Paper still has its uses (monitors make bad airplanes). Terminals are still appropriate for some tasks. WIMPs are best for the personal editing of document-like material. But for socialising or other forms of shared experience, we need something better.

Current 3D interfaces are 2D windows onto a simulated world. Controlling an avatar is not totally intuitive. One day we will have true VR, so the point of consciousness can coincide with the point of action; but for the time being we are stuck with the separation: avatars we must be.

Current demos of the technology are too twee for the business opportunities to be taken seriously; which is a pity, as 3D promises to do for social interaction what 2D technology did for document production: make the process vastly more efficient and (over?) productive. So what might a "sanitised for business" demo look like?

A distinct lack of purple dinosaurs and white rabbits, for a start. Each avatar will be soberly dressed, genderised (ideally personalised with regard to clothing), but recognisably human.
The face should be a real image of the person represented. The 'spaces' in which we interact should be office-like, with chairs and tables and whiteboards etc. Each avatar has a 'home' room, into which they can retreat for privacy.

Which brings me to the First Law of the Network Reformation: identity is the foundation of interaction. Any person or service that cannot be uniquely and accurately authenticated is not able to participate. But the Second Law is of near equal importance: That which I produce, I control; IOW the data streams that my authenticated identity produces are shared and stored strictly as I wish, or have actively agreed to.

The first law ensures trust, and the second privacy. Spamming, as an example, may never be impossible, but it should become traceable. As another example, when I vote in a national election, whom I voted for should remain secret, but that I was entitled to vote must be accurately ascertained. (These are issues I'll return to.)

Back to our demo. Imagine a virtual table around which are gathered several avatars with colleagues faces. The task before us, and how it might be performed, will be the subject of my next waffle.

[footnote] Although the boss sees the sober-suited avatars, as does anyone watching on a 'public' monitor, a neat hack would be to enable the participants to customise their view of the others' avatars. One participant might decorate the boss with horns and a pitchfork, another with a glowing halo, and a third change them into a donkey (albeit one with a human face). Only the customiser sees their preferences.

Whether the above is achieved or not is moot; but one critical ability will be colour control. Colour blindness is surprisingly common. Other accessibility issues, such as how the deaf will hear and the blind see some 3D equivalent are important issues. If the personalisation of the experience is designed-in from the start, much can be achieved.