Input device design (3)

Jim Gettys jg at
Fri Sep 2 13:54:00 PDT 2005

Certainly the computation I did a day or two ago for wireless links
convinced me we need a compact representation; even an X event structure
represents more than a millisecond on a 802.15.4 network.  

The problem with the X event structure definitions is that they get very
klunky for devices with more valuators.  Extensibility is also a

But whatever we do has to be pretty compact.
			- Jim

On Fri, 2005-09-02 at 15:36 -0400, Joe Krahn wrote:
> James Cloos wrote:
> >  >>>>> "JimG" == Jim Gettys <jg at> writes:
> > 
> > JimG> One of the nice things about XML formats is that this sort of
> > JimG> encapsulation of such information is possible, with a section X
> > JimG> understands, and something private to both client and input
> > JimG> device.
> > 
> > If xml is chosen, there is a binary format that is a one-to-one and
> > onto mapping of xml.  Using that on the wire/over the air may provide
> > enough ‘compression’ to alleviate any latency issues.
> > 
> > For reducing verbosity in the text representation, there is a sexp
> > mapping for xml, as well as json.
> > 
> > -JimC
> The more I think about it, the more I think XML and variants is not so 
> useful. Any information we want to pass will necessarily have to get 
> encapsulated into the X protocol, so binary packet definitions will have 
>   to be defined, where needed, anyhow. Why not use the same binary packets?
> Or, is there an idea of embedded XML strings into text Device 
> Properties, which are handled similar to Window Properties?
> Joe
> _______________________________________________
> xorg mailing list
> xorg at

More information about the xorg mailing list