[crawdog]

You could build a bloom filter with the profane words, then check your tokens for the page against that filter.

[teraflop]

That would make sense if your list of expletives was too large to fit in memory, which would be... impressive, to say the least.

[GordonS]

I'm Scottish, and I'll say it's definitely possible :p

[Keyframe]

if profanities were single words only. What if they weren’t? You’d have to have a giant list of permutations and build a huge-ass bloom filter. Still doable though, but then spelling errors (or not)...