I think that there is no danger in this: we (as astronomers) also use
common science software - from plotting to math libraries. And at least
I would try to ensure that the packages which are needed for astronomy
are in a somehow good shape -- independently whether they are in
debian-astro or debian-science or debian-python.
However, I have the feeling that many science packages have a bad
upstream development philosophy in common: they seem to tend to include
oldish, abandoned libraries, sometimes even patch them for the own
needs, not following the common SW development. It is hard for us the
take these libraries back out and replace them by dependencies, and to
continue maintaining these abandoned libraries. Grace is just the
current example here, other science, also astronomy programs may follow
with the same problem. But this is something only upstream can change
(or the maintainer, and a cooperating upstream), and often one just
cannot convince upstream at all -- even if other distributions will run
into the same problem sooner or later.
But this is independent of the science field, and independent of whether
we run our own mailing list or blend. Any idea how to solve this?