releasing.

Father…ooh…oh…oh…
I see the world, feel the chill
Which way to go, windowsill
I see the worlds on a rocking horse of time
I see the verse in the rain
Ohh…ohh…ohh…ohh…

Pearl Jam – “Release”

Mark Shuttleworth talked about what he thinks is a sensible
release rhythm. While you probably all know his proposal, here’s what I’m thinking about
releasing every six months (or in regular intervals anyway) and synchronising our releases with
the release of other Free Software communities.

Right now, we can say very little about KDE4. We will have to see how KDE4 pans out. My expectation is that it’ll
prove to be a reliable and stable platform that will suffer from little regressions after 4.0 is out. That means that
we should be able to put out a snapshot every so often and use some weeks (6, 8 maybe?) to stabilise it. Time-based
releases are possible in this scenario, but we should ask ourselves if we want to release on a regular basis (and in some cases
postpone shipping some new feature because it would’ve needed three more weeks to finish and polish).

On the other hand, KDE3 releases were very reliable. If I recall correctly, we only slipped once or twice slightly due to a
security problem (which would affect time-bases releases in the same way). So practically, we have time-based
releases for some time already.

What we can do to make it easier for vendors to ship the next version of our software is making it clear what needs to be done
before a release. That would mean: “Here’s a list with what the next release will look like, if you want to make it happen
earlier, help us hacking” This could be as easy as recording this in bugzilla, marking it as “showstopper” or “required
for x.y.z” and then make the query showing those entries easily accessible through, for example, a direct link.

As to a pulse and having Free Software communities release at the same point in time would
— from a marketing point of view —
be a very bad idea. It would mean two things: Press coverage for Free Software only once in a while,
and smaller projects being snowed
under by those that are doing better PR work. I’m actually wondering why this is brought up by Mark since Canonical’s
umbrella branding fails miserably in this respect — measured by the press coverage that a Kubuntu release has got during the
last cycles.