Hi Axel,
2008/2/3, Axel Dörfler <axeld@xxxxxxxxxxxxxxxx>:
> "Niels Reedijk" <niels.reedijk@xxxxxxxxx> wrote:
> > I would like to share with you a list I made of all the steps I think
> > need to be made from now to a release of alpha 1. Before I do this, I
> > would like an informed opinion from you of where we stand. I know we
> > have a 'it is done when it is done' mentality, but without actually
> > defining 'done', we could wait a long time.
>
> There is a number of open bugs that are related to R1/alpha here:
> http://dev.haiku-os.org/query?status=new&status=assigned&status=reopened&milestone=R1%2Falpha
>
> Whatever is missing there should be added in Trac as a central place to
> keep track of that information.
Yes. At the same time, all the bugs currently assigned to this
milestone should be re-evaluated. This sort of leads to a chicken and
egg problem: the timeline of the release depends on the open tasks,
and the number of tasks we can finish depends on the timeline.
You notice that I would very much like everyone to consider setting an
alpha 1 release date. Even though this project historically did not do
much with release dates, I think it might be necessary at this point.
It would get the energies focussed on a single goal. If that is
self-hosting, then I'm all for it. Remember, the 'end-users' are early
adopters, interested developers and testers.
I will compare this to KDE 4 (only once, because it really isn't that
strong). They were continually designing and redesigning their new
frameworks. Some frameworks were almost done, some needed more work,
others were still in the designin stage. The way development was
going, they could have taken years more. Instead, they picked a
release date. This forced them to choose their goal, to let go of
things that didn't work, and to integrate things that do. In the end,
there is an unstable KDE 4, but they are fully aware of what needs to
be done.
> > So the question is, is it time to start planning?
>
> Maybe, but not that formal IMO.
> I think we should see the alpha release process as a first try -
> keeping it simple and seemless should be our first goal. Then learn
> from the mistakes we made for the beta release, and so on.
I agree with simple and seamless, but at the same time I reckon these
things require at least a bit of planning and discussion. The proposed
IRC meeting between devs and Haiku Inc. could work, but remember, the
discussion is first and foremost about how to (or whether to) proceed
with the alpha development. Let's not cloud this discussion with
technical details.
If there is a consensus on doing one, then I think it is time to start
with the discussion on what should be in it (self-hosting, ohci), to
determine a timeframe. But only after the initial 'alpha 1' yes or no
decision this is of importance.
Niels
