Re: [Haskell-cafe] Proving stuff about IORefs