Re: Functional definition of a feature freeze.


Subject: Re: Functional definition of a feature freeze.
From: Martin Sevior (msevior@mccubbin.ph.unimelb.edu.au)
Date: Thu Oct 11 2001 - 20:36:07 CDT


On Thu, 11 Oct 2001, Dom Lachowicz wrote:

> Quoting Martin Sevior <msevior@mccubbin.ph.unimelb.edu.au>:
>
> > Hi Folks,
> > In order to get a 1.0 release out sometime we have to decide to
> > make a feature freeze at some point in time. This is a bit tricky for an
> > open source project like ours. We have no particular deadline that is
> > vital to be met. One deadline I would personally like to meet is to have
> > stable 1.0 release for the GNOME 2.0 platform. GNOME 2.0 requires at
> > least
> > Gtk 2.0. The unix hackers are very keen to port abiword to Gtk 2.0. This
> > will be non-trivial in many ways and will require some time.
> >
> >
> > As Tomas has said GTK 2.0 has been "Just about ready" for release for at
> > least 1 year. That being the case I propose that as soon as Gtk 2.0 is
> > released, abiword should go into feature freeze.
> >
> > >From then on all new feature to the abiword 1.0 code base should be
> > plugins. That way we can either include them or leave them out for the
> > actual 1.0 release and more features can be added post 1.0 by means of
> > plugins.
> >
> > This way we can have our cake and eat it too :-)
>
> Hi Martin,
>
> Basing our release deadline on another product is a bad idea, and GTK+2 is an
> especially bad 'other product' to base *our* release upon.
>
> I see no other way than to have Abi 1.0 be based upon GTK+1.2.x, and we should
> not focus our efforts on porting to GTK+2 *besides* our work to finish abi's
> features and close many of the 500 bugs remaining in bugzilla.
>
> I'm suggesting that we work on new features for another 3/2 -> 2 months. There
> are just some features that no plugin architecture will get us, such as your
> new sections features. Plugins are no panacia, and we don't really have a
> full-blown plugin architecture that would enable us to do cool things like
> handle mouse motions and clicks. Right now, we're pretty much limited to
> importers/exporters and dialogs from plugins, and *nothing* more advanced than
> that.
>
> I'm proposing that we change our release numbering to 0.9.99x when we hit our
> feature freeze. I'm suggesting that that take place 1 1/2 -> 2 months from
> now, regardless of other project's deadlines.
>

OK just so we're all clear. We could do a feature freeze NOW if we wanted
and just focuss on fixing bugs until we're happy with the state of abiword
to make a 1.0 release. Apart from "Borders and shading" we're well ahead
of the feature set orignially planned by the abisource guys for 1.0
release. Of course we're also way past their expected 1.0 release date
too.

I'm happy to have another couple of months to add features but I expect
this will also delay 1.0 by a couple of months. This means porting to
GTK 2.0, gnome 2.0, tables, frames and other cool features are also
delayed by this time.

Are there objections to this delay? It doesn't particularly bother me but
if it bothers others please speak up!

Cheers

Martin



This archive was generated by hypermail 2b25 : Thu Oct 11 2001 - 20:36:12 CDT