Mouse acceleration under Plan 9

[Brian Vito]

If you'll excuse another question from a non-programmer, does Plan 9
use a mouse pointer acceleration algorithm (if so, where would it be
documented?) or is it 1:1? Thanks so much.

[Rudolf Sykora]

Post by Brian Vito
If you'll excuse another question from a non-programmer, does Plan 9
use a mouse pointer acceleration algorithm (if so, where would it be
documented?) or is it 1:1? Thanks so much.

Having read the section 12.3 of `Introduction to Operating systems
Abstractions Using Plan9 from Bell Labs' by F J Ballesteros, I believe
there is some acceleration available.

Ruda

[David du Colombier]

Post by Brian Vito
If you'll excuse another question from a non-programmer, does Plan 9
use a mouse pointer acceleration algorithm (if so, where would it be
documented?) or is it 1:1? Thanks so much.

Yes. See mouse(3) and /sys/src/9/port/devmouse.c.

[erik quanstrom]

Post by David du Colombier

Post by Brian Vito
If you'll excuse another question from a non-programmer, does Plan 9
use a mouse pointer acceleration algorithm (if so, where would it be
documented?) or is it 1:1? Thanks so much.

Yes. See mouse(3) and /sys/src/9/port/devmouse.c.

however, i've found /hardware/ mouse acceleration to be a mistake.
at least some kvm eat or mangle these commands, which motiviated
me to consider the issue, and i don't see a reason why the kernel
shouldn't just do the acceleration (multiplication) itself. surely one
multiply should be of no concern on a 100Mhz+ computer.

:-)

the code's in the 9atom kernel, fwiw.

- erik

[c***@gmx.de]

probably more a resolution thing than the cost of a
multiplication.

kernel gets only deltas at some fixed sample intervals
where the mouse can do it for every encoder cycle.
(no clue how its really done or with optical mice, touchpads
or trackpoints...)

but kernels software mouse acceleration works good
enougth, so we also have it enabled by default instead
of using the hardware one (which caused problems with
some mice on laptops).

--
cinap

[erik quanstrom]

probably more a resolution thing than the cost of a
multiplication.

the multiplication is not linear. cf. dztab, so there is no
resolution tradeoff for slow/small movements.

- erik