I hadn't seen that note from him, thanks for the link. I think J solves a lot of the problems he makes of APL. It can be taught without a J interpreter, it doesn't require a special keyboard, and it offers abstraction capabilities. Also it's an answer to some positive things he talks about in this 1985 interview and elsewhere: http://www.cs.utexas.edu/users/EWD/misc/vanVlissingenIntervi... It supports both functional and imperative styles and it's deeply rooted with mathematics. It's also "hard" and isn't made for the layman developer. It encourages a very logical thought-process. I would guess he'd think Haskell was superior, but I still think he'd at least consider J in higher regards to APL even if he didn't like it.