Fedora.NEXT Products and the fate of Spins

Amit Saha asaha at redhat.com
Thu Jan 30 01:03:45 UTC 2014


----- Original Message -----
> From: "inode0" <inode0 at gmail.com>
> To: "Development discussions related to Fedora" <devel at lists.fedoraproject.org>
> Sent: Thursday, January 30, 2014 8:01:15 AM
> Subject: Re: Fedora.NEXT Products and the fate of Spins

> I guess I'd like those active in the spin community to make
> suggestions here. I imagine spins and other new creations built on
> Fedora to be things the project wants to promote, not push away. The
> reality may be that we can't do what we do now in support of spins,
> but I hope we can continue to do something that helps and encourages
> those making them.

I am the maintainer of the Fedora Scientific Spin and I speak wearing my Fedora community
member hat (not RH employee). 

A bit of background will put things into better perspective. The Fedora Scientific Spin
was first released during the Fedora 16 release cycle. I failed to fix a build failure 
due to package dep. issues during the F19 release cycle, so, we have had 4 releases since then. 
I haven't been bombarded with emails from users, but I have had emails from users who find it 
useful and have requested for things to be included (which couldn't be included for other reasons). 
One good thing to come out of the failed F19 release was that I came to know that people do care 
about Fedora Scientific, since I had potential users looking to download it, but didn't find an ISO.

Now, I will begin the pitch for Fedora Scientific to exist. I think Fedora Scientific is a good thing to have
in the Open Source scientific computing ecosystem. Thanks to the packagers, we have all the tools/libraries that
the ecosystem boasts of ready to use in Fedora Scientific [1]. We may be even attracting users to Fedora due to it 
(no facts to back it up), since the closest distro that aims to achieve what Fedora Scientific does is 
Scientific Linux. 

Some more points re. specific issues pointed out and what can be done about it:

- Starting with Fedora 20, all the spin maintainers were supposed to fill a matrix 
  (as someone else already pointed out) which was a validation that the test composes, 
  the RCs, alphas and the betas worked correctly (installation, applications packaged, 
  etc.). Overall I think it was a good thing to have as it helped the spin maintainers
  to fix issues with their builds. So, +1 to that and we should continue it. [2]

- I am not exactly sure about the cost of doing the builds, checking why they may have failed, etc.
  So, perhaps, this can be something that my be off-shored to the spin maintainers? It increases
  our responsibilities, but helps lighten the load from rel-eng.

- In my moment of "Is Fedora Scientific actually being used?" or, more recently, "Is it even a contribution worth a Fedora
  10-year anniversary T-shirt?", I have thought that perhaps, all the packages that Fedora Scientific ships
  can all be made into a yum group and the user can just do "yum groupinstall fedora-scientific". Yeah, perhaps
  it can be done, but I hope we don't have to, since it still needs the user to download a whole bunch of stuff
  after the install, the exact problem I wanted to solve using Fedora Scientific. 

Overall, I think Fedora Spins are a good thing to have, and we must address the question of how we can keep them
by sharing the work that goes into having them, rather than eliminating them or converting them into Remixes.


[1] http://fedora-scientific.readthedocs.org/en/latest/ (Needs a lot more work to make it complete)
[2] https://fedoraproject.org/wiki/Scientific_Packages_Testing

Amit Saha <http://echorand.me>

More information about the devel mailing list