Well apparently no one reads the thread before replying.

I think pseudo-multi touch would work but it would have several limitations. First of all, I am going to assume the software works like this:

1) default is single touch mode, remembering the last known position while the screen is being touched

2) if it detects a sudden change in location (faster than humanly possible to move your finger) then it will deduce that a second finger must have been used and the average value of the two touches is being "sensed". The two limitations here are that:

a) you have to put one finger before the other and
b) you have to put the fingers further apart than whatever distance is determined to be "too far" to move for a single finger in that time period

Once the software has determined that two fingers are being used, it is then going to switch to multi-touch mode.

3) If it detects another sudden change (2nd finger is lifted) then it will switch back to single touch mode. [I don't think 3-finger-touch will be possible]

As for the multi-touch mode, the software takes the currently detected location and, on the assumption that it is the average of two points, it can then simulate multi-touch. Obviously we cannot get two unknowns from just one equation so we must give it one fixed point. This means that the limitation for multi-touch gestures is that one finger (probably more intuitive to use the first finger which touched the screen) must not move.


So, to summarize:
I) Limited to two-touch
II) Must put one finger before the second
III) One finger must be stationary

This would still be useful for multi-touch gestures and could simulate the "pinch-zoom" everyone loves so much... just need to keep one finger still and pinch the other.