RFC: Spec versioning

[Mark McLoughlin]

On Wed, 2005-03-23 at 23:16 +0100, Waldo Bastian wrote:
> Hi,
> 
> I thought it might be useful if we could agree on how to version specs.
> In particular it isn't always clear when to update the version number. E.g. 
> when working on a number of changes to a stable spec you don't want to use a 
> new version number for every change, but you do want to change the version 
> number when making the first change to make clear that this isn't the stable 
> version any longer.
> 
> To accomodate that I propose the following:
> 
> version:	<major>.<minor>.draft-<draft-level>
> status:	work in progress, subject to change,
> 		not to be used for any implementation
> website:	http://standards/foobar-spec/foobar-spec-<major>.<minor>.html
> 
> version:	<major>.<minor>
> status:	Finished version, ready for implementation
> website:	http://standards/foobar-spec/foobar-spec-<major>.<minor>.html
> 
> version:	<major>.<minor>.<patchlevel>
> status:	Functional equivalent of <major>.<minor> but may contain additional
> 		clearifications or examples to improve understanding.
> website:	http://standards/foobar-spec/foobar-spec-<major>.<minor>.html
> 
> When making a functional change to <major>.<minor> or 
> <major>.<minor>.<patchlevel> the version should be updated to either 
> <major>.<minor+1>.draft-1 or <major>.<minor+1>

	Yeah, I think that makes some sense. A little complicated and hard to
follow, but it does make sense ...

	Perhaps we could just drop the patchlevel part? Is you're not making
any functional changes, why change the version number at all?

> Questions:
> Can <minor> be larger than 9 ? Is 0.10 the next version after 0.9 or is 0.10 
> equivalent to 0.1 ? 

	We just released GNOME 2.10 ... 0.10 comes after 0.9 :-)

	(That seems perfectly natural to me, but I know others were totally
insistant that 2.10 made no sense and we should have bumped to 3.0 -
perhaps this is a locale/cultural thing?)

Cheers,
Mark.