> As long as we're setting up a new protocol, why not reserve a space for
> tactile feedback from the device?
Sorry, I don't know exactly what you mean (indeed, what are computer games? :)
If `tactile feedback' is a number, it can be encoded as Z: this
is for example what I did with the wacom tablet. If it is
a single bit it can be encoded as an extra button.
Well, if it doesn't fit, these' a `device-specific' protocol extension:
applications can retrieve extra binary data once they know which is the
underlaying physical device (the name/features of the device can be
asked to the kernel module).
The basic idea behind kmouse is that applications reading a pointer
are mainly interested in having information, without too much care
about *how* the information is generated. Otherwise you have
situations like "this program needs that pointer, and that program
needs this pointer". The device-specific hook is to be used when
application and pointer match and extra features can be expoited, but
what if I want to drive a car and I don't have a turning wheel?
My own X is currently reading the tablet as it was a mouse (with no gpm
overhead) and I can handwrite in the xpaint canvas.
I added an almost-formal protocol specification to the kmouse package,
and the wacom tablet is fully supported (has anyone some spare strange
pointer to loan? :).
Kmouse-0.47.tar.gz is available from
tsx-11.mit.edu:/pub/linux/ALPHA/kmouse (currently in incoming...)
iride.unipv.it:/pub/linux
ftp.systemy.it:/pub/develop
Regards
/alessandro
-- __ o La forza dei forti sta nel traversare le traversie con occhio sereno _`\<, (Paperino) __( )/( )__ alessandro.rubini@linux.it +39-382-529554