen Moeller:
> The question is indeed a valid one. For larger compute clusters, "stable"
> indeed seems to mean "invariant". And while as scientists we always want the
> very latest, we also want stable compute clusters. This is a dilemma.
Indeed. That's why I backport. It seems to be the best compromise. This
whole problem is intensified by the fact that bugs in old versions might
make results become void. So having the latest versions kind of becomes
a neccessity.
> Would there be a way to auto-backport packages that are flagged in this
> respect? Similarly to some non-free packages being explicitly allowed to be
> auto-built? Basically all packages in debian-science and debian-med are very
> likely candidates, IMHO.
Yes. I like the idea but we simply can't rebuild everything from the
task pages of these blends since there are also tools from KDE or GNOME
which would mean to backport quite a lot of unrelated stuff. Also,
packages with code in interpreted language can almost always be used
directly from testing. But I think auto-building could work for a
well-defined subset of packages.
If there is interest in this and several (~5?) developers would
volunteer in maintaining these backports, we should get our heads
together and try to build something like that. I could probably donate a
VM for this purpose, and would volunteer to talk to the backports.org
people. But given my time constraints at the moment, I will need support
from interested people (even non-DDs) to make this happen.
Best regards
Manuel
Reply to: