Chapter 25: The URL on Your Cereal Box

In Chapter 17, I limned the creation and development of the Web. In a subsequent chapter, I'll talk about the geographical spread of Linux. But first, I want to look at the spread of the Internet and the Web that depends on it.

The ARPAnet became functional in 1969: at the end of that year, there were four nodes. In January 1976, there were 63 (so much for 5- or 6-bit addressing). Five years later, in August 1981, Host Table #152 listed 213 hosts. In May 1982, Host Table #166 listed 235.

The great switch to the domain system occurred on January 1, 1983. It was none too soon. The 562 hosts of August 1983 just wouldn't have been feasible under the older protocols and the older scheme.

Here's the growth over the next few years:

I'll stop there for a while, for several reasons, not least because this marks the advent of both the Web and of Linux. But this marks several other things as well.

On the political front, the Department of Defense relaxed its notion that only government, academic and research sites could connect. This was partially the result of recognizing the expansion of private networks (like IBM's VNet and Prodigy) and of networks distributing news and mail (UUNET, Bitnet, Fidonet), as well as the recognition that the network of networks was already vastly larger than the US and its "allies."

On the "engineering" side, the advent of the desktop machine and the commercial modem, meant that individuals could have a computer at home and plug in to their telephone lines. All for under $3000!

In 1971, BBN estimated 10 users per host. It was extrapolating from 617000 to one million and multiplying that gave the Internet Society the 10,000,000 users that TIME magazine claimed in 1992. In an informal poll a decade ago, I found that I knew individuals who owned a dozen domains and who were sole user on most (0.1 user/domain?) and that IBM's T.J. Watson Research Center had over 4000 users and one domain. I didn't know then and don't know now what the "correct" ratio is. Craig Partridge estimated 5/domain in 1994. John Quarterman estimated 3.5. These sound more reasonable than double or triple those ratios.

The advent of NAT (Network Address Translation) makes all the rations and estimates yet more unreliable: we have no way of determining in any accurate fashion how many desktops on a LAN are lurking behind a single address.

But in 2003 (the last dates for which the numbers seem feasible), there were about 175 million domains. Using the growth rates of 1995-2002 for 2002-2006 would mean 600-700 million domains at the end of 2006, and about 2,000,000,000 users worldwide. (This counts students in schools and people in libraries, of course.) Just under a third of the world's population has access to an Internet-connected device.

There were about two dozen in January 1970.

In 1990, when there were about 300,000 hosts, three students at McGill University in Montreal set about writing a search engine that would poll FTP archives on a regular basis. Alan Emtage, Bill Heelan and Peter Deutsch called it archie. Soon there were other search engines, too: jughead, veronica, WAIS [Wide Area Information Server].

They were very useful, for a brief period of time. The University of Minnesota, where gopher was developed, wanted to profit from it. Tim Berners-Lee offered the Web free. Though he's now Sir Timothy, he's not a millionaire. But, by 1994, the World Wide Web had swept away all the other browsers.

Why is this important? Well, in 1973, when Dennis M. Ritchie and Ken Thompson gave the first UNIX paper, there were about 200 people in the room. There were just over 40 hosts on the Net. Word of mouth and then the CACM paper were how the word got out. When Linus Torvalds posted his note on the comp.os.minix newsgroup, there were about 200 groups. There were over 600,000 hosts connected to the Internet. The potential audience for Linux was enormous. And it was virtually instantaneous.

I wonder what my long-ago Toronto colleague, Marshall McLuhan, would say; today's Internet and Web are "hot" media far beyond his notions.

As I wrote in Chapter 17, try buying a candy bar that doesn't have a URL on it.