You could call it a complete failure...
It was meant to be a physical model of a string you could pluck and tune etc. What it turned out to be was a sort of interference generator. Basically, what you hear is interferences caused by asynchronously writing to and reading from an audio buffer, and distortion caused by me not understanding Ulaw sound encoding.
What the pluck, who cares as long as it sounds groovy?
The basic sound comes from the string model, the rest is stunt programming...
start opens the stream
pluck plucks the string
stop closes the stream
base tunes the string
range sets the amplitude / amount of distortion (Ulaw...)
feedback the damper (the higher the value, the longer the sound. set to 16 to get a continuous sound)
for starters you might set feedback to maximum, base to 9 or 10, pluck and sit back...
note: this one's quite heavy on the cpu!