I really enjoyed reading this paper last night - it gives a new way of looking at how graphs can be represented in programs. One of those neat changes of view that's so obvious in hindsight.
It's also thought-provoking as an example of how a non-trivial implementation can be hidden behind a neat interface.
The resulting library is here (ML + Haskell).
Posted to functional by andrew cooke on 5/10/01; 12:27:57 AM