[r-t] Bobs only Erin Triples (was: Composition search)

[Don Morrison]

On Tue, Nov 24, 2015 at 12:20 PM Alan Burlison <alan.burlison at gmail.com>
wrote:
> That can be done in a hierarchical way, for example if we have
> established that any given sequence is true, we can cache all it's
> ordered sub-sequences as being true. For example, if (A,B,C) is true
> then (A,B) and (B,C) are true.

There may well be something useful here, but I don't think it's going to be
as easy as this example seems to imply. You rarely run into falseness
locally, it's in more distant places in your path through the graph that it
becomes an issue. For example, ANY three, or even four, division sequence
in Erin is guaranteed to be true, and only one of the thirty-two five
division sequences is false. Things probably don't get interesting until
you're up at thirty or forty division sequences, and caching the billions
of them that there are is not going to be practical (and I doubt there's
any useful scheme for flushing most while retaining the most useful; LRU,
for example, probably doesn't help much). Note that in saying this I'm
assuming a one part; if you've modded out by a part head things might get
more interesting earlier.



-- 
Don Morrison <dfm at ringing.org>
"Money is like muck, not good except it be spread."
                  -- Francis Bacon, _Essays (1625)_
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://bellringers.net/pipermail/ringing-theory/attachments/20151124/c6910aec/attachment-0003.html>