Oh my.
I’m in an OSCON talk about the Parrot Compiler Toolkit. It is just way too cool and way too right for a compiler class. There’s support so that all you really have to write is the parser and code to tag action methods. Those are the big things I want my students to do. So I really have to figure out the PCT.
Now, someone might say instead I need to figure out yacc, since that’s been around for forever and does the same thing, but I just won’t listen to those people. In fact, since you can get to all the parts of the compiler (like the OPcode tree, or POST), I think I could justify PCT being better.
And anyway, yacc’s been around for forever. Isn’t that sufficient reason to try something new?
Tres cool sidebar…I entered all my sessions in Google calendar (very easy, since Google imported the calendar file created by O’Reilly) and I just got warned of my next session. So gotta go.
But I’ve saved the link for my poor compiler kids! (Really, pity them not…this has got to be easier than compiling to .NET!)
Laurie White's Blog 
Nathan Moore said
There’s a whole slew of parser generators out there. Some of them combine the tokenizer in the same source language, which intuitively I think would allow the FSA and PDA could be combined, possibly yielding faster processing and the ability to deal with things like >> within a C++ template where , >>, and << are meant to be brackets rather than compare and shifts.
When there’s a parser generator that really helps to create actions and recovery for error states let me know.