Re: About GNOME 2.0 - The end of a dream

[Dietmar Maurer <dietmar ximian com>]

Daniel Veillard wrote:

> On Sun, Jun 17, 2001 at 03:21:19PM +0200, Dietmar Maurer wrote:
> > Daniel Veillard wrote:
> >
> > > > The above statement give the impression that we have made some massive code
> > > > duplication, which is simply not true.
> > >
> > >   This is not the point.
> > > Do you agree that we should have a formal process to get a new piece of code
> > > added to the Gnome platform:
> > >    - yes
> > >    - no
> > >
> > >   Don't make convolution around it, it's the real question at stake !
> >
> > I disagree, we do not need an formal process if someone writes a new bonobo Moniker,
> > because I think that we already agreed to use bonobo. So what piece of code your are
> > talking about? We have added nothing to the Gnome platform, only reused the existing
> > PropertyBag interface and Monikers.
>
>   Basically the decision was to remove GConf.
>
> s/get a new piece of code added to/change the design/

And that is simply not true. Yes, we have removed the GConf compile time dependency,
because it is no longer necessary. We have turned it into a dynamic dependency using
monikers. So I consider this mail as totally pointless. Instead of doing something useful
you construct artificial problem. Please notice that our change has absolutely zero
influence to any application. It is only a change in how we access configuration inside
libgnome.

> > Such decisions can be taken by the people actually working on the code. And
> > everybody is able to see what those people are doing, because we use CVS. We can
> > even revert changes if someone made a bad decision, due to the use of CVS ;-)
>
>   I will take a similar example to show you that is not that simple....
>
>   Okay, since libxml is in CVS you know where it had been going since libxml1
> was forked. Good, I know at least one person won't have troubles switching to
> libxml2. And since those are discussed in the xml gnome org (of course I know
> nobody from Ximian/Redhat/Eazel ever subscribed to this list, except Jacob
> when he was still using his university address) I can argue that you were all
> aware where libxml2 was going for all this time, and protest if anyone
> argue that something has changed.
>
>   I'm a coder too, and I also went through a lot of process in my previous
> job, so I know how much it can be a pain to follow a process. But stating
> before hand where development is heading, how it is supposed to be achieved,
> how long it is supposed to take, and negociating if necessary at the planning
> stage with the other projects depending on it is necessary if we want to go
> forward without loosing people, focus or the community sense.
>
> > Using a formal process only removes the responsibility from the maintainer,
>
>   No it makes him responsible for following the plan, i.e. give him more
> weight actually. If the gconf change had been planned before hand then Martin
> very simple answer would have been "we decided this, this was the plan",
> and he would have been in a far better position to defend himself.
> In fact nobody would have objected at that point, and this accident would
> not have happened.
>
> > and I
> > think this will block the whole development.
>
>   Maybe your experience and my experience of large number of cooperating
> teams working toward a given direction are different.
>
> Daniel
>
> --
> Daniel Veillard      | Red Hat Network http://redhat.com/products/network/
> veillard redhat com  | libxml Gnome XML XSLT toolkit  http://xmlsoft.org/
> http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/
> Sep 17-18 2001 Brussels Red Hat TechWorld http://www.redhat-techworld.com