[r-t] Man and machine

[Mark Davies]

Simon Humphrey has sent me this very interesting question:

"What's really interesting is how on earth does the composer's brain (well,
some composers' brains, anyway) manage to eliminate vast numbers of node
searches.  Which it must do, to produce results in such relatively short
times.  Intuition, perhaps, but what does that mean? If it's something to do
with pattern matching it ought to be codable !"

I tried to come up with an answer, which I've given below. I suspect others
may have further thoughts - let's hear them.

=================================================

I think a composer's brain eliminates nodes, as you say. There are two ways
I can think of that a human can do this, which computers currently don't.
One is to make use of experience: for instance known patterns of calls and
coursing orders that work and give good music. A human composer will have
built up a palette of such things, which can be applied to new situations in
a way that a computer is unable to do.

Another is simply the rejection of bad nodes which contain no music. A
computer will plough through lots of unmusical nodes, in the hope that
eventually the good stuff will appear, making the composition good overall.
A human is more likely to abandon a line of inquiry if after a couple of
coursing order changes he has not got anywhere good. The human is going to
miss stuff with this kind of search, but often that won't matter, as good
compositions will still be found.

I think a computer search could be designed to make use of the latter
technique - it's a type of music pruning. It struck me after discovering my
"no duffers" Bristol Major composition that a computer could have found this
very quickly, if it adopted a pruning method where branches of the search
tree were abandoned if no music had been produced after say 3 nodes.
Generally this technique will miss good compositions, but the compositions
that it does produce will have the benefit of a good distribution of music
and no long "duffer" stretches. This is definitely the next pruning
optimisation technique that I intend to program.

However! I think in general that it is a good thing that humans and
computers differ. The best composer is not a human nor a computer, but the
combination of the two. They provide different abilities:

The human

1. The ideas - what music we want, what method, what plan of composition.
2. The intuition and experience - these calls will be good, this part end is
desirable, this block is an essential part of the finished touch.

The computer

1. The accuracy - the ability to exhaust a search space and not miss 
anything.
2. The speed - the ability to explore big ideas very quickly.

The creativity of the human and the brute force of the computer is a perfect
partnership. Again my 5056 Bristol Major provides a good example. Mine was
the idea: few calls and short courses to give equal weight to front and back
and to move the bells around quickly, plus an emphasis on LB music and
5678/8765. I created some of the core blocks of the composition, for
instance the palindromic "super-optimal" block. The computer linked the
blocks together for me in the optimal way, using an exhaustive search.

Another excellent example of partnership between man and machine was I 
believe the production of bobs-only Stedman.

=================================================

MBD