Thanks Frank.

And, no, my boss is not a DB expert (but he unfortunately has dealt with DBs and thus has some knowledge which makes him dangerous).

As for performance, he would rather trade headache for better performance, so if he sees anywhere that multiple joins would be slower, than he will insist on removing as many joins as possible and just adding extra columns instead.

Also, it's only about 50k records now, but that will probably be 100k within a few months and several hundred thousand in 6 months. Would that be enough to start noticing performance differences?

Personally, I would prefer the normalized version for flexibility, but I'm afraid that it won't happen unless I can find some reputable evidence that either A) there is no performance penalty, or B) the penalty is worth it.

So, if anyone has any other links that could help, I'd love to see them. If not, thanks for the help so far!

Dave