-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512
There have been a bunch of scattered responses---good for discussion, bad for knowing exactly what to respond to!
My thoughts now...
- - Turns out there are some people using Fedora Jam. Maybe we're not doing as badly as we thought! But even they want a package group, which makes perfect sense. Does anybody know how to get one of these? If not, I can do some research about that.
- - Interesting how my previous comments were interpreted. I didn't actually mean we should abandon Fedora as the base for our work, and I also didn't mean we should abandon the Fedora community (even though, you know, that is what I wrote). So the rest of my points are a revised version of what I was thinking.
- - If we use the openSUSE Build Service, we can build packages for any Linux distribution. We can also branch packages from existing version on the OBS, or volunteer to share maintenance with contributors to other distributions.
- - The OBS lets us set our own packaging guidelines, and it's much easier to become a packager---if this is our primary concern at the moment, then this may be a solution. It also means we can decide to ship things that violate Fedora's licensing guidelines, and that we could ship "content" (like sample sets) too.
- - If we make a remix that basically consists of Fedora + OBS_repo, that seems pretty easy. If we don't want to bother with the remix, then we won't.
BUT.
I wonder if I'm just getting cabin fever. I'd like to know what Brendan thinks, because he's the one who currently does the most work.
Maybe we should all just shut up and try to pick up five of the packages he maintains.
Christopher
On 26 January 2014 13:11:47 you wrote:
On Sun, Jan 26, 2014 at 11:39 AM, Christopher Antila crantila@fedoraproject.org wrote:
[snip]
Problem 1: with a "Spin," we can't accomplish everything we wanted (e.g., no realtime kernel, no MP3 support).
Is the real-time kernel still an issue? Sure, in the days of slow machines, 32-bit x86 and low-end sound cards it might have been a problem, but really - with a 2 GHz 64-bit chip and audio cards that do the heavy lifting, is the real-time kernel still needed?
MP3 - the other distros get around that by paying money to other businesses. Google and Mozilla are slowly but surely hacking away at the need for MP3. We'll get there, maybe even in my lifetime. Firefox OS doesn't support FLAC, for example. ;-)
Problem 2: we didn't get the additional contributors we hoped for.
Packaging is labor-intensive and capital-intensive. That's why I don't do it.
Problem 3: even the contributors don't use or recommend the Spin. I installed the F19 spin, but I wouldn't do it again, because it's just not what I want, in the same way the Desktop Spin and the KDE Spin aren't what I want.
Well, I have two hobbies - algorithmic composition / digital sound synthesis and data / computational journalism. So no single spin covers both and my machine is basically Fedora Design Suite plus R and R Studio plus the packages from the Jam spin.
Problem 4: we don't have enough contributors to do what we want---or we want to do more than we can.
Nobody has enough 'contributors', where 'contributors' == 'people who work for love instead of money' ;-)
Problem 5: pulling ourselves into the Fedora community necessarily means we restrict what we can do, both technically and legally.
I don't think that's really a problem - I'd rather have Fedora's restrictions than Ubuntu's chaos and all the other distros' lack of already-packaged audio solutions.
I'll propose some solutions too. Because the spin didn't and can't accomplish our technical goals, because it hasn't accomplished our social goals, because we don't seem to be using our own work, and because it's causing additional effort when we can't afford it, let's drop the spin. That's an easy first step.
I'd still like to see a yum package group/groups salvaged from the kickstart file.
[snip]
But what comes next? (Or: what comes .next?) We may need to leave Fedora.
Like Brendan, and probably many of you, I really appreciate the Fedora community's published values. However, through my time here, I've begun to realize that Fedora is the upstream for RHEL, the Fedora community is for innovating in the cloud, and if something isn't going to make money for Red Hat, there usually isn't enough initiative to make it happen (or: these initiatives get clobbered by the contributors who have more time because they're paid by Red Hat).
Let me clarify that I honestly believe everybody is acting in good faith, and that Red Hat's influence is overall a positive thing for the Fedora community and Linux in general. This is a "tyranny of the majority" situation: what's best for accomplishing our needs and desires is different from what's best for accomplishing the needs and desires of most of the rest of the Fedora community.
If we 'leave Fedora' there are essentially two options:
- We join the throng of 'labor of love / no product-market fit' audio
distros based on Debian or Ubuntu, competing for scarce resources with them, or 2. We migrate to openSUSE, where there's a huge infrastructure in the OpenSUSE Build Service and SUSE Studio, and try to find a paying market.
I'm seriously considering moving my CompJournoStick Fedora Remix back to SUSE Studio where it was 'born'; the appliance creation / distribution process there is a few years ahead of building ISOs on my workstation and I don't have the business model to go to Red Hat and say, "you should invest in computational journalism". I guess they'll figure that out on their own if I stay on Fedora. ;-)
I hope to encourage a discussion here. We need to consider how to "speak up" to get what we need. We have technical goals and community-building goals that aren't currently being met. Will "Fedora.next" allow us to meet our goals, or should we try to establish ourselves independently?
The good news is that F21 is going to have a long release cycle - at this point it looks like August and I'm guessing some big upstream must-haves will slip it a bit. So if we stay with Fedora, we "just need to marshall resources". ;-)
Which requires more effort,
and how much effort do we have to work with? If we leave, should we make a remix or simply offer an additional repository and a supportive community?
I don't see a point in another repository. Planet CCRMA and RPM Fusion have just about everything the Debian/Ubuntu repos have. If we're going to make a remix, I'd seriously consider moving to openSUSE - the licensing / branding stuff is pretty much the same as Fedora's but the SUSE-supported packaging and media creation infrastructure on the web is much better. And their KDE kicks major butt.
If we stay, how can we encourage the rest of the community to accept our admittedly-disruptive desires?
I don't know that they're all that disruptive. We have the same wishes all open-source projects have - robust, low-cost software without a bunch of lawyers getting between us and our users. ;-)