multi-touch-dev team mailing list archive

[Henrik Rydberg <rydberg@xxxxxxxxxxx>]

>

> I believe the plan was to treat single-touch (touch-begin and touch-end)
> gestures as the source of contact information, and the whole
> combinatorial thing can be used to cross-link between N-contact events
> and individual contacts.


Yeah, this sounds sensible, in particular if the actual state can be extracted
at will.

> 
> On second thought, that sounds hard from an application perspective and
> from an implementation perspective.  I'd rather just keep the current
> way of sending contact information as attributes of a gesture event.


In what way would it be hard? The current approach is expensive bandwidth-wise,
and seems less extensible. Referring to the touch state (as in Qt) would make
things quite easy to use, and save both memory and cpu.

> Does this answer your question?


It raised the issue, yes, thanks. :-)