On Wed, 16 Mar 2011 08:50:19 +1100, Helen South wrote:
One of my issues with major/minor is that we then have the pressure to conform to Major/Minor. The progression must be followed, but what if upstream have no major changes and nothing much is happening with the kernel?
Well, I think part of this comes from the underlying idea that the version has to be bumed by a 'major' increment every 'x' amount of time. Hence why that decision (on product version) IMHO is more of a marketing discussion than a technological discussion.
What constitutes 'major'? Is it seen as partisan to mark a significant development in one desktop or another? Can one major follow another immediately if some major development occurs?
Indeed, these are fair questions.
My impression is that often change is a gradual, almost organic process with a lot outside of our control. The 11.4 to 12 step could be regarded as a major/minor, or it could simply be an abbreviated incremental process. I gather it is seen somewhat as a step to 'major' but it concerns me that this means that to go to 12, we are actively looking for reasons to make it major (without stepping on KDE toes over Gnome SHELL, for instance).
In my formal role at Novell, I've had this discussion internally (I work in the training department on testing & certification - the discussion has been in the context of 'internal' to my small team, not with the SLE team itself) with regards to SLE, and ultimately it seems that Linux in and of itself is an evolutionary platform. It's pretty rare for there to be a major change (such as the change in look from KDE 3.5 to 4.0 or from GNOME 2 to GNOME 3 - both of which IMHO are major changes that immediately come to mind) within the product itself. The evolutionary nature of the Linux platform itself does tend to raise the kinds of questions that have been raised (now repeatedly) about how to version the openSUSE product. That's one strong argument for not continuing the current scheme, because every release ends up with a discussion around "so what version will we be releasing next?". That's also a big point in favour of using a YYYY-MM or YYYY-XX style convention rather than a sequential convention.
I worry about painting ourselves into a corner by formalizing the major/minor idea. Some sort of flexibility would need to be built in, possibly to the release schedule to allow for desktop or kernel delays.
Not necessarily negatives but just a couple of issues that would need to be addressed if we went that way.
Agreed.
I agree somewhat with Chuck; this is something that should probably resolved quite quickly.
One might even argue it should have been decided some time ago, but yes, it is something that certainly needs to be worked out before the next release. This is one case where I think tradition probably is not a very strong argument - "we've always done it this way" isn't a good enough reason if it means we have this kind of discussion every release cycle. Jim -- Jim Henderson Please keep on-topic replies on the list so everyone benefits -- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org