> Parsers that work on simple input are easy. Parsers that get all the edge cases and diabolical input right are stupidly difficult.
Not to mention parsers which work on a vague and totally unspecified grammar - like "real world" HTTP and HTML parsers, which have to account for misbehaving clients and servers or malformed web pages (is this what you meant by 'diabolical input'?).
Not to mention parsers which work on a vague and totally unspecified grammar - like "real world" HTTP and HTML parsers, which have to account for misbehaving clients and servers or malformed web pages (is this what you meant by 'diabolical input'?).