Re: desktop-devel-list Digest, Vol 24, Issue 43

[Dave Neary <dneary free fr>]

Hi,

Gustavo J. A. M. Carneiro wrote:
>   And calling gtk+ 3.0 for marketing could backfire.  Developers will
> realise gtk+ 3.0 is basically just the same as the latest 2.x and thus
> be disappointed.

This is an unfounded argument. With a 6 month release cycle, the next
release is *never* going to be anything more that "the same as the
latest 2.x" with some extra stuff. Firefox 2.0 is not going to be
anything more that Firefox 1.5 with some extra stuff, Firefox 1.5 was
Firefox 1.0 with some extra stuff. But the version numbers evolve anyway.

Here's your yardstick: Look at GNOME 2.0. Now look at the latest GNOME
release. Are these releases radically different? I'd say yes - we've
improved the user experience with better integration, automatic
detection of a bunch of devices with DBus and HAL, and a load of core
applications have either been dramatically changed or replaced (Evince,
Nautilus, Ekiga...).

So here's the question - are we sending a message to people by sticking
with a 2.x version? Would we be sending a message with 3.x? Is there
anything wrong with bumping a major version number on an incremental
release, if that increment means we're miles away from the .à release?

To those questions, I'd say yes, yes, no. The message we send sticking
with 2.x is a project that isn't changing very fast, the message we'd
send with a 3.0 release number is of greater change, and no, I don't
think we need radical change to allow ourselves to recognise that the
desktop has come very far since 2.0 came out in 2002.

That said, I'm not hugely passionate about version numbers, I have a
tendency to think that bumping major version numbers when we get a
little tired of the old one is a good idea, but between 2.16 and 3.0,
I'm happy either way. It's not like it'll change the applications if we
bump.

Cheers,
Dave.

-- 
David Neary
bolsh gimp org