Metcalfe's Law

An interesting recent article in IEEE Spectrum has a nice discussion of how Metcalfe's law -- the maxim that the value of a network of size N grows quadratically (that is, proportionally to N²) -- is wrong, and why the actual growth rate is probably closer to N ln N. The intuition is that Metcalfe assigned equal value-added to all possible connections in the network, when in reality the marginal value of an additional connection between two nodes is decreasing in the number of possible connections. The paper suggests some interesting questions about the valuation of networks of various kinds, including those of interest to ELSers.