[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]
Re: science category
> On Apr 19, 2017, at 8:30 AM, Jason Bacon <bacon4000%gmail.com@localhost> wrote:
> The fact that the collection is continuously updated actually makes it unsuitable for a lot of scientific research, though. Many researchers need to keep a software installation static for the duration of a study since newer versions might alter results.
> One of the beauties of pkgsrc, though, is the inherent ability to install multiple snapshots on the same machine under a prefix of your choosing. For research tools, we have snapshots available over NFS in /sharedapps/pkg-2017Q1, etc. to provide continuity. Users just load the module for a particular snapshot and they're on their way. This is a lighter and simpler solution than containers or other methods people often use.
> On my clusters, I also keep a pkgsrc tree on the local disk of each node in /usr/pkg for system tools (SLURM, subversion, git, editors, etc). Your binary packages should work well for this once the necessary packages are committed and vetted.
> My plan is to build packages for the /sharedapps snapshots and provide them to the community. It's all approved and in the works, but it's just a matter of man-hours. I'm 1 of 1.5 FTEs supporting research computing for most of the university and I spend most of my time responding to immediate needs. The project is inching forward, though…
I think the outcome of this thread is that adding the “science” category is an ok thing to do, as long as there is a set of packages to populate it with.
Jason, given that we both have things that might reasonably go into such a category, why don’t we collect the packages and create the category as soon as a reasonable collection is ready to be committed. We are in the process of hiring someone to work on this and so should be able to help expedite the process of updating and checking the existing science collection.
Main Index |
Thread Index |