Yeah, increased the friction, and put in thread timing to limit to 50 updates per second. (the world is flat, so no slopes… yet)
Now the friction works quite well (if a bit high)…
the real problem seems to be in the physics library itself - the ball will pick up an unexpected drift. This might be due to lateral friction interference, but it feels really weird.
also - in response to some things you have posted earlier, ODE does do friction (the ball rolls - i give it force, not angular velocity) - but it is in a 'perfect world' - eg there is no energy loss when friction is applied.
…if friction is applied energy must be lost. :-o I think that's a basic rule of physics. 
time to start a flame war (not really)
yes - energy is lost - typically as noise or heat. I agree, energy should be lost… and thats what the callback does - it removes energy.