Random thoughts in the shower

[handslive]

For some reason I woke up this morning thinking about a combination of "nanotech" ideas I've read about, specifically thinking of Diamond Age and Deepness in the Sky. My memories of these books may be muddled since I think it's been three years and two years respectively since I read them.

Both books talk about communications between air borne micron or sub-micron sized "machines". In Deepness one gets the impression that a large number of passive sensing systems are available and some method of wireless communication networks them together, provides energy, and maybe even a degree of active sensing/feedback. Similarly, Diamond talks about a whole class of information warfare based on having "machines" that seek out other "machines" in the air, producing a rather dangerous dust-like pollution during conflicts.

I was thinking about this specifically in terms of several supposed traits in SF of micromachines. I'll avoid thinking about it as nanotech for reasons I'll give below.

• First off, there's usually an assumption of some networking capability to allow for complex behaviour. Complex behaviour requires coordination, so there has to be a communications net.
• Secondly, some degree of environmental interaction is available. This might be a control task, active sensing (as opposed to passive sensing), or providing what I'll call macro-level communication (that is, communication outside of the local micromachine environment).
• Finally, some repair or growth capability. Usually the ability to produce new micromachines either of the same type or for specialized tasks.

This is no surprise. Our fictional thinking about futuristic uses of micromachines reflects our thinking about organic systems to some degree. Complex, specialized, and interactive.

So, I'm thinking to myself about the human-micromachine interface this morning. I believe in Deepness that Vinge suggests the need for the main character to maintain a "trance-like" state that allows him some degree of interaction with the micromachines. Basically, he sends a specialized signal into the network that can be picked up by the passive sensors. This results in a behaviour change so that certain micromachines will attach to his body in ways that let him interact. The degree of interaction is severely limited because

• (a) the machines don't have a lot of power to blow off,
• (b) Vinge understands that there's a limit to the degree of programming complexity available.

But I think it imposes too high a degree of skill on the main character to continuously interact with the network in this way. No way would it work, unless you have modified humans. No way would you be able to train anyone to do it. There's a high degree of risk, if the complexity of the network is truly limited as we've already said, that the system will randomly recognize other untrained humans as trying to interact directly through this interface and then screwing things up.

So, what would a glitzy air-borne interface look like? Let's forget the Star Trek "coherent light" idea (and you have to wonder why they don't build interfaces that use this tech on their bridge for instance -- totally mobile displays, fewer exploding circuits, completely comfortable and personally configurable control "couches" for staff, support for non-humanoid aliens).

The basic idea is a display that "coheres" out of a fog of micromachines. The display could react to a variety of inputs from the user, from light pens to gestural controls. There are a number of problems just in terms of physics with this idea.

• Networking. Even in Vinge's universe, there's no good explanation for how the network functions. The machines are small enough to be mistaken for dust, but a true nanomachine doesn't have enough space to form any kind of transmission. You can't "interact" if communication is only one way. Hence, micromachines. Supposedly, someone is already working on designs for micromachines that look like very large dust particles. They still have problems with air buoyancy issues (you need a pretty stiff wind to keep them off the ground), but they do incorporate a very small, simple, but short range network for passive sensing. Further, it might be possible to play with assumptions about the environment the micromachines live in to propose purely molecular ways of passing information between machines. This suggests an organic model, though, which probably doesn't work that well in air.
• Power. Assuming we can work out the communication problem, where do we get the power to move information around? Even if we think we can pass along information by dropping molecules from one place to another or changing molecular state, this costs energy. Vinge suggests microwaves as a power source, but I wonder if you could do this without cooking the humans in the vicinity. Further, if we're going to display information (the easiest form of communication with humans), we're talking about emitting light which takes a lot of energy. If the light comes from a secondary source and we're just discoloring/obscuring it for display purposes, then you have to ask why we don't just use flat-panel displays.
• Materials. We're going to need to replace these little buggers frequently. They'll get chewed up by ventilation systems, organisms in the environment, other micromachines. Vinge and Stephenson both assume we can build more machines on the fly in some sort of self-replication mode using either a time-elapsed method or some notion of "network density". Leaving aside the issue of manufacturing error in this process (viral mutation of the machines anyone?), we need a source of molecular material if we're talking true nanomachines or of very specific pieces of raw material if we're talking larger micromachines. There's an energy question here, too, but the underlying assumption seems to be that raw material is readily available because we don't think about the content of dust that much. Unfortunately, dust with the right kind of raw materials here would probably be toxic to humans. This leaves us with "vats", I guess. Some sort of predetermined construction area that micromachines leave from to join the network.
• Location. In order to provide a coherent display, we need to keep particles within a particular space. Even if we assume that we'll light up only particles passing through a particular volume of air so that no one particle has to stay in place, we still need to get the micromachines into the volume of space directly in front of the human. This starts to sound like a rather ugly dust cloud, though. Standing columns of air? Chimneys?

On the whole, I decided it wasn't feasible, but it calls a lot of other nanotech fiction I've read into question. There's a lot more handwaving in the SF that addresses this field than I had guessed. Anybody got other nanotech fiction they'd like to recommend?