Off the wires, we are used to two basic modes of inter- personal communication - face-to-face, and writing, in order of decreasing directness. One provides less sensory input than the other, but allows more time for measured, intelligent and therefore useful responses. Electronic discussion in the form of e-mail or even real-time textual chat is unusual in that it provides both a sense of immediacy as well as the space required to organize one's thoughts. However, telecom giants are hoping that consumers of the future will simply cart their real-world conversation paradigms into cyberspace, making videoconferencing the rage. This is unlikely in the long run - even novice users of the Net discover the power of pure text early on.

Physical proximity during a conversation encourages awareness of subliminal messages, through body language, inflexions of the voice and the physical context of immediate surroundings. It also distracts one's attention with a flood of details, sometimes useful but often irrelevant to the purpose of communication. Face-to-face interaction is very good at putting people at ease - they feel reassured that they are talking to someone who is really listening, and they get to know what other participants are "all about."

E-mail, when best used, gets to the point. It does not carry any subliminal stream of unintended communication that is beyond one's control; instead, it encourages focussed, well thought out and relevant content. Of course those who have nothing to say also use e-mail, and the lack of content in their messages is more apparent, more quickly, than in face-to-face communication.

Supporters of videoconferencing believe in virtual reality. As customers didn't consider talking to disembodied faces very appealing, we've progressed to torsos, sofas, embalmed-Pentium paperweights and even the odd potted plant. The idea is to create a sense of physical proximity, carrying people into each other's virtual living rooms. Unlike real living-room meetings, the barrier of distance and the camera lens (headset, datagloves, whatever) haunts any VR conference, so any closeness is clearly artificial. So you end up with the disadvantages of too much immediacy, of forced "spur-of- the-moment" responses, without most of the advantages of facing real people.

Face-to-face communication and attempts to replicate it over fibre make the basic assumption of definite identity: all participants in a conversation are real human beings using their true names, rather than intelligent agents, pseudonyms or dogs. This is almost certain not to be the case, as concerns over privacy mount - a hypothetical Infosphere Secret Police could, from traffic in public forums, compile dossiers to shame the Stasi. One could imagine pseudonymized videoconferencing, with ski-masks digitally painted over faces, but unless you enjoy being subject to constant voice-stress (or tilt-of-head) analysis, you'll stick to text.

Of course, realistic videoconferencing will have a niche market. Even though Granny, in the famous example, can't hold her newborn grandson over a video-phone any more than over a VCR playing a home video, she can see him gurgle in real time. High-flying directors of multinational corporations might like to hold board meetings while cruising off Nauru - but the truly foresighted are already realizing that the suits, accents and gestures transmitted so well over VR phone do not imply competence or expertise. Coherent conceptualization, which is all that textual e-mail permits, does. The sooner we understand this, the better equipped we will be to enter the information age.