William Stein wrote:
Hello,
I officially propose removing DSage from the Sage distribution.
Why?
1. Nobody has worked on the code for years.
2. The guy who originally worked on the code is gone, and can't
work on the code even if he wanted to (due to conditions of his job).
3. So far as I can tell, DSage is completely and totally broken,
and has been for *months* in shipped versions of Sage, so absolutely
nobody could possibly be using it. See
http://trac.sagemath.org/sage_trac/ticket/7975
If you care, please vote.
[ ] Yes, I don't use DSage (don't even know what it is...), and
think it should be removed from Sage.
[ ] No, I think DSage should stay as is. I love using DSage!
[ ] Sort of -- delete Dsage for now, but if a maintainer/developer
comes along and fixes it so it actually works, then DSage gets back
into Sage for free.
I vote: Remove it, and don't ever, ever allow it back in, even if
somebody wants to actively develop it.
Rationale:
- For simplistic parallell computing on a single machine there's now
other, much simpler solutions implemented
- For running on a real cluster, I believe the approach is wrong. Most
clusters used in academia is run by a professional support staff and
have a sophisticated setup for scheduling jobs. I'd never get DSage set
up correctly on my cluster without pestering a lot of admins. However I
*can* of course integrate Sage with the cluster-provided job scheduling
system in user space if I want to. That's the way to go IMO.
- Running on a small home-brew cluster should a) fall outside of Sage's
scope, and b) use existing libraries rather than home-cooked to a much
larger extent than DSage does (use OpenMPI or Google's Map/Reduce or
something in-between).
Dag Sverre
--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to
sage-devel+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org
