Apart from the parsing tables only one character per parser state is required to control the correction process. The method is applicable to a wide class of stack automata including LL(k), LR(k), SLR(k) , and LALR(k) parsers. It is shown that for LL(k) grammars error correction can...