X configuration paradigm, and a proposal

[Avi Alkalay]

> In fact, much/most of the configuration should be outside of the X
> server's direct control.
>
> For example, input devices should be handed to the X server by the
> operating system: this is a property of the operating system's desktop
> policy (input devices might be used for other things than the window
> system).  This and hotplug implies that they should not be configured
> into the window system statically at all.


Not sure if I understand it completely.....
I think I got that is a desire to delegate as much as possible to the
OS. But even then some configuration glue must be done, on X level,
say, mouse1 is the real mouse, and mouse2 is only for scrolling.


> One can argue similar things about even graphics hardware and monitor
> resolution, though it isn't quite a clear cut situation.

The OS may say to X: "I have here 3 monitors for you". But it is this
configuration glue the responsible to define that monitor1 is on the
left of monitor2, etc. Correct !?


> So yes, we're planning on major changes in this area indeed...
>
> If you have time, talent and energy to put toward the engineering tasks
> ahead of us in this area, it is a very good area to work in.

I'm interested in this point. What are you guys planning? At least to
test my engineering talent :-)

I'm interested on how to improve this configuration glue. HW discovery
is not really X responsability, but OS's.

Regards,
Avi