[r-t] Big searches

[Mark Davies]

Don writes,

> your statement that you don't care about what the part ends are until
> you're done surprises me.

Yes, that's just a result of using the rotationally-sorted search. As well
as getting the rotations only once, you can do general multipart searches
very easily. But, you don't get the node-collapsing optimisation of a 
fixed-part-end search.

Regarding the number of compositions, I've only looked at the very first one
you quote Don:

> 5088 (14235678): BW3H W BWH B2H WH

It's certainly not in my search results, but Ringing Master tells me that
it's false, too. Might be worth checking that out!

MBD