🚀 go-pugleaf

In article <1992Oct7.165306.29274@m.cs.uiuc.edu> carroll@cs.uiuc.edu (Alan M. Carroll) wrote:
> In article <JWZ.92Oct6210542@thalidomide.cs.cmu.edu>, Jamie Zawinski <jwz@cs.cmu.edu> writes:
>> Clearly the right thing would be to add a minimal number of WM primitives
>> to the emacs C code, and start your window manager as "emacs -q -l wm.elc".
> 
> I must strongly disagree with this. It's hard enough to handle the X
> events Epoch already deals with. It'd be a nightmare to add real WM
> capabilities.

Well, I was mostly joking (mostly.)  If it's harder to handle events in Epoch
than in GWM, then that just means that GWM has a better design for event
handling, right?  They're both lisps.

The reason I think that such a thing could be reasonable is that emacs-lisp
is pretty close to common lisp, reasonably efficient for simple things, and
comes with a pretty good development environment.  And, in emacs, most of the
work is done.  I think that writing a new lisp (with a brand new idiosyncratic
syntax, no less) was a pretty big mistake.

> I think that the GWM approach, while perhaps not the best in implementation,
> is the superior solution in the long term. What you can already do is
> communicate in a high-bandwith way with GWM from Epoch and make requests of
> the window manager.

I'm just talking about a customizable, extensible window manager.  (I think
that the way that one can offload the work of popping up menus, etc to the
window manager is a bad idea, but that's a different matter.)

	-- Jamie