The easiest way to solve a sudoku is to map it to an exact cover problem and then apply an exact cover solver, which for sudokus that have one solution is equal to finding colums that are a strict superset of another column and removing the rows that have a 1 in the superset and a 0 in the other column. You need far less code for describing this than the code presented in the article.
The easiest way to solve a sudoku is to map it to an exact cover problem and then apply an exact cover solver, which for sudokus that have one solution is equal to finding colums that are a strict superset of another column and removing the rows that have a 1 in the superset and a 0 in the other column. You need far less code for describing this than the code presented in the article.
The article isn’t about solving Sudoku, though. It’s about using functional programming to design hardware. Sudoku is only used as an example.