At the request of the FCAIC, I recently attempted to nudge spin owners to see which spins are being actively maintained. The ones on the F29 spins page with a last-update date are the ones where a human confirmed they read my email.
With cwickert saying at Flock last week that he's not going to own spins anymore, we need to look at how we want to manage spins going forward. Some options:
I vote for 3.
There has not been an active spins wrangler in at least 4 years likely more. We moved new spins to use a standard change process to be reviewed and approved a couple of years ago. The thing that we need to make sure happens is that people are actively maintaining the spins and that they are tested and working. We have done a bad job of doing that last bit.
I think if we added some milestones where the FPgM checked in with the owner of a spin that they have validated it is working and we have a plan to remove a spin if its not been validated. Ideally we could have something in fedora messages that informs the owner when their spin fails to build so they look at it, and we work with the spins owners to get testing of the spins in autoqa so that the checking is not needed often would be best
Can we maybe introduce a manual step (a dead person's switch) that Spins need to perform once per release to make sure that there is still someone caring about the Spins? Just using 3. will probably mean that there will be unmaintained Spins in the end. I am also fine with 1 or 2.
I like the idea of #3 with @till "deadmans switch" but I'd also like to see us able to have spins publish outside of normal release cycles to let them update to a current release when it makes the most sense for their target. They get dropped if they aren't on a current release.
I'd be OK with Till's suggestion as well.
Ultimately, I think Spins should be replaced with a tool that allows people to create media (ISO, container, whatever) with a recipe they maintain elsewhere. Then people could generate their spin on-demand with the latest package sets.
I'd be OK with Till's suggestion as well. Ultimately, I think Spins should be replaced with a tool that allows people to create media (ISO, container, whatever) with a recipe they maintain elsewhere. Then people could generate their spin on-demand with the latest package sets.
So we'd be circling back to something along the lines of Revisor, Canvas, or KIWI?
The only issue with this model is that it makes it hard to show that Fedora is more than the three Editions. It also cuts out a huge avenue for the community to be involved for other desktops.
I think people may be more involved if they are allowed to make things the way they want versus getting it prebuilt from someone else's decisions. They may still end up using the basic 'kickstart' everyone else uses but they can also swap and talk about why they do X or Y versus the default. It also allows them to talk about what they want to do with their tool.
Most people don't really care about the OS anymore than most people do not care about their car. [Sacrilege to car enthusiasts as much as OS enthusiasts but that is the reality.] They care more about what they are doing where there car (or bus or cab or hired car) took them. That is what they want to talk about and get involved in. The spin was a way at a time to get them to do that, but it isn't anymore as much.
At the moment, people come up with a new spin idea then find they don't have the time to keep running it. It then piles up unfixed items until someone tries to tamp them all down so the spin won't go away. That isn't involvement, that is fire-fighting.
I'd be OK with Till's suggestion as well. Ultimately, I think Spins should be replaced with a tool that allows people to create media (ISO, container, whatever) with a recipe they maintain elsewhere. Then people could generate their spin on-demand with the latest package sets. So we'd be circling back to something along the lines of Revisor, Canvas, or KIWI?
Sure.
I disagree. Why wouldn't the community maintain the recipes needed? Why wouldn't we be able to have e.g. xfce.fedoraproject.org point to some other server where this is well curated and maintained and hosted?
Exactly. We have a few well curated Spins and the people doing them put a ton of time in. The rest seem more like fleeting interests at best, despite best intentions.
I'd be OK with Till's suggestion as well. Ultimately, I think Spins should be replaced with a tool that allows people to create media (ISO, container, whatever) with a recipe they maintain elsewhere. Then people could generate their spin on-demand with the latest package sets. So we'd be circling back to something along the lines of Revisor, Canvas, or KIWI? Sure. The only issue with this model is that it makes it hard to show that Fedora is more than the three Editions. It also cuts out a huge avenue for the community to be involved for other desktops. I disagree. Why wouldn't the community maintain the recipes needed? Why wouldn't we be able to have e.g. xfce.fedoraproject.org point to some other server where this is well curated and maintained and hosted?
If someone wants to build a system that allows people to do that, have it built, and then offer it to download for people, then that's fine with me. I always wanted something like SUSE Studio for Fedora.
jwboyer: If the kickstarts are maintained the tools already exist for people to create their own isos (livecd-creator and live-media-creator) the Respins Sig have been using them for years to generate updated isos for the community.So far i am likeing Tills' idea
I suspect that http://weldr.io/ will play a part in any technical solution however that is likely something that FESCo would tackle once the council gives it recommendations and helps set direction for spins going forward
+1 to this.
Marking this closed. FESCo formally approved this.
Metadata Update from @bcotton: - Issue close_status updated to: approved - Issue status updated to: Closed (was: Open)
Login to comment on this ticket.