Proposing a Silverblue SIG

Project Discussion
, ,
1

Currently Silverblue sort-of exists in a “gap in the sofa”, so far as I can tell, between two groups:

  • @tpopela and @siosm tend to do most of the visible “Silverblue things” (responding to tickets, requests for help, etc) but have told me on many occasions that the “steering” is done by the Workstation working group.
  • Michael Catanzaro (can’t see his handle on here :eyes:) was very surprised to hear this when I finally approached them about this :upside_down_face:

As such, I think it makes sense to either formally place Silverblue within the purview of the Workstation working group (like Kinoite is for the KDE SIG), or to create a Silverblue SIG to drive development and steering of the edition (and I do think we need to start thinking of Silverblue as an edition, if it is a serious contender for the future of Fedora).

(it is worth noting that not all Workstation working group members currently see a future for Silverblue, or where it should “sit” in relation to Workstation - which is fine, but would underline the value of a SIG that can make the case for Silverblue)

Creating a SIG could provide important decision-making capacity to handle issues such as:

On that last point: there seems to be a push away from using rpm-ostree to layer things in CoreOS, which I don’t really understand - CoreOS/Silverblue are definitely container-forward operating systems but the unique “selling” point of CoreOS/Silverblue is that you can still install packages to the base layer if it’s the best option (e.g. the nvidia drivers on Silverblue, or Cockpit on CoreOS). I would go as far as to say that I don’t understand the point of CoreOS/Silverblue if I can’t leverage the Fedora RPM ecosystem, which is one of the crown jewels of the Fedora project.

Finally, I would like to pitch an idea that I think “places” Silverblue very cleanly within the Fedora Project. The Fedora project has a mutable stack that’s split into Server and Workstation editions; I see Silverblue as the immutable Workstation edition of CoreOS, and (if this becomes an objective for the next five years) a “future” of Fedora that people can start using now. Something like this:

Now Next
Fedora Workstation Fedora Silverblue
Fedora Server Fedora CoreOS

Obviously, Silverblue (and CoreOS) isn’t ready to replace mutable Fedora - but if this is where Fedora might want to be then I think it makes a lot of sense to create a SIG that can progress on documenting and solving the problems that stand in the way of an immutable future.

7 Likes
2

I think this is a great idea, well argued, and I especially love this table:

Timely, but @dvlv has just ported the silverblue page to the new site, check it out: Fedora Silverblue | The Fedora Project.

Is there a process planned towards getting it promoted to being a main edition, like CoreOS is?

4 Likes
3

Hey, that looks really good! I have some thoughts, but the primary one is let’s get it up there ASAP to replace the ageing current iteration :sweat_smile: Great work dvlv (another person I can’t find on here - huh :eyes:)

4

I’m in if we end up creating one! The process is described in https://fedoraproject.org/wiki/Creating_a_Fedora_SIG.

You might also want to post a link to this discussion to the fedora-devel mailing list to raise visibility.

3 Likes
5

I would only ask: please don’t create a new mailing list.

1 Like
6

I’ve been involved with Silverblue off and on over the years, including in the beginning when we thought a lot about the product vision, and on the design side working on the user experience.

If a Silverblue SIG happens, I’d definitely be interested in participating on occasion, but sadly don’t have a huge amount of spare time so can’t make a firm commitment.

2 Likes
7

I would be willing to participate too if the timezones work.

8

Okay, I created a wiki page - and if it has a wiki page it’s real, right?

https://fedoraproject.org/wiki/SIGs/Silverblue

Very much a starter for ten, even with the meeting schedule (it’s set to once a fortnight, starting next Monday at 1500 1400UTC) - I am guessing at what makes a good time, so we can change it up (with the caveat that it needs to work with my schedule if I am to commit to running/minuting it).

Any suggestions, please fire away :slight_smile:

1 Like
9

I think the first thing on the agenda should be…
What is Silverblue and what’s it’s target audience.

When I first heard of Silverblue it was on version 30, I read it was atomic and immutable. This to my understanding was that this OS does it’s own system administrated work and I shouldn’t be able to destroy it, which I found attractive.

To me, the target audience is the user

Naturally I was curious to break it on purpose. If anything I would just reinstall this system. Rinse, wash, repeat. Only thing that stopped this cycle was setting up the home folder over and over again. If you distro hopped you know what I mean. Thus, my interests in homed.

Even thou we have the options to overlayer on the OS, I think the target audience is still the User.

In my general opinion, Fedora, like many other distros concentrate on system administrators and not really on the User. This makes Silverlue more appealable to the User. Straight to work OS, just like GNOME’s desktop.

10

It’s been brought to my attention that the Flatpak SIG is at 1400UTC on alternate Mondays, so I’m going to move the proposal to that and bake it in for this cycle, unless there are any objections. We can always change it after a few cycles, but I think it’s more important that we have a time than that we spend time trying to work out when to have one :slight_smile:

I’ll create a new thread for the upcoming meeting, to bring together a list of topics to discuss. We will also use that to drop the minutes and act as a forum for the discussion to continue.