> Wouldn't be possible to make Dynamic derivable or something (better: > with a magic default definition of typeOf, so the instance can be > made after the type definition)? Yes, I'd like to do this in GHC at some point. Simon