I'm not sure the project has the person-power to make that happen, without making other sacrifices. But this system really doesn't work that well, and it never has, in the entire fifteen years of the project. It seems to me that continuing to do that same basic thing is more inertia than anything. Every cycle, everyone grumbles about how bad that cycle just was, and every cycle, it gets worse. The system works, but it really kinda sucks, and a major overhaul seems very much in order.
So I think the sacrifices necessary to make this happen might be a very wise investment indeed. I'd be perfectly happy with a number of developers working on this instead of other stuff, and then a substantial chunk of them working on actually implementing it, once the design is hammered out. Yes, this means things will slow down, but slowing down for a year or two, in order to speed way the heck on a permanent basis, seems like a complete no-brainer, at least from the outside.
Now's the time to be doing this; if it doesn't start really soon, it'll be at least two more years of Debian being always behind the other distros. Debian's stability is legendary, I love it to death, and I use it everywhere, but it's stability at the cost of being a little obsolete all the time. The approach that Lars and Russ are proposing seems like a good way of preserving Debian's inherent robustness, the promise that "this software works!" Debian keeps that promise probably better than any other distro, except possibly RHEL, but I suspect this system would let it keep that promise while avoiding freezes almost completely.
This is a goal worthy of significant time investment, even if it hurts Debian during that time. I think the project would benefit so much in the long run that the short-term pain would be utterly worthwhile. And now's the time to do it.
Oh, and then one more thought: it seems to me that this might even get pulled into the Social Contract. I'm not too up on the wording, so just take this as a very rough idea, but adding something like "Consistent with our other goals and ideals, the purpose of the Debian Project is to make stable releases, so that non-developers can have easy access to the fruits of our collective labors." Something vaguely like that, anyway.
Maybe that isn't actually a goal of Debian, but as one of those end-users, and one who's been using it since '98 or '99 or so, it's always felt like it was. If there's real disagreement in the ranks about releases being the goal, then a big argument about "Why the heck are we doing this, anyway?" might be in order.
If the Debian community doesn't actually care much about releases, you guys are sure putting yourselves through hell to make them.
Copyright © 2017, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds