On Thu, 19 Aug 2004, Chris Poole wrote: > > --On 19 August 2004 10:10 +0100 Chris Poole <poole at maths.ox.ac.uk> wrote: > > > > > Why not map the nodes to edges (and vice versa). > > > > How? You can't make this work in a one-to-one way - try it. ... don't you get round it by using hyperedges?