The Web may be huge, but it won't turn you into a billionaire

We at gulker.com have been reading lately about something called the "small worlds" concept and how it applies to the Web. The reading in question is coming from a new, short tome from MIT Press The Laws of The Web by Bernardo Huberman. He is a Hewlett Packard Fellow at HP's Palo Alto research centre. Around these parts, "Fellow" means somebody who really smart people regard as really smart.

Huberman has applied a tool called statistical mechanics, which physicists use to understand very complicated systems, to his study of the World Wide Web.

Interestingly, Huberman has discovered that despite the seeming chaos of the freewheeling Net, the Web actually has an underlying structure based on mathematical laws that are well known in natural systems. "Well known" if you're a physicist, that is.

Physicists often deal with systems that display what's called non-linear behaviour. These tend to be complicated collections of atomic particles, where behaviour depends on thousands or millions or quadrillions of arcane interactions. Trying to figure out overall behaviour by calculating each and every interaction would drive the biggest supercomputers nuts for trillions of years, so physicists rely on tools that look for patterns in big systems, and that can offer a glimpse of how, on average, the system will behave.

One of the things Huberman noticed was that the distribution of Web pages, and links in Web pages, follows something called the power law, which describes the general behaviour of a lot of things, especially in terms of the distribution of the very big and very small.

Take the Web: it has a few very large sites, and a huge number of very small ones. Or wealth: the world has very few Bill-Gates-sized billionaires and lots of very poor people. Basically, the power law tells you that even if you're born in a land of the free, you're unlikely to become rich.

Huberman found other phenomena common in power-law distributed systems, and determined that the Web exhibits a behaviour called "small worlds". One popular example of small worlds is the "six degrees of separation" theory, which holds that any two strangers are separated by a chain of about six mutual acquaintances. The "six degrees" concept was originally discovered in the Sixties by the justifiably famous Harvard sociologist Stanley Milgram. He asked a group of randomly chosen people in a Midwestern town in the US to mail a postcard to a Boston stockbroker, using the rule that they must send it to someone they knew on a first-name basis and who they thought might be closer to the Boston guy either professionally or geographically.

That person must send it on to another "closer" first-name acquaintance, and so on.

The average path was six; the conclusion, that most people in a large population are connected by only a short chain of acquaintances. When Huberman applied the principle to the Web, he discovered that any two Web sites in the world are separated, on average, by a chain of between four and 19 (at most) hyperlinks.

Other "laws of nature" may also be operating on the Web if you want my opinion (and, trust me, no one around here refers to me as "Fellow"). One of those may be an observation drawn from studies of Darwinian evolution, which holds that given enough time and a stable environment, one species will come to dominate a given niche.

Even a small advantage, operating generation after generation, will eventually crowd out even the slightly less qualified. Which, translated from Darwinian by way of physicist, means that putting up a Web site is even less likely to make you rich than being born is.

cg@gulker.com