Re: alt-tab enhancement

[Elijah Newren <newren gmail com>]

On Sun, 17 Oct 2004 22:12:52 +0200, Jan Nieuwenhuizen <janneke gnu org> wrote:
> Elijah Newren writes:
> 
> > though if we get working implementations of any of these ideas we
> > could then perhaps bring it up on the usability list too.
> 
> Have you tried my keywise.c hack?  It has 86390, 136666, while 102656
> would require a one-line change, and 97725 is on my whishlist too, but
> I just don't know how to code it.

I haven't.  It sounds really cool that it incorporates those other
abilities.  But it looked like it would require manual configuration
to define "group", which looked like something I didn't want to mess
with (and that I doubt most users would either).  I know you've put
lots of work into this, but that's a deal-breaker for me--even with
the other enhancements.  I'd like to be more positive than that,
though, so let me try to suggest some possibilities you can try:

- determine if implementing the enhancements requested in 97725 and
94682 would be enough for you; if so, work on implementing them (note
that 94682 sounds very close to what you are requesting--so much so
that it may suffer from the same grouping problem you do).

- try to make it possible to define "group" without manual user
configuration.  This would probably require application and toolkit
support so that they could provide a hint to the window manager about
which group to place each window in.  This is something that would
need to be discussed on the wm-spec-list.  But, it may be useful for a
number of things--there's a variety of bugs against metacity and
libwnck requesting stuff that needs intelligent grouping (e.g., your
enhancement request, bug 94682, bug 141171, bug 51574, bug 59692, and
bug 97491).  Personally, I'm a little skeptical that grouping could
Just Work, but if someone proposed something that looks reasonable and
covered all the cases I'd be willing to listen.

Hope that helps,
Elijah