Re: WM SPec purpose (was: Re: wms offsetting XMoveWindow() coords)

[Paul Warren <pdw ferret lmh ox ac uk>]

On Wed, 29 Sep 1999, Nathan Clemons wrote:
> On Wed, 29 Sep 1999, Paul Warren wrote:

> > I think that the decision to use CORBA or hints should be a purely
> > technical one.  [etc.]
> 
> We have two contradictory considerations here: technically superior
> ideology and acceptance. If it is technically superior but is not
> accepted, then this spec has _failed_. If it is accepted but is not as
> technically well-done as it could be, we have gathered people under a
> banner that can later have revisions incorporated into it. Most window
> managers do not include CORBA, and we should not assume that any window
> manager authors want to support it. As the case is, as good as CORBA may
> be, I don't see but very few of the people on this list advocating it, and
> several of the authors decrying it's use.

I think that one thing that we have discovered with Gnome's "use whatever
WM you want" philosophy is that most existing standalone Window Managers
do not fit in well with GNOME/KDE.  The feeling that I get is that there
is now a significant number of WM hackers out there, desperate to write a
WM that Really Fits Well with Gnome, and will implement pretty much
whatever we come up with in this spec.  I am more concerned that we should
come up with a spec that is technically brilliant, than ensuring that its
simple enough to fit in with the plans of standalone window managers.

I would much prefer to have 3 spec-compliant WMs whose raison d'etre is to
work properly in a Desktop Environment than a dozen standalones that "just
happen to support" the spec.  And, in my eyes the former would be a
success for the spec.

> Frankly, I don't think CORBA is an acceptable option to the populace. WM
> Authors only may correct me if they feel otherwise. We don't need to
> convince either GNOME or KDE, we need to convince the WM authors, and in
> my opinion that is (*at least currently*) a lost battle.

FWIW, I am a WM hacker, and if this spec says "use CORBA", I'll go
away, learn it, and hack it into whatever WM project I end up doing
(and I'm rapidly reaching the conclusion that the only WM that I'll
ever be truely happy with is the one I write myself ;-)...

Technically, I don't know enough about CORBA to say whether we should use
it or not. I am more concerned that we should be aiming this spec for the
right target.

[WM Config tools]

> In my opinion, and I do not necessarily feel the spec should cover this, I
> still believe we should have something along the lines of a GNOME .desktop
> file for WM's that specify the name of the WM, the wm executeable (with
> full path), the current version, compliancy to the spec, and the wm config
> executeable (with full path, or NONE). 

Sounds good enough to me.  And I see no reason why that shouldn't go in
the spec.  Call it "optional", or a "recommendation", or whatever you
want. It is an essential ingredient to making the window manager work
properly with Gnome/KDE, and such behaviour *has* to be defined somewhere.

Declaring your config util, spec compliance etc. is an interaction between
Window Manager and Desktop Environment.  What is this spec for if not for
defining these interactions?

yours,

Paul