This problem is, unfortunately, more difficult that it appears
on the surface, more than we even realized when we designed X11.
It is one of (maybe the) biggest regret I have about X11's design.
Keith Packard and I have been conspiring to work on this problem over
the last year or two. Some of the prerequisites required to make it
feasible are now (getting) deployed in XFree86 4.x, with Keith's
new frame buffer code, and the RandR extension: more work is needed.
I'd recommend looking at Keith's and my Usenix paper of last June to get
a handle on the issues. We hope/intend to be doing serious work on remaining
problems over the next 6 months, and would like to demo this at this
summer's Usenix, if we can. I suspect serious work will start up after
the holidays.
Work is needed in a bunch of areas: e.g. getting RandR implemented in
all servers (it is needed as part of the notification of movement process,
and to inform client toolkits which visuals are accelerated), and toolkit
work (we believe the best way to do this is at the toolkit level, for
reasons explained in the paper). Proxy servers are also useful, but
more limited.
And there will may need to be some futher server hacking, though
we think FB and RandR go along way toward what is needed. Certainly,
some ICCCM sorts of work need to happen.
The vision includes moving apps from a small handheld screen to
a wall of projectors, and to have the "right thing happen", which
may involve apps displaying completely different user interfaces when
migrated: with modern UI interface builders such as glade and the like,
becoming more used, this vision is, we believe feasible.
So if you want to work in this area, there is plenty to do...
- Jim Gettys
--
Jim Gettys
Cambridge Research Laboratory
Compaq Computer Corporation
[EMAIL PROTECTED]
_______________________________________________
Xpert mailing list
[EMAIL PROTECTED]
http://XFree86.Org/mailman/listinfo/xpert
