[r-t] Predicting search size (was FKM)

[Don Morrison]

Mark Davies wrote:
> 2. Step through the current composition node by node.

I had to stare at this for a minute before I could make sense of it, thinking 
"what on earth does he mean 'current composition'?"

It reflects  how different your world is than mine. I don't normally expect to 
have a "current composition." Depending upon how I've set up my problem I may 
come up with no compositions at all, even after several hours of search. I guess 
you always generate lots and lots of them, and sort them out afterwards. I, on 
the other hand, am trying to get as many unproductive paths to terminate as 
early as possible, so that f and when finally I do get a composition, it's as 
close to The One I Want as possible.



-- 
Don Morrison <dfm at ringing.org>
"I like to whine. It relaxes me and I ring better."
                  -- author's anonymity preserved to preclude reprisal