Adrian and Shlomi are doing very well without my interference. They'll be
even more productive when they (we all) can agree on a mutual vocabulary or
understand what the others mean by their own terms.
Try this notation:
pA > pB > pC
means position B was created from position A by applying one of the legal
moves of position A to position A (and the same for position C wrt position
B). This is Shlomi's "recurse."
pA > pB > pC < pB means one of two things:
1. Pos C has been reached for the first time, but pos C has no legal moves.
2. Pos C has been encountered before and will not be explored again if it is
reached from above in the position tree.
In both cases the automated solver backtracked to the position from which
position C was created.
pA > pB > pC < pB > pD means
pos B was created from pos A, and pos C was created from pos B, but pos C has
no remaining legal moves or has been seen before. So the solver backtracked
to pos B from which pos C was created, applied the next legal move in pos B
to pos B itself, which created pos D. If pos D has no remaining legal moves
or has been seen before, the sequence would be
pA > pB > pC < pB > pD < pB > pE.
If pE cannot be explored and pB had only 3 legal moves when it was first
seen, then
pA > pB > pC < pB > pD < pB > pE < pB < pA.
I've used pA > pB > pC. Some will suggest A > B > C is enough. I'm being
cautious because later we may need A, B and C in other freecell contexts,
such as mA, mB and mC or tA, tB and tC. Hmmmm. Presumptuousness seems to be
rampant here.
//----------------------------------------------------------------------------
---------------------------
Adrian and I had a brief discussion about his (or Woods' or Keller's) word
"intractable." I don't like that word because it means "resistant" or
"difficult," both of which are inaccurate and weak. I prefer "unmanageable,"
which in this case would mean "with this solver running on this computer,
there are not enough resources to decide the question."
//----------------------------------------------------------------------------
----------------------------
I have put up a straw man if noone ever takes a position from an automated
solver and uses it as a starting position in a "virgin" solver.
more today...
Bill Raymond
Received on Tue Jul 02 2002 - 10:35:41 IDT