On Notation - Pat Harmann at RacketCon

The talk covers a lot and the questions are just fantastic. Pat Hanrahan won 3 Academy Awards, started CUDA, cofounded Tableau and Pixar, won a Turing award etc. etc. While he does all his math in Haskell now, he talks about K, APL and Agda while at RacketCon.

The SETL language compiler sounds very interesting, choosing underlying data structures (timestamp) for the most efficient execution of the semantics. He similarly noted that when looking at a notation, we don't just think of the AST etc. and the accidental complexity or occasional drive for mechanical sympathy just get in the way. Why don't we have more such implementations?

Zhang and Norman's Work on the Representation of Numbers also seems fascinating. (timestamp)