[RFC] MPX XInput Protocol extensions - big issues....

Peter Hutterer mailinglists at who-t.net
Mon Mar 3 23:35:03 PST 2008


Jim Gettys wrote:
> How will we associate users with input devices?  We'll certainly be
> dealing with input devices that can identify who is using them (e.g.
> Diamond Touch, remote devices, etc.) 

I don't think this should be part of the X server. I did try to find a 
solution to this problem and it's a bit bigger than expected. ATM, I 
think adding it to ICCCM/EMWH is the proper solution.

The reason is simple: X doesn't really care about the semantics of how 
the devices are used, it'll just send events when appropriate. When you 
introduce a user object, you suddenly have to worry about a lot of other 
things:
- is the user an operating system user? or just a guest?
- can the user even be identified? or can you just differ between users 
(e.g. DiamondTouch can only differ, but not identify)
- If two devices are to be used by the same user, which one is in the 
dominant hand? Is there a concept of a dominant hand? If more than two 
devices are used, the dominant hand may change depending on which 
devices are to be used.

All these are very context-dependent problems and can change, even in 
between applications. The idea I had to sort-of solve this add a user 
object to ICCCM. Assign device ids to the users, assign login names if 
available. The information is there if apps need it, otherwise it 
doesn't matter. ICCCM allows us to be a lot more flexible too.

I don't have a complete solution to this yet. I ran into the roadblocks 
mentioned above and they need sorting out.

But my opinion is that it should not be in the server.

> How will that interact with Eamon's security framework?

Eamon's work has been merged into MPX. there were a few occurances that 
needed sorting out, Eamon helped me with that. But, bar various bugs, it 
does not interfere with the security framework.

As for the user object, see above.

Cheers,
   Peter



More information about the xorg mailing list