ive reinstalled xp and now there is very noticeable negative accel. my old setup though bloated with broken software and registry etc, had perfect mouse.
my setup:
g5 setpoint 3.1.116
ati radeon x700 latest drivers (new 'control center' required .Net 2.0)
xp corporate
setpoint- game detection on, keep mouseaccel, keep mousespeed, speed slider on 3 mouseaccel none, dpi 2000, khalmnpr.exe and setpoint.exe in realtime priority
redistributable directx 9.0c (59mb) (i tried to download full (~120mb) from microsoft, but i couldnt validate. maybe they have crippled the free version?)
a whole lot of tweaks and reg hacks, but none that ought to affect qw.

my old setup differences:
setpoint 3
ati radeon older drivers, control center didnt need .Net
xp professional
no .Net
directx 9.0b or c full version

ive tried many different things but im stumped.

also why does ezquake not using independent physics feel so different to fuh.31? its very noticeable

did you lower the sensitivity in setpoint?, if u keep it at the default setting I don't think you'll experience any negative accel. also you could revert back to the old catalyst drivers

yea ezquake feels really "heavy" i don't know what they did to it but it's hard to use.

I had settings identical previously and it was perfect though. Hmmm i think ill reformat. and use my driver cds instead of newest drivers.
This setpoint version (latest 3.10.116)has configuration panel bugs, like buttons disappearing, and theres a non-logitech one on softpedia called setpoint 3.15 32bit, so maybe logitech ruined it. what driver versions and setpoint settings are people using?..........
Thanks squeeze. Id like to know from Tonik and other coders exactly what has changed to make it so different? (not using independent physics)..........blah

tried -dinput vs no -dinput?

-dinput vs no -dinput used to cause a significant (1/2?) change in sensitivity. I needed the combination of -dinput -m_smooth(m_rate 200) -noforcemaccel for my ms intelli ball mouse.
//(-noforcemaccel used to need -dinput, from the ezstart behaviour i deduce this is no longer the case. in fact it means that either
//-noforcemparams etc cmds are irrelevant with -dinput or they cant be used concurrently?)
(current cmdline is -mem 256 -democache 65536 -zone 2048 -conwidth 800 +set s_khz 44 +set cl_confirmquit 0 -sndbits 16 -dinput -noforcemaccel -bpp 32 -width 1024 +cfg_load foez.cfg +set cl_independentPhysics 0 +set vid_displayfrequency 85).
no -dinput makes gamma too dark in ezquake and varying gl_gamma has no affect (vid_hwgammacontrol 0 makes it a bit lighter but gl_gamma is still broken). its something in my cfg though, incidentally a good feature would be cfg_load_linesatatime "5" to execute cfgs in segments to isolate problems.
So has the -dinput code/behaviour changed from fuh .31 to ezquake 1754?
-dinput vs no -dinput still has the negative accel problem in fuhquake although the negative accel is less without -dinput, and sensitivity is increased slightly (not just the affect of less negaccel)

Then use -dinput and try different drivers for your mouse. If it still doesn't fix it, then it's probably all in your head.

And be careful with matches next time.

Thanks for your time Tonik.
I wasn't clear enough(was very tired writing it), this mouse thing isnt an ezquake thing and i wasn't expecting you(or anyone) to have the solution.
Its not pretend 8 ). But, _incidentally_, something that is more subtle and subjective is the far less perceptible milli/microscopic difference between fuh.31 and ez 1754 (its probably very dependent on the computer)
Ezq definitely feels smoother and more continuous (this is ezq without cl_independentphysics 1 ofcourse) so its probably just me favouring the more jerky\erroneous graphics(& inputs maybe) of fuhq because I'm used to it.
From memory you were intimate with Fuhrer and fuhq code (and of course you made zquake), so i was just looking for any quick insight you had to offer like what you took from zquake/fuhq and what you changed/reimplemented/added that might affect graphics and input.