[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Is there going to be a release soon?
From: |
address@hidden |
Subject: |
Re: Is there going to be a release soon? |
Date: |
Fri, 21 Nov 2008 08:05:58 -0800 (PST) |
User-agent: |
G2/1.0 |
On 21 Nov., 16:24, Adam Fedor <fe...@qwestoffice.net> wrote:
> On Nov 20, 2008, at 3:15 PM, David Chisnall wrote:
>
>
>
> > P.S. The stable/unstable version numbering system really confuses a
> > lot of people. Everyone expects 0.14 to come after 0.13, not
> > before. The odd-unstable/even-unstable system is used by a load of
> > projects (including the Linux kernel) but no one I've come across
> > else has them the same way around as GNUstep. I realise the
> > rationale for this was to encourage people to use the latest stable,
> > but this is usually a decision made by packagers (who go with stable
> > or unstable, depending on their target audience, and know which
> > order they come in) while people compiling from source usually just
> > use svn and don't care about versions.
>
> If it's a new release cycle (e.g. X.X.0), then the stable and unstable
> releases come at the same time. So if we release 0.14.0 stable, we
> release 0.15.0 unstable at the same time.
>
> I guess I didn't bother to release gui 0.15.0 (because it would be the
> same as 0.14.0 anyway), but I probably should have.
>
> Right now we're in the middle of a series, so the unstable and stable
> releases could come in any order depending on when the "release
> manager" wants to release them, but in any case the unstable release
> is always an odd minor number greater than the stable release. So I
> guess I am confused because you seem to think the opposite is happening.
I would like to propose a quite dramatic change of this.
IMHO, we should think about making it more similar to
e.g. gcc or glibc or many others.
There, the (most) stable releases are always of the form
"X.X.<maximum>" i.e. if you go to gcc 3.4 you will find gcc 3.4.6
(i.e. has bugs fixed)
And the instable one is the one with the highest available numbers.
I know this is a quite different view that there are intermediate
versions that are bug-fixed
and become more and more stable while work is already being done on
the next higher number.
So there is no need to "release" a stable version. Just start a new
X.X.0 when new features are
introduced.
And one thing makes me wonder a little: why is it still a 0.15
release? After so many years of work.
>From a marketing view this looks like an instable experiment and not
as something that can be used in daily work.
Linux started as 0.01 and went pretty fast to 0.99 and now we have
2.6.something. This looks mature.
So my proposal would be to go to 1.5.0 as the next step instead of
0.15.0.
Just my 2 cents,
Nikolaus