[r-t] Man and machine

[Don Morrison]

Richard Grimmett wrote:
> When I computer generated all possible 7 parts of stedman triples using 
> singles only I found none better than those already put together by RWP 
> in terms of distribution and number of calls.  I am not sure why.  He 
> was very pleased when I told him.

I think many of us have done searches for 12 parts of Grandsire Triples, and 
ended up in awe of Parker's production of his eponymous composition, created so 
long ago. It really is a marvelous, unique optimum in that space of possibilities.

Both regard to it and to the RWP singles only Stedman: in hindsight it is 
perhaps not so surprising, though that in no way diminishes the value of the 
achievement. A human is naturally going to focus on schemes and bits of schemes 
that seem to be producing the most pleasing result, by whatever metric we're 
eventually going to judge it. So when he or she does get to an end product, it 
would seem there's a good chance it's much better than the average of all things 
that are possible. It is still, however, a tremendous achievement to get to such 
an optimum; it is only unsurprising in that the human, if we assume he or she 
has sufficient good taste, will have refused to settle for anything less.



-- 
Don Morrison <dfm at ringing.org>
"To know a thing does not always enable us to prevent it, but at
least the things we know we do hold, if not in our hands, at any
rate in our minds, where we can dispose of them as we choose, and
this gives us the illusion of a sort of power over them."
                 -- Marcel Proust, _À la recherche du temps perdu_,
                    tr. C. K. Scott Moncrieff