Now a week on the 3x+1 problem:
Now with soundfiles or recorded sound:
clicknilsonaug15th2006.rtf (missed rehearsal, extra catchup on Thurs 17th morning)
Primitive roots modulo n
Starting to think more about long term change
Main discoveries (initial thoughts, I haven't looked back through the sessions extensively at this point):
1) It's very tiring, particularly after a day of computer based work. It's not relaxing/alternative stimulation, and for that I miss piano/violin practice, which I forsook while I've been doing this.
2) Quality varies very much, often correlated to what time of day I might have been doing the session, but also based on the theme I choose for the day. One week I challenged myself to explore the 3x+1 problem, and the negotiations with the algorithmic theme probably became more refined as the week went on, though still varying in their musical effectiveness because of the further issue of psychological mapping etc. I found myself choosing between musical output and mathematical thinking. (perhaps there is a way to have one person map while another explores the maths?)
3) If I chose a tough theme, there would be a break at the start of the set while I tried to get some basic functions coded- sometimes 5 minutes before any sound! All the sessions were from scratch live coding, so either a forgiving connoisseur audience, a duet partner, or cover material for the early stages would help in dramatic presentation.
4) However, I could often reach a critical point of complexity and breadth of material where unexpected and interesting results would flow, and there were enough voices to play with to maintain a musical exploration simultaneous to the algorithmic exploration. The Aug 4th session is a good demo of this happier state.
5) I got better at it, by introducing various shortcuts, by having certain synthesis and algo comp tricks in the fingers ready for episodes, and just by sheer repetition. I did try to use jitlib a few times, perhaps successfully on one occasion, though the habits of jitlib actually interfered with my standard non-jitlib habits, so I'd need a month on jitlib alone to cover that properly. I was happier without it (sorry Julian, though I did use Tdef every day!) just in the way that you get used to your own way of working, even though it might not be the best general approach.