I've actually made several lexers and parsers based on the joe DFA style of parsing. The state and transition syntax was something that I always understood much more easily than the standard tools.

The downside is your rulesets tend to get more verbose and are a little bit harder to structure than they might ideally be in other languages more suited towards the purpose, but I actually think that's an advantage, as it's much easier to reason about every production rule when looking at the code.