On Thu, 7 Dec 2000, Philip Nemec wrote:
> There's fairly good joystick support in Linux [including serial
> joysticks] with a uniform interface (variable number of axes and buttons
> - all scaled/calibrated to give consistency).
Indeed. For what it is worth, I have an early prototype for a replacement
of pfuFlybox that I call pfuJoystick. It handles just Flyboxes on IRIX
and both Flyboxes and regular PC-Joysticks under Linux. It is implemented
at a C++ class that also has a C-wrapper. As such, you can access
multiple joysticks / flyboxes (one per instance of the class). It is also
possible to re-map the axes and buttons on the joystick (for example to
try to emulate a flybox). There is also a ::print() method that'll output
what buttons / axes are being used and what their values are. It would
be nice to add support for more than just Flyboxes under IRIX.
> I'd say that the pfuXformer model is a bit broken and an alternative
> that includes support for joysticks would be worthwhile... It seems to
> be that a matrix mapping axes (including from a mouse) and buttons to
> XYZ should be general enough to support different motion models as well
> as different input devices.
(again, for what it is worth), I have a modified version of the
src/pguide/libpfui/fly.c that uses pfuJoystick and the FlightStick motion
model (I also have a twoJoysticks.c program). The frustration I had with
the motion model was the lack of a throttle (which i blame on pfiXformer).
Instead, it seems, one only has control of acceleration or deceleration.
Which doesn't work too well with a throttle. It would be nice to modify
pfiXformer to accept a throttle-like input.
Replacing pfiXformer with something more generic and flexible (as I
interpret your suggestion to be) sounds quite desirable as well.
well, those are my two cents :)
"Mongooses are famous for their snake-fighting ability, and are
almost always victorious because of their speed, agility, and timing
and also because of their thick coat."