v-2.org

"Ubiquitous" computing, "pervasive" computing: phrases that are beginning to appear beyond the narrow technical circles of their origin, as an arcane idea surges toward daily embodiment, its hour come 'round at last. But beyond the clash of branded buzzwords - those originating at Xerox and IBM, respectively - what's the hype all about and why should you care?

As first articulated by the late Mark Weiser over a decade ago, the field still cohering under the rubric of ubicomp concerns computing without "computers": what happens when the access to functionality and processing power we associate with these slabs on our laps and desks evaporates, becomes part of the everyday environment?

It’s also, by implication, about the fusion of orthodox computer science with wireless networking technologies, "artificial intelligence," possibly robotics, and maybe even nanotechnology. And inevitably, as a consequence, it’s about all the transformations of our notions of space and time, self and other, that result when digitally-encoded information is everywhere around us like the air, available for the asking.

Why am I interested in it personally, to the extent that I invested in sending myself halfway around the world to a conference devoted to the field? Well, I think it would be virtually impossible for a Web site concerned with "interface + usability, architecture + urbanism, media + culture" to not be interested in ubicomp, since it touches upon every single one of those themes.

The relevance to the evolving discipline of interface design is obvious, and I think it’s not too difficult to see that developments in the field have the potential to radically and permanently alter the experience of citying, or of belonging to a polity. (Ubicomp has a curious resonance with practices of urbanity that I want to explore further.) Beyond that, I don’t think it’s a stretch to assert that computational ubiquity will supplant most everything we currently think of as "media."

strange attractors

The ubicomp field, from an outsider’s perspective, appears to be at that stage in the evolution of a paradigm where every passing year brings a consolidation in concepts, and an event such as this conference - the fourth annual such - serves primarily to unify subsequent research around those metaphors and methodologies that show the specific promise of being able to function as the nodes of such an accretion.

Despite a truly pointless keynote speech - a gee-whiz spectacle of uncritical extrapolation delivered by a lower-echelon science fiction writer, and amounting to "someday we’ll be able to do really cool things with matter" - the Ubicomp 2002 conference felt like it did just that.

I say "felt" because I approached Ubicomp with, clearly, very little understanding of what an academic research conference is all about - parameters and formats for position papers, accepted modes of written speech, and so forth. It didn’t seem to matter much; the session organizers were unfailingly gracious about allowing this clueless guy to come stumbling in and waste their time with his randomish observations.

I had traveled to Goteborg with every intention of discussing the promise of gestural interfaces between humans and computers, having been invited to present at one of the conference’s breakout workshops, and indeed that’s what the paper I carried along with me on my laptop was all about. At the last minute, though, I got cold feet. I began to feel as if nothing that I had to say about gesture was particularly original, and that at best what I could offer was a historical overview of the work that had been done in that area.

As it turns out, I wound up discarding even that. One predictable consequence of so much travel so unduly drawn out is a truly impressive sense of dislocation, almost discorporation. I wound up sleeping fitfully, waking into a dullish haze sometime just north of 05.00 on the morning before the conference was to begin. And as I struggled into something approximating consciousness, two words kept alternating in my mind, blinking like the display on an unplugged VCR: panoptical and forgiveness.

It’s those two words, and their implications, that I wound up talking about at the workshop, and that I’d like to continue thinking about here.

As you know if you read this site regularly, I’m not a developer. I’m not a technologist or an engineer, and my technical background is limited to say the least. No, I’m an information architect, and part of what that means is that I get to spend a great deal of time with the people who use complex technological systems, devices and artifacts - systems, like the very one you’re reading this on, that have a great deal in common with those being dreamed at Ubicomp, the "context-aware wearables," "face-responsive interfaces" and such.

There’s a family resemblance among these artifacts, and therefore I think it’s not unreasonable to assume that insights gleaned from the studying the interaction between human beings and the World Wide Web might prove useful when planning a generation of technologies that enter our lives at a still-more-intimate level. They’re not particularly profound, but I think they are real and important. I’d like to share a few of them with you.

through a scanner darkly

Human beings both fear and long for total surveillance. Many of you are familiar with Jeremy Bentham’s figure of the Panopticon - a notional prison in which every cell is exposed to perpetual surveillance by an observer located in a central tower - and how the philosopher of power Michel Foucault, in his book Discipline and Punish, used this as a model for the workings of control in societies that have evolved past the messy need for physical sanction.

The Panopticon is usually trotted out as a polemical figure, but what those who do so generally fail to understand is that many, many people like the idea that they’re always being watched, because they equate that watching with always being cared for. If the most accepted model for pervasive devices to date has been the Assistant, we should never forget that a competing model - one that holds strong appeal for a great many people - is the Superintendent.

"Someone to watch over me," indeed: someone to correct and smooth over our all-too-human blunders, to immediately suss out that a toddler’s gotten into the dishwashing liquid, and call Poison Control (or better yet, an ambulance) before the bottle hits the floor.

