pointer (touchpad) aceleration and threshold too low in 11.3.* builds

[Paul Fox]

eduardo h. silva wrote:
 > 2012/5/17 Paul Fox <pgf at laptop.org>:
 > > eduardo h. silva wrote:
 > >  > xset q tells that in 11.3.1, the pointer is configured with the values:
 > >  >
 > >  > acceleration 7/4 threshold: 1
 > >  >
 > >  > This takes 2, 3 and sometimes 4 swipes to move across the screen. It's
 > >  > quite a difference with a typical touchpad on bigger laptops, and
 > >  > although the experience can't be duplicated because we can't expect
 > >  > kids to be already highly trained with a touchpad, I think it could be
 > >  > improved.
 > >  >
 > >  > I think the balance is to have a larger threshold, so that slow
 > >  > movements are allways acurate, and a higher acceleration, so that
 > >  > movements across the screen are sufficiently fast (very important in
 > >  > Sugar to access the frame for example). Managing to drop down the
 > >  > amount of swipes needed to move across the screen (like 2 swipes)
 > >  > would be excellent and more in par with the older pre 11.3.* behavior.
 > >
 > > so you think it was better, previously?
 > >
 > > i wonder what has changed.  those xset values were chosen specifically
 > > to be appropriate for the original ALPS touchpad on XO-1.  it's
 > > entirely possible, now that we're two laptops and at least as many
 > > touchpads beyond that, those numbers are incorrect.
 > 
 > I meant to compare between the setting of the builds back in 2008 and
 > 2011 when I began using the XO after a hiatus of years.
 > 
 > I found a discussion started about the change from "acceleration 7/4
 > threshold: 1" to "acceleration 7/4 threshold: 0)
 > (http://lists.laptop.org/pipermail/devel/2010-January/027245.html ). A
 > few messages later, it's shown in xset man page:
 > 
 > "If the `threshold' parameter is provided and 0, the
 >     `acceleration' parameter will be used in the exponent of a
 >     more natural and continous formula, giving precise control
 >     for slow motion but big reach for fast motion, and a
 >     progresive transition for motions in between.  Recommended
 >     `acceleration' value in this case is 3/2 to 2, but not
 >     limited to that range "
 > 
 > Which is exactly what I was looking for. A bit later, Richard A. Smith tells:
 > 
 > "2 of the apps that our deployments spend a lot of time in are scratch
 > and etoys. Both of these are not completely sugarized apps  and thus
 > some of the UI elements on the 200 dpi screen are quite small and hard
 > to use with a fast accel." Though I don't know if he used higher
 > acceleration values, or with threshold set to 0.
 > 
 > So the thread ends with:
 > 
 > "Find someone with kids and figure out whats the best settings for them."
 > 
 > Next time I am with my nephews I'll be sure to do this, but I already
 > find threshold 0 a lot better. Perhaps the problem Richard was having
 > could be solved with a lower acceleration, like 3/2, but still with
 > threshold 0.

i'm sure you're right, and i think we had consensus around that
several years ago.  it's my belief (after skimming old mail) that the
very earliest X servers shipped on XO-1 a) did treat a threshold of 0
as special, but b) did not do it correctly.  so using '1' at that time
was deliberate.  but after a time we (i, certainly) agreed that we
should be using 0.

we also had a discussion about lowering the 7/4 -- at the time i thought
3/2 was too slow, and suggested 165/100, but i suspect that that's
splitting hairs.  (i.e. 1.5 vs. 1.65).

after verifying, we should at least fix the '0' threshold as a bug, in
12.1.0.

paul
=---------------------
 paul fox, pgf at laptop.org