Delta-based scroll events

[Peter Hutterer]

On Sun, Jul 06, 2008 at 02:00:51PM +0100, Alexander Jones wrote:
> 2008/7/6 Peter Hutterer <peter.hutterer at who-t.net>:
> > true scrolling events are nothing but events on axes other than x/y. the X
> > server doesn't really care what the semantics are (i.e. scrolling). And here
> > is where the actual issue lies, there is no standard. The third axis is
> > generally assumed to be pressure, beyond that it's anyone's guess. IIRC we
> > support up to 32 axes per device, but the main issue is to know which axis
> > does what.
> 
> Right. So if we could define some pair of axes, for the purpose of
> this example 4 and 5, to be scrolling axes, then all I would need to
> do is update the synaptics driver to generate those events and then
> program GTK to interpret them as scrolling? Is that it?
> 
> > and of course - fixing this in X is only one step, you'll then have to go into
> > every toolkit and fix it there as well. It's not an overnight project,
> > unfortunately.
> 
> While it doesn't sound like an overnight project, it doesn't seem like
> /that/ much work to me. Maybe that's just my naivity showing through.
> Seems that half the problem is just defining a standard and getting
> people to agree on it.

yes, the coding part is easy. the people part is the hard part.
that aside, rather than defining two axes for scrolling, it'd much rather
prefer something like axis labelling, so that a device that provides x/y +
scrolling only has 4 axis and not e.g. 8 where 4 of them are dead anyway.

how this labelling is done is the another question of course.
 
Cheers,
  Peter