Also at said dinner, I found myself in the position of having to describe to humanities people what a math paper is like, and I failed miserably. Why would you solve a problem that's already been solved and for which good references already exist?