“If a piece of calculation leads into an ever-denser thicket, nature probably did not intend you to go that way. Try a different approach.”
David P. Stern: “All I really need to know”
Most VM decipherment attempts can claim initial success, but then begin to stall, and get bogged down. To make any more progress at all, ad hoc rules are then invented which add further stages, exceptions or modifications to the originally supposed scheme. As we all know, this can result in ludicrously complex fabrications.
At this point I usually ask myself: If the algorithm you suggest is really that complex, how could you ever claim initial success with your simple startup version? And if you couldn’t, how could you deduce the ever more complicated steps from a simple start which would never have worked?