[Haskell-cafe] feasability of implementing an awk interpreter.

I'd like the community to give me feedback on the difficulty level of implementing an awk interpreter. What language features would be required? Specifically I'm hoping that TH is not necessary because I'm nowhere near that skill level. An outline of a possible approach would be appreciated. I am using http://www.math.utah.edu/docs/info/gawk_toc.html as a guide to the language description.