As the name suggests you lose context, you can't "remember" anything you've seen in the token stream. So you can't really parse languages that use indentation.

As the name suggests you lose context, you can't "remember" anything you've seen in the token stream. So you can't really parse languages that use indentation. Although you can circumvent that by processing indentation in the lexer or inbetween lexer and parser.