>could be saved by your approach. You could measure this by hacking Xt
>to log all transactions with the X server.
There's already a program that runs between X client and server and logs
all transactions. (Don't remember its name just now but i can dig it up if
someone is really interested.)
>And does this traffic really matter with modern networks?
Not everyone has modern network. I sometimes run X programs over 14400bps
modem line, but I would do this much more often, if it just would work
faster.
>implementing it as an extension to X would help.
Compatibility to run also X programs is mandatory. At least at source code
level.
>previous message, is whether you see this fatter display server as
>part of the OS (i.e. the kernel) or not. I hope that you don't
>actually propose that such a server be part of the kernel.
But using GGI sounds like a good idea for me.
Actually, there's one major problem with X in my opinion: X server tries to
be both windowing system _and_ a video driver. It would be good if we would
have a windowing system which would be complete having all widgets
(as modules, so that widgets could be replaced with another ones, so
that all programs would start using the new set then) but _not_ being
a video driver.
GGI library in user space would be then the "video driver" and GGI part in
kernel would protect anyone crashing the system/screen.
| Tuukka Toivonen <tuukkat@stekt.oulu.fi>
| Homepage: http://stekt.oulu.fi/~tuukkat/
| Try also finger -l tuukkat@stekt.oulu.fi
| Studying information engineering at the University of Oulu
+-----------------------------------------------------------