I'd believe that you could create a hyper-optimized Sudoku solver that uses bitwise operations to transform a board state into a bitset of possible moves. I just don't think the search space is large enough to really warrant it. And since Norvig was clearly going for a pedagogical approach, I think in his case it would be counterproductive.
I'd believe that you could create a hyper-optimized Sudoku solver that uses bitwise operations to transform a board state into a bitset of possible moves. I just don't think the search space is large enough to really warrant it. And since Norvig was clearly going for a pedagogical approach, I think in his case it would be counterproductive.