...

...

Is [marshaling functions] something absolutely impossible in Haskell and by what reason? Just because of strong typing (forgive my stupidity ;)? Or are there some deeper theoretical limitations?

If you're interested in some recent work here, have a look at Clean (similar enough to Haskell), specifically their work on first-class I/O. I gave a reference in another thread: http://www.haskell.org/pipermail/haskell-cafe/2003-October/005243.html Then follow the references. "orthogonal persistence" is another nice search key..

...

The big theoretical issue is whether it would provide an Eq or Show instance for -> by the backdoor. Careful API design could avoid the worst of this.

once you can do IO, you can inspect the GHC heap for "equality", or map memory into Haskell data structures, so that's not really new ..

What's the problem with a Show instance for ->? (Or, put another way: what makes the current Show instance in the Prelude harmless compared to a fuller implementation?)

have you tried to show a function?-) an equality based on that would be very rough, a proper equality would be undecidable, and there are whole worlds of other "equalities" in between. Cheers, Claus