On Mon, 4 Dec 2000, Stefan Seefeld wrote:
> uhm, wait. The way it is done now suggests that the coordinates as reported
> by an event are just pixels.
No, they bear no relation to pixels or anything else. They happen to
be pixels, when you are using the X target, but they can be anything, really.
No scale is implied, though you can probably figure out the scale if you
are receiving evAbs events rather than evRel. You don't even have to open a
visual along with the input -- so how could they?
> I'm proposing that GGI allows scaling factors to decouple both concepts,
> much in line with the original question about 'accelerating'
> the mouse movement arbitrarily.
Well, considering there is no relationship guarenteed by the API,
they can't be more decoupled then they already are :)
> Event filtering, again, is a complex thing, and needs to be done in the
> context of a 'scene graph', or a set of windows, or whatever concept you
> use in your UI. It's the region manager's task.
Finding a good balance point where an event filter would more of a
use to, rather than a hindrance to, a complex region manager is a
challange, I do admit.
--
Brian
