Until 1999, these two groups shared a common goal: Keep government out. They both rallied around the idea that the Internet should grow without government getting in the way. Whatever the context, they worked together to resist regulations, particularly those on pornography, encryption and Internet commerce. For each of these issues, an alliance was formed to resist government's efforts to intervene.

Early Internet surfers wanted government out because they liked the freedom of cyberspace. The code of that world made certain liberties possible: to speak and access speech freely, to browse with relative anonymity, to copy and distribute content individuals liked, and to build at will without regulation. The Net made this freedom possible, and the enemy of this freedom - this community thought - was government.

E-commerce advocates wanted government out for a very different reason. They longed for the freedom to innovate without interference, and more importantly, without taxes. If government could be kept out - if the rhetoric of early adopters could be used to keep government out - then commerce would flourish. Commerce would thus stand with Net freedom-rights advocates to keep government away.

This happy union began to fall apart early in 1999, and no event signaled the breakdown more clearly than the debut of Intel (INTC)'s Pentium III microprocessor, which could link a particular Web transaction back to a specific machine. The identification became possible through a change in the chip's code. No longer would a computer stand mute as the Net tried to identify it; now each machine had a name (like any device on an Ethernet network or, for that matter, like the mother of the Macintosh, Lisa), and it would reveal its name if properly asked.

Online activists went wild. What right did Intel have to make a computer part of a system of Net surveillance? they asked. Privacy groups sent letters and press conferences exploded. Fairly quickly, Intel backed down, at least for the moment.

But the event would mark a pattern. The original Net was magic - no one could deny that. But there was much about its original design that e-commerce groups found less than ideal. It was too hard to collect data about surfers; too hard to know how individuals really used the Net. In short, it was too easy for people to use it however they liked. These were "bugs" in the Net, and business sought to fix them. It worked to introduce, through code or contract, changes that would make the Net safe for commerce.

The Pentium III was just one of the first examples; the year was filled with others. In February, it was Free-PC announcing computers gratis to anyone who would turn over his or her personal data. April saw a wonderful battle at the annual Computers, Freedom and Privacy conference between the recording industry and MP3 activists, with the former seeking any way to kill the latter by building copyright controls into the technology. The month of May saw FCC Chairman William Kennard pressing censorware products: software that filters content on the Net according to lists compiled by companies that call themselves names like "Net Nanny." August saw the announcement of Amazon.com (AMZN)'s "purchase circles," which made it possible to see what employees at other companies were buying. In October, Xerox (XRX) showed off its technologies for monitoring workers by firing 20 employees who allegedly had been viewing porn sites at work; the New York Times did the same the following month to employees who wrote "improper" e-mails. But RealNetworks (RNWK) won the prize in November with the announcement that it had built code into its RealJukebox software to report data from users' machines back to the mother ship.

In all these cases, the story is the same: Commerce seeks to build an Internet that is different from how the Internet originally was. Through code or through contract, it seeks to change the design of cyberspace to make it more commerce-friendly.

Now, there's nothing necessarily wrong with removing bugs. But for those who considered "bugs" like relative anonymity to be features, the year taught us two important lessons: The Net's citizens could not count on the Internet remaining as it was, nor could they count on an "ally" in commerce. The Internet was changing, and those who would resist the changes would have to find ways to resist the power of the business community.

Obviously, it was not just commerce that sought to change the Net. The FCC's Kennard represents many in government who hawk censorware to filter the Net in ways the Supreme Court has ruled the law cannot. Even more alarming was the effort by the FBI in November to convince the Internet Engineering Task Force to modify the Net's protocols to make it easier to wiretap. The IETF said no, resoundingly so. But one imagines federal investigators are not terribly upset with this loss. Now they have a clear example to take back to Congress, to urge legislators to regulate even more directly.

The point is that both government and commerce are in a position to change what the Internet is. Both can modify the Net to make it fit better with each vision: cheap and efficient enforcement in one case, cheap and efficient commerce in the other. Early adopters of the Internet began to understand this over the past year and started struggling with what might be done in response.

The intuitions of the early Net users here are not very clear. These folks tend to think of themselves as "libertarians," and thus ignore threats to liberty from anyone but government. So when conservative columnist William Safire wrote of the dangers to privacy that the new e-commerce architectures would present, technolibertarian Declan McCullagh scolded Safire for his anticommerce views. Apparently privacy is important only if it is privacy from government.

Many people will take from this conflict a basic point: If there were values of the original Internet that are values worth defending, then they are worth defending against whatever the threat. If government would compromise these values, then government should be resisted. If commerce would compromise them, then commerce too should be resisted. This is not because the original Net is good by necessity, but because where it was good, the good should be defended.