How is this more functional than IO in Haskell? Your approach seems to be the same as I/O handling in Haskell minus the convenience of monads and plus caching.

How is this more functional than IO in Haskell? Your approach seems to be the same as I/O handling in Haskell minus the convenience of monads and plus caching. Not all that different.