>Both systems are complex enough that essentially no one completely understands >them.
this touches on an important point. the first introduction of regular expressions to editors was great, because it took some formal language theory and made it useful in an `every day' way. conversely, the theory provided significant underpinning (in a structural sense) to that practical application. now there are big books on `regular expressions' mainly because they are no longer regular but a big collection of ad-hoc additions. (i think there is a similarity to Perl's relationship to the history of programming languages.) the result really is hard to reason about ... because it hasn't got that underpinning and the algebra has gone. it's worse than that: the code is a mess too, i supposed in consequence.
