Ether oracle: Robert M. Metcalfe
By Ross Kerber, Globe Staff
These days, even cyberdunces seem to be throwing around the term Metcalfe's law, the theory that describes how the value of networks grows quadratically as they connect more people. Robert M. Metcalfe, at the center of Boston's technological solar system, invented the computer technology known as Ethernet in 1973. He founded 3Com Corp. and led the company as it popularized devices allowing PCs to communicate with one another quickly. Today, he is a columnnist for InfoWorld, as well as vice president for technology at InfoWorld's publisher, International Data Group of Framingham. Metcalfe moves between a Back Bay brownstone and a farm in Maine devoted to preservation of rare animal breeds. Metcalfe spoke with Boston Globe technology reporter Ross Kerber last month.
Is networking likely to remain a major force of change over the next 1,000 years?
A large part of the history of mankind is of transportation getting faster. Last century, communications started getting faster, too. As that happened, it started to displace transportation. The biggest trend I can see is the continuing substitution of comunication for transportation. So, as the Internet, or something like it, continues to evolve, we'll have more and more communication and less and less transportation. The upside of that is obvious: We'll have less reliance on fossil fuels and less need for public works. And people won't have to be traveling away from their families so much, since they'll be able to communicate more.
Sounds like the basis for Metcalfe's law.
I didn't even name it. But the concept is real. When I was selling Ethernet as a concept, I had to explain the fact that people were trying networking, connecting their PCs together, and the experiments were failing. Potential customers were failing to see the value. I was trying to sell them e-mail and printer sharing via my company, 3Com. What I'd often do is persuade people to hook up three PCs, but I wouldn't be able to close the sale. Customers would say, ''What do I need e-mail for? I don't send that many e-mails.'' I'd say, these achievements haven't achieved critical mass. Some percentage of their communication would have to go on line to get there; just three out of 1,000 people is below the critical stage.
As a society, have our communications reached critical mass yet?
With the Internet and the telephone network, we have. That's because each additional person who has a phone gets to talk to all the other people who are already on it. These are ''network effects,'' and they go way beyond Metcalfe's law. It's really how you get these things to propogate. The reason that Yahoo and AOL are so popular is that they understand how to propogate - how to create epidemics on the network.
Why isn't this just a phase, a short-term phenomenon?
There's no way you can call it a fad. Tulips and hula hoops were popular at one point, but they weren't useful. The network, by definition, is useful.
So we're on the cusp of something big?
Yes. Looking at the microprocessor, the fractional horsepower motor is a strong analogy. There are today about 50 small motors in every home. Now we're at the point where we have dozens of microprocessors in homes. The irony there is that the smaller microprocessors are far more powerful than the ones they replaced, unlike the smaller motors.
And the analogy is that the smaller motors had the effect of spreading out the power network?
Yes, and the microprocessors are bringing in the Internet to the home. With more and more networking everywhere, we can be telepresent more frequently, which is the real benefit. Most people think that because they've used a phone and maybe videoconferenced that some day it'll be almost as good as being there. I prefer to think that teleconferencing is better, because, well how often do you go to a meeting and wish you had all your stuff with you? Your notes, your papers?
So how long will it take until the network reaches that point?
You could say we've already gotten there, in some cases. Already, I don't want to fly to England to talk to people that often if I don't need to.
What about class issues and the inequalities that arise as the benefits go to the rich first?
That's a perennial problem. It's not just the Internet. The introduction of hygiene was such a problem. Even having clean water was a real have and have-not thing. So let's fix that perception first before we start thinking about how the Internet is different. I mean, fewer than half the people in the world have ever made a phone call. So I don't know what we're supposed to do about the Internet - hold it up while everybody catches up?
Or one could regard the Internet as a tool for general development, which ought to benefit everyone, in theory?
Exactly. But most people who ask the question about inequality invariably have some stupid thing they want to do to answer it. The best way to let the Internet proliferate is to let it.
Sounds like you regard the Internet as the ultimate manifestation of capitalism.
It's just the next stage, where frictions in value exchange [transaction costs] are lower. There's the old theory that totalitarianism and networking can't last, so the best thing we can do for the people in China is to install telephones there. You can sum it up by saying more communication is better, and I subscribe to that theory.
Sounds great. So let's say it's the year 2999. What's networking going to have done for me lately?
I'm not sure whether we should call it ''networking'' or ''the Internet'' or what. But because of it, the power is shifting from governments to markets, and the markets are getting better because frictions are going down. Centralized planning is waning in importance, and the nation-states are going to atrophy. They're usually just an excuse for picking a fight, anyway.
You mean we'll just have one big network?
There will be many networks. There will still be some small tyrants, too, and roles for big governments. But those roles will be things like preserving markets, where people can voluntarily exchange value. Also defense, law enforcement, and antitrust.
You're leaving out health care and education, I notice.
Well, government isn't performing so well in those areas. The ultimate point is to look out for the have-nots. So how better to take care of them than to let them get the benefits of the market?
In other words, put them on the network?
Exactly.
| 
