Someone in an earlier message suggested that it might be a good idea
to keep a list of various interesting move combinations. My
particular set of algorithms for solving the cube includes one which
flips two side cubies in place, and leaves all the others fixed. The
method I had been using to do this up until now was 26 twists long,
and I think that it was fairly standard -- find a set of moves which
flips one edge in place and randomizes the rest of the faces, then move
another side into the same position and invert it. Others that I
talked to since then have indicated that they used the same method.
Anyway, today I found a 22 quarter-twist version which does the same thing.
I would be interested if anyone has found shorter ones; I am also
interested in seeing the various other primitive transformations that
others use to solve the cube. Using the FBUDRL notation, here is my
-- Tom Davis -------