On Wednesday 13 June 2012 20:57:21 Martin Schlander wrote:
Onsdag den 13. juni 2012 20:33:08 Jos Poortvliet skrev:
On Tuesday 12 June 2012 13:41:50 M. Edward Borasky wrote: <snip>
If *I* were making the decision, I'd say "let Ubuntu and Fedora fight it out for the mass market and switch to the Debian / Gentoo model". Because I think that's the way openSUSE is evolving anyhow - a strong, sophisticated community of IT users and technologies like Open Build Service and OpenQA. I'm using openSUSE and not Debian or Gentoo because of that.
I'm entirely on the same page with you.
A 1 year release schedule and using OBS and Tumbleweed to satisfy the needs for those needing more up-to-date software seems perfect to me.
Let's be honest - if you want the latest, 6, 8 or 12 months - it's all too slow so you use Tumbleweed anyway. If you want stable, 12 months is better than 8 is better than 6. Simple.
if it happens to help our release engineering, awesome too ;-)
Easy for you to say since you're already not using openSUSE but rather Tumbleweed.
A 12 month schedule does not automatically make it more stable, unless maybe if you also increase the various freeze periods. But that'll just make developers/packagers and testers even less interested in testing/using factory than is already the case - and move even more activity away from factory to little toy OBS projects - and hence away from the distro that actually matters.
As Coolo already said, this might be mitigated by making Tumbleweed a more integral part of how we work. Moreover, I absolutely agree that our distro gets not automatically more stable with a longer release cycle. Increasing the freezes and being even more conservative would, for me, certainly be part of this plan. Basically, we move openSUSE even more in the direction of a stable and slightly boring workhorse, while the 'cool' stuff happens in OBS repo's and Tumbleweed - all feeding back to the new stable release once a year.
I personally think 8 months is the perfect balance. It's not clear to me what exactly has gone wrong with 12.2 anyway. In my brief tests 12.2beta seemed pretty good. Usually the beta is the worst milestone of them all, cuz it's just after feature freeze.
There's certainly something wrong but I guess this is something Coolo can better address than I can.