Re: [Enigma-devel] CVS update and mouse controls

[Daniel Heck]

Jeremy Sawicki wrote:

> At 02:23 PM 1/6/03 +0100, Daniel Heck wrote:
> >I tried it. Frankly: it feels horrible :-) I'm not sure how to
> >describe it... maybe "unbalanced" is the right word.  Either the
> >marble accelerates too much, even when you move the mouse only
> >slightly, or it doesn't move at all.
> 
> I don't understand -- you mean it feels horrible with my patch?  The
> patch wasn't intended to affect the behavior at all.  It was only
> meant to affect how you set the options to get a given behavior.  

Mhh, somehow it *did* feel horrible with your patch. I probably didn't
find the right parameters then. But I took your suggestion to heart and
removed the mouse damping altogether.  It may still need some fine
tuning, but in general I really like the result.  What's even better,
the OS/CPU dependence of the previous scheme seems to be gone.

What does the rest of you think, does the marble movement "feel" ok? Can
we keep it that way?

> So you can see that multiplying the mouse speed setting by a given
> amount or dividing the damping by a given amount will have a similar
> impact on the mouse sensitivity.  Changing the damping also has other
> effects (which are more subtle in my opinion).  

Indeed they are, or rather were.  The whole algorithm depended in subtle
ways on the size of the time step chosen in the physics code, and since
that adapted to the speed of the marble, the result was a little
unpredictable, depended in subtle ways on the granularity of the system
scheduler (I think that's why it felt different on Windows and Linux)
and things like that.

Daniel