Ben Kraal

By: Ben Kraal

Nov 18 2004

Category: Uncategorized

1 Comment

Ricky takes issue with the assertion of Gene (that I stole from Matt), that ubicomp is a 100 year problem. Excellent.

Ricky says:

There are many, many problems to overcome before the true dream of ubiquitous computing arrives. But in a limited sense, it’s already here.

And I would agree with him. I’ve made the same point in lectures I’ve given. I have a small computer on my bicycle. I pay for things using EFTPOS in preference to cash. In Sydney and Melbourne cars have electronic tags fixed to their windscreens to collect the toll on privately funded roads. Ubicomp is here.

Ubicomp is also, emphatically, not here. Ubicomp, as many people describe it, is a sort of idyllic place where:

These tags will be able to automatically communicate with each other via information networks, and accomplish the everyday tasks that we normally do for ourselves. We will live in a computer-controlled environment, in which computers will function without a human being having to manually perform each and every little operation. With the advent of this “ultimate” in computing, our lives will be made dramatically more comfortable, convenient and efficient. (source)

And, in this ubicomp world, nothing ever goes wrong or is used for malicious intent.

Ricky also says: “UbiComp is not about handheld computers anyway; it’s about hiding computers from sight altogether.” By which I’m guessing he means computers in the walls, in the furniture, in the crockery.

I don’t doubt that some aspects of the technology that will make ubicomp work exist already (see crockery above). I have doubts that ubicomp proponents have considered the downsides of the actual integration of computing into every aspect of our lives. Adam Greenfield considered smart furniture and said:

Consider a piece of furniture whose physical shape adapts to changes in its environment: say, a bed capable of adjusting its contours in response to its occupants’ weight, the ambient temperature, or whatever other factors you can imagine (and I’m sure you can imagine quite a few). Then suppose that there’s a power surge, circuitry fries, and the onboard software crashes while the bed is morphed into some outré configuration.

Now you’re locked in, right? Now you’re living a little vignette of Phildickian slapstick: forced to try falling asleep on a surface jacked up twenty degrees to accommodate reading, at least until the repair crew can get out to your place. Even assuming that good designers – and they’re not all good, are they? – would build in some kind of physical failsafe designed to return the furniture to a “neutral” configuration in the event of a shutdown, who knows whether that configuration is the most congenial to the way you generally interact with the object in question?

But Ubicomp is supposed to be even more than smart furniture. It’s supposed to be everywhere and in everything. I have doubts about the value of “self-documenting, self-tracking, self-extending” objects. Do I want to live in the distopian future of, to take a bad example, Minority Report, where I am addressed by name by advertisements for bland Japanese luxury cars (Ok, that’s probably not the actual example from Minority Report)? Does my backpack need to acquire new functions automagically? Does every can of baked beans need to contain a record of it’s entire production process?

The point of the 100 year problem is that although the technology exists, it’s fragile, and no one really understands what it’s like to use in the real world. The Prada store example in the Fred’s House article spends some time on handhelds but also addresses RFID closets that don’t recognise their tags, automated doors that are broken and more. It’s just one example, but it’s a good one. The use of technology in the real world is greatly divorced from the same technology in the laboratory.

Ubicomp is here, now. It works. Ubicomp is also 100 years away and right now know one knows how to make it work. It’s time now for a redefinition. Ubicomp is a bunch of technologies like high speed networking, RFID (or any sort of IDing), RDF etc. They work. Ubicomp Applications are things like smart rooms, spimes, e-tag, smart cards, and so on. Ubicomp applications are something that we, as the IT/High Tech academic community have little understanding of at the moment. We don’t know what works, what doesn’t and why or why not. We especially don’t understand how Ubicomp applications work (or don’t) in the real world, and I suspect we won’t for a very, very long time. The increasing incidence of ubicomp and ubicomp-like applications is interesting but Ubicomp, like Speech Recognition and Agents is still something that I consider to be, at best, a Long Way Off.