IEEE Spectrum has an article titled "Metcalfe’s Law is Wrong" that was published in the July 2006 issue arguing that the basic math behind "Metcalfe’s law" is wrong.
Of course it is, how could it not if you think about it? The infuriating thing is that this still seems to be news!
I blogged about that very thing more than a year ago, when this made news last time around, saying:
Of course it’s overstated! If I have a fax machine and somebody in central China, who I have never heard of and will never interact with, buys another fax machine, the value of the network will not grow proportionally to N (say, hundreds of millions: the number of fax machines in existence today) but by some much smaller number such as a couple of hundred at the maximum (the number of fax machines that person will ever send a fax to or receive a fax from)…
I recall giving an internal presentation mentioning that very issue in 2000 at my previous company…