thanks oleg and heinrich for the solutions; i'll definitely take a look at the vault package. i'll probably end up using unsafeCoerce though, it's too tempting (shame on me.)
On May 2, 2012, at 2:33 AM, o...@okmij.org wrote: > It seems you actually prefer this solution, if it worked. This > solution will entail some run-time check one way or another, because > we `erase' types when we store them in the log and we have to recover > them later. actually i don't prefer this solution. i'm interested to hear if there are other solutions (or ways to avoid the problem entirely.) i don't know what to search for, something like interpreters with environments with heterogeneous types. i find the circuit-diagram / functional representation the most interesting, but it seems unfortunately syntactically impossible. best, ben _______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe