[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Proposal: Debian release numbers



On Wed, Jan 08, 2003 at 09:52:30AM +0100, Emile van Bergen wrote:
> There is also no good reason whatsoever to get rid of the current
> scheme. It can do everything that a major-only scheme can do, plus allow
> smaller releases to be made in the future.

I disagree with your last point.  First of all, there's no reason that a
minor release can't be made and given a .x suffix.  In fact, if the
release really was "minor", by some standard, then a .x suffix is
perfectly logical.  Second, in the scheme we're currently using, point
suffixes are not given based on whether the release is major or minor or
whatever.  Both slink and potato were "major" releases; that's why they
were given new codenames.  Yet they were only given "minor" version
numbers.

The way I see the current release scheme is as an artificial *deflation*
of version numbers.  We know what our major releases are by their
codenames, but we somehow give less weight to the change in versions
between slink and potato than between potato and woody.

> If for, say 4 consecutive releases, only X.0 releases are made (with
> their 'r' fixes), perhaps /then/ it would be time to say that Debian
> never uses the ability to indicate a smaller difference to the previous
> release and that we may want to get rid of it. 

Except that, in the current versioning scheme, the version is chosen
completely arbitrarily by the release manager.  He could, if he chose,
easily kill this idea simply by declaring that the next 4 versions will
be point releases.  The proposal we're discussing now would give actual
meaning to the version numbers.

> But IMHO, that time is not there yet, not by a long shot. It doesn't
> matter that it's been years, it matters that the major release just
> before this one (2) had several minor releases following.

And just what was it that made those releases minor?  There was no logic
behind it, it was totally arbitrary.  I'd be willing to bet that over
90% of the packages in Debian were changed between any one of those
minor releases.  Plus you must consider all of the new packages
available.  The same holds true for a major release like woody.

> In short, I see no benefit whatsoever in changing the scheme. I'd rather
> that more software adopt this flexible scheme.

The benefit is that the new versioning scheme would give actual meaning
to the version numbers.  Right now, they are arbitrary.  Simply put,
that's the point.

noah

-- 
 _______________________________________________________
| Web: http://web.morgul.net/~frodo/
| PGP Public Key: http://web.morgul.net/~frodo/mail.html 

Attachment: pgpImkQyOAS02.pgp
Description: PGP signature


Reply to: