After I read the chapter on microsounds in Formalized Music, I posted a few questions, comments and observations:
. . . I object to his characterization of analog electronic music as static. Did he listen to the same people that I did? Was he not aware of oscillator drift?
Well, since the article was written in 1969 or 1970, he had actually not listened to the music that I have, because it had not been made yet. Voltage-controlled synthesizers are certainly not static. But static oscillators are. The problems he described were present at the time.
The solution that Xenakis proposed is something that's called Gendyn. It uses a bunch of probability algorithms to make chaotic waveforms. He made three pieces with this method. They sound almost exactly like data bending, a technique wherein you take data such as a data file or an application program and open it with a sound editing program, such as Sound Hack and play it as if it were a sound file. I think the similarity between his algorithms which are designed to produce chaotic effect and actual data is actually very illustrative of the nature of data. His algorithms can be limited in certain ways so that the output is bounded between a minimum and a maximum. In the same way, data is often bounded. The similarity is striking.
There exists an os 9 application to do gendyn-type synthesis, called gx. Also, there exists a few Synthedefs in Supercollider, coded by the fabulous Nick Collins, who is, iirc, a Wesleyan alum. These synthDefs are called Gendy1, Gendy2 and Gendy3. They have helpfiles and also reference the appropriate pages in Formalized Music.