Sorry folks, I missed one possibility.
In case 2,my algorithm does not account for the configuration reached by
L R D^2 R' L' F' B' D^2 F B D^2 14qtws
(FDL,RDF,BDR,LDB) => (BDR,LDB,FDL,RDF)
Either add this transform to the list given, or when you get to
this particular configuration execute C3 twice 24qtws, but still inmy
range and still along the general lines of the rest of the possibilities.
Chris Worrell (ZILCH@MIT-MC)