There was nothing inherently wrong with using an optical encoder but rather that the pot was probably a much simpler solution once we were aware of the application.

Also the issue of optical encoders oscillating would not have been a major issue as it normally occurs when you have servo systems where there is mechanical and electrical noise from the motor and mechanical systems and where the encoder is held in a particular position. This application had a known maximum angular velocity and was not held in a position for a great amount of time - the sampling rate and software decode would have negated most of any high frequency oscillation that may have been present. If you play with fire -expect to be burned, if there is no fire the likelyhood of getting burned is somwhat diminished!