hmm, yes w some modification it should be possible to do as u describe.
(pretty smart actually!)
But i was looking at that construction on that site and could not figure 
out how the wind-direction was measured... It seems to have a simple
rotary sensor and no other stuff to detect it with.

------
Yes, you can.
In real, the rotating sensor is made of 3 or 4 arms and cups. If the
rotator generates one pulse per arm, and if one of the cups is bigger
than the others, you can measure a small time difference between pulses,
the time from certain pulse to another will be short, this happens when
the bigger cup is being pushed by the wind. If you install another
sensor (one pulse per revolution) in the rotator that identify a know
position (home position), it would be easy to identify in percentage
(from 0 to 100% of a complete rotor revolution), so the direction of the
wind with great accuracy. The only problem is that the rotator would
vibrate, as out of balance, extra adjustments with weights could be
necessary. 

> ok, how the &Û%&Û% does it tell what direction the wind is comming from ?
> >From the rotation of the 'wind-wheel' it can only determine windspeed,
> not direction ?!? (north/west/south/east and so on)