Re: Task: parallel-* ?
On Wed, Jan 23, 2002 at 09:31:19AM -0500, Adam C Powell IV wrote:
> Martin Quinson wrote:
> >typically, *-dev is used to designate the packages containing the headers
> >and other stuff to compile things yourself. So, I'll name the first of your
> >tasks 'parallel-computing-master', or something like that.
> >We could also make a third task called '-dev' containing all .h and so on.
> Good point. The -dev task has traditionally had things like lam3-dev,
> etc., but those are not for the master/head node. I'll gladly change
> the description to reflect that.
> So what does go on a master/head node, but not on the others? Is there
> a need for such a -master task?
I dunno. I though the title of your task was wrong, regarding to your
description. But it appears that it's the description which was wrong, and
the title ok :)
In fact, I have some ideas for this -master task, like the nameserver and
memory if you have NWS, or the agent if you have NetSolve or Ninf. and so
on. But the fact is that NWS is not in Debian yet (although you can find
some packages on my page) and that NetSolve and Ninf are not packaged yet
(although I think I'll do that one day).