On Thu, Nov 11, 2010 at 10:53 AM, Ketil Malde
..and you are able to tell the difference. Am I wrong in thinking that this could be made to work if serialization was to/from an opaque type instead of (Byte)String, so that the *only* operations would be serialization and deserialization (and possibly storing to/from file)?
This was my first thought as well! However, reading to/from a file would of course be in IO, at which point you'd be free to read the file back in through normal means to get at the representation. So in that respect, this is equivalent to (a -> b) -> IO String. Outside of IO, it would pretty much have to be limited to serializing and deserializing. You'd be able to create opaque tokens representing functions, pass them around, and/or extract the function in order to apply it. Conveniently, it turns out that Haskell already has support for this, you can implement it as follows:
module Serialize.Pure (OpaqueFunction, serialize, deserialize) where
newtype OpaqueFunction a b = Opaque { deserialize :: a -> b }
serialize = Opaque
Toss in some existential types as desired, if you want to hide the function's actual type. I suppose one could object that this isn't actually serializing anything at all; to which I would respond that, in pure code, how do you expect to tell the difference? - C.