Yes, I understand. So I guess the question is, what is the short and long term goal of OpenSuSE? Personally I think the long term goal automatically = the short term goal without a lot of people realising it. When I first started using linux, the idea of building brick upon brick and improving on previous releases was and still is fantastic, I think it actually drives a lot of people to keep doing this, but inevitably changes mean bugs and bugs mean some of the bricks get kicked out below us but most people don't think far enough ahead to realise what is happening here.
The way I see it, the manner in which we are making these changes means the goal of having a truly stable OS will never be achieved which leaves me wondering what is the point of all this? Where is the next step for distro's in the opensource movement? How will a community distro ever reach mainstream with such unpredictable results? Let's face it, even post release there are no gaurantees that anything unfixed will be fixed, maybe in the next release but chances are something different will be broken there too. At times we will get close, other times will be bad, but there will still be no way anyone will know how good a release is unless they try it for themselves which I think you'll agree is not a very good advertisement for a distro.
From personal experience there have been 2 good releases since the 9.x
series only, SuSE 9.3 and OpenSuSE 10.3 and even they were not perfect. That's 2 out of 8 releases or 25% of the great work that is done. It seems like a waste to me.
I think you'll also agree that there are some things that you just cannot release with if they're not running right. Two examples would be no X or no bootloader. There are many more, these should truly be blockers, maybe we could start by tasking someone with identifying those and laying down the beginnings of some absolute rules?
All in all, if this isn't already mandated somewhere perhaps a compromise could be that no release is considered final until an x.3 release and maybe even only that release will continue to receive the best backporting of fixes, stability improvements etc? That way people who want the latest can always upgrade to any release, bug test etc, people who want stable know to get a .3 release which comes around every 2 years. We could effectively consider x.0, x.1 x.2 = RC1, RC2, RC3 and the x.3 = final.
I think these two points would start to give to point us in the right direction, maybe even start giving us an advantage over other distros including some of the more stable ones.
What do you think?
On Thu, 2008-09-25 at 18:08 +0200, Andreas Jaeger wrote:
"Quentin Jackson" Quentin.Jackson@exclamation.co.nz writes:
Well I found the priority documentation at http://en.opensuse.org/Bugs/Definitions#Bug_Severities, it seems quite good, so further to below, I feel we should not be releasing a distro with any outstanding job having a criticality over Normal. Obviously there would be exceptions, but it seems that we do indeed release a distro with 'Major' severity outstanding calls correct? So by proxy that would be saying that we're OK releasing a distro that has "Major loss of function"?
We include priorities as well - not have any P1 bugs.
In the past we had as policy no blocker. No bugs with severity above normal is not feasible - we would never ever release ;-) or release with an outdated distribution.
There's indeed quite a difficult balance between time and quality - and we do this also with limiting the changes we allow to make to the distribution at some time,