In the article you say “although I’m still not sure what’s intrinsically special about functional languages that would make them more appropriate for combinators.”

The answer is they’re very easy to see in purely functional terms. A parser is a function that converts a stream of characters or tokens into a list of possible parses. Each possible parse is something like a parse tree plus the remainder of the stream that wasn’t parsed. A parser combinator is a higher order function that takes one such parser function and applies a function to the parser to produce a second, more complex, parser.

Monads happen to be a really easy way to build just such higher order functions and they come straight out of the functional world. As coincidence would have it, my next article on monads will demonstrate how a parser combinator library is built under the covers in Scala and I’ll use arithmetic as my example.

Now, higher order functions can be emulated in straight OO languages by using classs. It just might be a bit clunkier. JParsec is in fact a monadic parser combinator library for everybody’s favorite J language: http://jparsec.codehaus.org/.

Two other interesting links on OO and parser combinators come from Gilad Bracha, formerly of Java fame but now more into Smalltalk and derivatives. Here he gives an overview of the OO view of parser combinators: http://gbracha.blogspot.com/2007/01/parser-combinators.html.

Here he talks about a distinctively OO way of reusing one base parser to produce different kinds of output: http://bracha.org/executableGrammars.pdf. Unfortunately, he concludes that the technique isn’t very compatible with static typing.

This tutorial is not for those completely new to compiler development: although I give a brief high level description of compiler internals, if you don’t understand what a “lexer” is before reading this article you’ll get very, very lost and I’d suggest you do more reading. I’ll update the prerequisites to reflect this.

Might I suggest that you stick a few links in the article to more in-depth definitions of some key things that you briefly define (such as ASTs, Lexical Analysis, etc), even if they’re just links to wikipedia? I know that in my travels I’ve found it very helpful to have those links handy while reading things that I don’t have a good knowlege of. Having them there increases the chance that people will make an effort to read up on it if they don’t quite get it.

As I said, great write up. I enjoyed reading