Just as the real world knows several different types of crime, so crime in cyberspace too has variety. Both worlds punish depending on the severity and social perceptions of an act thought to be wrong - although cybercriminals are often impossible to trace, leave alone imprison. And as the former does not treat murderers and pickpockets alike, the latter should distinguish a cracker from a pirate.

Should - because there is no established order of crime and punishment in cyberspace, and one can only theorize. One can do so on the basis of formal law, hoping that it will colonize the infosphere; or one can accept the inaptness of real-world legal systems, and build a model of the order the infosphere may evolve. Such a model would revolve around three classes of crime - of knowledge, of information and of data.

Data crime is the most obvious. Data criminals include all the famously arrested perpetrators of famously revealed plots to destroy the Internet with a computer virus, break into the Pentagon's computers and steal (or modify) people's credit card records. Data crime is so overt because it almost always lies on the edge between cyberspace and the world outside, where law enforcement sometimes works. Its real-world cousins range from trespassing to burglary, and its best antidote is, unsurprisingly, the on-line equivalent of a locked electric fence - better security.

Information crime is more serious, to discover and to punish. It relates to what is not just explicit, but implied, and includes violation of privacy, defamation, and forgery. It is a hazy world of reputations, context, deceit and hostile intentions. Its perception as crime is highly dependent on social codes and no list of specific acts is likely to be widely acceptable as criminal. Even if discovered, the damage done by such crime is usually irremediable. On the other hand, information crime is especially vulnerable to preventive technologies, most based on some form of encryption.

This is in stark contrast to probably the most economically important category, knowledge crime. This is not about the movement of raw data, but what people do with them - for knowledge is expertise and also its output, intellectual property. This controversial term carries with it controversial rights, and its real-world equivalent may include - depending on your point of view - breathing the clean air over another's garden, stealing someone's car, and slicing fingers off a violinist. Disagreement among societies in cyberspace on precisely what is criminal, and what is practically enforceable, will hamper the protection of copyright, patents and trade secrets for some time to come.

Technology, yet again, may be the answer. But more than anything else, the treatment of crime in cyberspace is in essence social. It is in essence heterogeneous, thanks to the diversity of the infosphere - even though some classes of crime may well receive common punishment everywhere. And it is in essence informal, based on 'goodwill law' - quite different from the formal legal systems that surround us today.