Debian and Scientific Linux (was: Removal requests submitted for CERNLIB packages)
* Steffen Möller <steffen_moeller@gmx.de> [110305 12:31]:
> My hunch is that eventually we will see the efforts behind Scientific
> Linux merge with some major Linux distribution. And the reason most likely
> will be to prove to their funding parties of their impetus to give back to the
> world as much they can - independent from their hunt after some quark.
> This may be Fedora or Debian, but they will come, I am sure. That they
> have Scientific Linux and are Open Source already is per se already quite
> remarkable.
Hi,
it is important to understand that Scientific Linux is not just a
Linux distribution build from scratch by Fermilab, CERN and other
scientific labs. It is a RHEL clone, that is rebuild from the
RHEL source rpms that are provided by Red Hat for free. The main
goal of Scientific Linux is *not* to be as scientific as possible,
e.g. in terms of the number of scientific software packages
included in the distribution. (As a matter of fact, Debian comes
with much, much more scientific software packages than SL does).
The main goal of Scientific Linux is to be as close to RHEL as
possible in terms of binary compatibility, i.e. identical
kernels, identical library versions, everything build with
identical compiler flags etc.
The rationale behind this is that the majority of commercial
software packages for science and engineering are tested and
certified on "enterprise" Linux distributions only (in some
cases: on RHEL only). So it is more likely that these packages
will also run hassle-free on SL as compared to other
distributions that are untested and uncertified by the vendors.
This is indeed a valid point as most vendors reject any claims
for support if their software runs (actually: doesn't run or
doesn't run correctly) within an uncertified enviroment. If this
happens to the users, they are usually at a complete loss, even
if they've paid tens of thousands of Euros to the vendor for a
maintenance contract. Scientific Linux aims to minimize these
kind of risks for its users.
> What may be helping to speed up that process could be
> * identify a series of industries that should have some interest in the
> technologies maintained through CERN (car industry for crash tests
> maybe?, geologists?, astronomers? ..) and a blog about it
> * more personal contacts between our distro and CERN-affiliated scientists
> via conferences maybe?
> * some larger research group that decides to use Debian rather than
> scientific linux
> * maybe that group could also get some industry/research money to
> describe the CERNLIB and friends for regular industries
> * ...?
I think it's not so much an issue of the awareness/acceptance of
Debian among researchers. It's more an issue of the commercial
vendor's perception of Debian as a suitable platform for their
proprietary software.
Best regards,
Juergen
--
GPG A997BA7A | 87FC DA31 5F00 C885 0DC3 E28F BD0D 4B33 A997 BA7A
Reply to: