After reading Zeldman's post regarding the W3C validator, I decided to check out the validator's source.
Ouch.
A quick glance shows WAY too much HTML mixed in with perl -- even some XML! Sucking some of that content out via the templating package du jour would be an improvement. Besides, the file is 3000+ lines long, perhaps some refactoring is in order? =D
However - it works! ...and that's always nice.
It also helps that there's an article covering the implementation details. It would have helped me more if I'd have realised that it was there...
-Dom