Plasma/Kwin cooperation to achieve efficacy of Activites spatial metaphors

Lucas Murray lmurray at undefinedfire.com
Wed Feb 11 04:33:47 CET 2009


On Wed, Feb 11, 2009 at 12:17 PM, Jud Craft <craftjml at gmail.com> wrote:
> On 2/10/09, Lucas Murray <lmurray at undefinedfire.com> wrote:
>> For the proposal to work in every situation there will need to be two
>> desktop grid/workspace managers: One built into KWin so it can do all
>> the fancy stuff and one built into Plasma so it can be used
>> cross-platform and when using another window manager such as Compiz in
>> KDE.
>
> Sounds like approaching impracticality here.  Plasma was never
> designed from the start to be a flexible environment that plugs into
> all sorts of platforms.  It was designed uniquely to be a KDE
> workspace environment.  It is a KDE pillar.
>
> It has no obligations at all with non-KDE platforms and technologies,
> since this obviously is DIRECTLY competing with providing an
> integrated KDE experience.  You either devote your development power
> toward integrating with Kwin or you do not.  Development time/power
> are a zero-sum game, and Compiz is not part of the KDE platform at
> all.  (I don't hate it, it's just not part of the platform.)
>
> And why, why, why build an abstraction _just_ for the sake of working
> with third-party compositing WMs when KDE already has a more than
> sufficient first-party WM?  You are looking for interoperability where
> it is not necessary, and probably even harmful (to Kwin/Plasma
> integration efforts).
>
> You cannot perfectly integrate with everything.  You must choose.

If you really, really don't mind dropping support for Compiz and
platforms that cannot run KWin then it would make the whole thing
simpler. There would still have have to be two modes to support users
that can't use/don't want compositing, the improvement would be that
everything can be handled by KWin allowing more code reuse.

I still don't see how dropping support for what we already do as a
good idea though, especially when there is relatively little extra
work required (Still some extra work though I do admit) to support
them in the grand scheme of things.


More information about the Plasma-devel mailing list