It would be an interesting day if someone could make regular expression searches over a large data set as fast as keyword or phrase searches against a sorted and indexed database.

They also need to make sure they use an implementation that doesn't suffer from exponential time[1] regular expressions.

[1] - http://swtch.com/~rsc/regexp/regexp1.html

I know this isn't exactly what you meant... http://thor.duckduckgo.com/?q=regexp+/(.*)/+ddg

Edit: Looks like yc cut off my link there--copy and paste the full thing.