Mailinglist Archive: opensuse-project (465 mails)

< Previous Next >
[opensuse-project] Re: openSUSE versioning scheme
  • From: Andreas Jaeger <aj@xxxxxxxxxx>
  • Date: Tue, 6 Jul 2010 21:20:24 +0200
  • Message-id: <201007062120.25421.aj@xxxxxxxxxx>
On Tuesday 06 July 2010 17:19:05 Michael Loeffler wrote:
On Tuesday 06 July 2010 16:16:32 Andreas Jaeger wrote:
On Tuesday 06 July 2010 15:54:21 Michael Loeffler wrote:
+1, I still don't see what's wrong with the current scheme nor real
benefits from a new scheme.

See the page that vincent wrote, my pain points are
* in the software world a .3 release has only minor changes, the major
gets released for major updates. For us there's no difference.
* I've seen to many confusion with SLE and openSUSE numbers

Michael, just answer the following honestly:
* What is the next release called: 11.4 or 12.0? Why?

11.4 because according to the current scheme 11.4 is after 11.3

* Why did we call 11.0 11.0 and not 10.4?

for 10 and 11 the .0 version was always the one prior to a SUSE Linux
Enterprise (SLE) release and .1 was the one SLE was based on. Why should
we give up the relation to SLE? Isn't it a thing to be proud of? At
least people telling me they are.

Which brings us to another point: Nobody else is aware of that! If you say
this is the status quo and we go for it, it needs to be documented! Henne
told Vincent: We always go to .3 and then increase major.

I only remember that discussion as: Let's increase it in time for SLE11 -
we even discussed basing SLE 11 on openSUSE 11.0 ;)

As we deliver a huge amount of software for many areas it will be more
then difficult to go to an .0 release because of any major change as it
won't be easy to set the rule what a major change (major desktop
version, major kernel, major or new application etc) is.

Exactly, this major/minor renaming is bad.

Btw. if we go to the SLE numbering, I propose to change it so that SLE is
based on the *last* minor release - so, once SLE is out, we increase the
major.
That gives an even stronger signal: SLE12 is the ultimate continuation of
openSUSE 12 ;)

Otherwise there's too much going back and forth when to release SLE and
openSUSE and when to increase the version number - just imagine SLE
suddenly decides to get released earlier/later than planned.


* When and why do we end the 11.x numbers and go to 12.0?

I'd propose to go to 12.0 prior to SUSE Linux Enterprise 12 and base SLE
12 on 12.1. In that case we probably would have a 11.5 or higher -
don't know the schedule for SLE 12.

And I'd like to get that unknown release date out of the equation.

And we're just so honest let me add some argument from my side why I'm
oppose to change the versioning scheme:
- do we gain more contributors or users through a new versioning scheme

Not gain - but avoid confusion and help planning.

I'm glad that we're now at a strict 8 month release cycle - one headache
less. I'd like to see a similar predictable version number.

Neither of us can say with authority today what's the name of even the next
two releases is and that's bad! Let's get the same predictability into it
like the 8 month release cycle!

? - we had twice a me too approach by giving our releases names
(lizards and philosophers) and this didn't turn out to be a success

Part of that was that we always spoke about version numbers.

- I don't stick to the confusion argument because with software there is
always confusion and because of the fact that you normally hear only
from defects, complaints or confusion. The huge majority (I think way
more then 90% in our case) you never hear from as they are okay with
our product - the stability/quality argument I don't buy neither as we
normally know after the release about stabitliy and quality when its
used on a kind of endless number of machines
- still the issue with current versioning and the benefits through a
change are in my opinion weak

Michl, look at the discussion going on here, e.g. what Per is saying -
software developers interpret the version numbers different than the
openSUSE distribution does. There is confusion!

- Assuming there is a real and harming effect of the current versioning
I see way more things around our project to focus on to improve the
project over all:
- strategy
- grow our community together (eg. ml vs forums)

I consider a constructive discussion where we learn how to discuss and work
together community building as well ;)

- get more contributors to openSUSE
- lower the bar for people new to openSUSE
- developer documentation
- tools integration

If we fill the project with life, activity, good stuff and tools we will
succeed independent of any versioniong scheme.

Yes, we have to focus - but that also means we have to define some things
properly and the discussion here has shown that there's no clarity today
which allows us to focus.

Let me summarize what my opinion is:
* A version scheme needs to be documented to avoid these kind of
discussions - we failed so far in this arena
* A version number scheme needs to be predictable - we fail with your
proposal basing SLE12 on openSUSE 12.1. My counterproposal to increase
directly after SLES12 to 13.0 is better here...
* I would prefer to have a superior version scheme but I haven't seen it
brought forward. If some great idea comes up, let's do it. But I agree
with let's not change it just to change it ;)

One proposal for change that I liked was to wait until 2012 and then use a
year.release or year.month scheme. That would give us for 2012 and 2013:
12.1, 13.1, 13.2 (preferred IMO) - or 12.7, 13.3, 13.11.

Andreas
--
Andreas Jaeger, Program Manager openSUSE, aj@{novell.com,opensuse.org}
Twitter: jaegerandi | Identica: jaegerandi
SUSE LINUX Products GmbH, GF: Markus Rex, HRB 16746 (AG Nürnberg)
Maxfeldstr. 5, 90409 Nürnberg, Germany
GPG fingerprint = 93A3 365E CE47 B889 DF7F FED1 389A 563C C272 A126
< Previous Next >