glib in kdesupport: yes or no?

[Havoc Pennington]

On Mon, Mar 10, 2003 at 12:46:16AM -0500, Navindra Umanee wrote:
> Havoc Pennington <hp at redhat.com> wrote:
> > Choice vs. Fragmentation
> 
> I seriously think Red Hat did a disservice to choice when they (you)
> decided to make the KDE and GNOME desktops look and feel the same by
> *default* in its distribution.
> 
> The point really is not that the KDE and GNOME desktop environments
> should look, feel and act the same but that GNOME applications should
> work well and integrate properly into a KDE environment and
> vice-versa.

Well, the two points are partly related and partly not.

The reason we wanted them to look and feel the same is that we wanted
to start proactively designing our user experience, and we didn't want
to do it twice. Owen goes into that here:
 http://people.redhat.com/otaylor/rh-desktop.html

My feeling is that if we'd done the whole custom theme, and migrated
to fontconfig, chosen "best of breed" apps, and all that neat stuff
only for GNOME, that people would have felt we were screwing KDE (well
also, my opinion is that we *would* have been screwing KDE). So we did
all the same work for both desktops. For every TODO task filed in
bugzilla, there was one bug for GNOME, one bug for KDE; almost
religiously symmetrical. We put the same amount of work into each.

> So given Red Hat's track record of side-lining core KDE apps by
> default and removing any trace of distinctiveness, it seems like they
> (you) are not at all interested in promoting choice.
> 

If we weren't, we would simply remove one of the desktops and lots of
apps, or something.  Basically that was the discussion we had prior to
8.0. Fragmentation is seen as a problem by customers and internally,
so we were forced to take some action. A lot of people felt that
fragmentation was intractable and we must choose sides. Preston and I
made the counterargument that the interoperability path I'm advocating
would be better. We argued that choosing sides would lose much of the
open source community's expertise, annoy users, and not really help
the Linux community as a whole anyway. We wanted to solve the problem
of docs/specs, benefit from apps no matter which toolkit they used,
and avoid a protracted fragmentation war with other Linux
distributions. We also felt that we could find a way to do work only
once (design the UI, write config tools, etc.)  while still having
stuff integrate reasonably well.

I don't agree that KDE apps are sidelined more so than GNOME apps. The
defaults for the "big three" apps are Mozilla, OpenOffice, and
Evolution. These were arrived at on technical merits. Galeon is
sidelined, as are Gnumeric and AbiWord. If we think OpenOffice is
better than KOffice, it seems to me that defaulting KDE to KOffice and
GNOME to OpenOffice in fact makes KDE look worse than GNOME.

Is the whole thing perfect? Not at all. We have to iterate over
time. However, customer feedback has been very positive, and the whole
thing is just default settings, anyone can change it to be however
they like, the choice is 100% still there.

Of course things did not go well in terms of communication with the
KDE project, which is why I'm now on this list and much more
proactive.

Havoc