Someone to preemptively slow the car before the as-yet-unseen motorcycle can emerge from the blindspot alongside the bus. Someone to warn us when the cumulative cholesterol level of the groceries we’ve placed in the cart exceeds the recommended allowance in our current health profile - or even to protect us imperceptibly, instantly adjusting the properties of contact lenses or makeup to react to the increase in UV radiation when we step out of doors. This is one promise of ubiquitous computation.

More immediately, I think of the work family friend Dave Kutzik has done around Web-based assistance for elderly people who choose to remain in their own home: mediations and interventions that promise to keep us independent for longer into life, with more dignity and apparent autonomy. Is this surveillance? You bet. Will we understand it that way? Don’t be so sure. Many of us will want this sort of panoptical surveillance, badly, and once we have lived with it, we will wonder what we ever did without it.

amnesty international

The flipside, of course, is that human beings absolutely require moments of disconnection, of amnesty. For various ethical, political, social, psychological and even spiritual reasons, we should be careful to design moments of silence, of unawareness and disconnection, into our ubiquitous systems.

In this regard, I have a story I like to tell about a former coworker of mine in Tokyo, a fellow American. It only speaks to a single aspect of the desire for disconnectedness, but it can stand as synecdoche for the rest.

She had been suffering with ongoing back pain for some weeks, and finally decided to seek out a massage to relieve her discomfort. Now, the neighborhood of Tokyo in which she was living, Roppongi, is indeed full of enterprises that advertise massage - the only trouble being that these are not precisely clinical healthcare establishments.

Unable to read the language, or pick up on any of the semiotic coding in the ads that might have saved her some confusion, my coworker went ahead and booked an appointment with one of these salons. She did note the hostess’s puzzlement when the customer turned out to be female, and shrugged it off; it wasn’t until the young lady who came into the room with her began to remove her own clothes that comprehension started to dawn.

A few moments of awkward, half-gestural explanation ensued; the young lady put her clothes back on. And an hour later my coworker paid and left the salon no worse for wear, not particularly competently massaged but certainly with a new story to tell.

And tell it she did, the next morning, self-deprecatingly and charmingly. Ordinary mistake, ordinary confusion: one of those "it could happen to anyone" miscues that merely makes for a few laughs at one’s own expense.

My coworker chose, of her own volition, to share this story with us, but there are any number of reasons why she might not have wanted to. Perhaps she wouldn’t care to make herself look a little foolish. Perhaps she wouldn’t want to let on to what might be construed as a condition of chronic debility. It’s even conceivable that she might have simply been curious what it felt like to pay a woman to take her clothes off and climb onto the table alongside her, and have no desire to disclose this curiosity to anyone. Regardless of her motivation, the current consensus in most sectors of Western culture is that it is assuredly her business should she choose not to explain herself, or relate the tale at all.

pleading the fifth

But consider what might have happened if this anecdote had taken place in a world where every shoe and every stoplight has an IP address, where presence and location and transaction are all things that are infinitely transparent and knowable - as if Metatron, the Recording Angel, could access credit-card receipts and GPS traces, facial-expression scans and airborne pheromone gradient telemetry.

Family, coworkers, friends. Management. Insurance companies. Law-enforcement agencies. If they’re all tied into one seamless fabric of perception, analysis, and response, it doesn’t take a whole lot of imagination to see that human beings, in the ordinary course of their everyday lives, will accumulate not a few instances in which they would prefer to remain unseen, even at the cost of their temporary inability to use those other services a ubiquitous network affords them.

(I’m thinking, too, of those assembly-line workers caught up in a recent controversy in America, regarding their company’s demand that they justify the timing and duration of their bathroom breaks.)

In fact, we probably don’t want literally "always-on, real-time, anywhere" access, at least not if this access goes both ways - which it has to, by definition, for any of the other good things we want to be enabled. We want places to hide from the all-seeing eye, to be forgiven of the obligation to report ourselves, explain ourselves, inform on ourselves.

This is something we’re already seeing as prologue with mobile phones: "coldspots," impromptu exclusion zones where clever hacks locally deny the transmission or reception of wireless signals. I came to Goteborg hoping that ubicomp could avoid the usual tech industry "arms race," or co-evolutionary cycle of provocation and adaptation; I wanted to ask instead that we design moments of forgiveness and amnesty into our architecture from the very beginning.

What I was very happy to see at Ubicomp was that the leading theorists in the field are well aware of this need, to the point that in retrospect it seems a little presumptuous of me to have "asked" such a thing at all. Privacy - and the differences between legitimate and illegitimate ways in which a ubiquitous system can be used to determine whereabouts and infer intention - popped to the surface as a concern as soon as the first Active Badge installations were a reality.

Indeed, the discourse is several steps beyond that already, with privacy well understood as something that can kill the prospects for ubicomp all by itself if not fully and transparently addressed - including the subtle and thorny issue of how to avoid managing to incriminate ourselves by the simple act of seeking privacy.

ubiquitously yours

Assuming these tensions are addressed and held in balance, what will the daily experience of a ubicomp world feel like?

The Japanese designer Naoto Fukasawa speaks of "design dissolving in behavior," of interactions with designed systems so effortless and well-thought-out that they abscond from awareness. This is a goal I’ve (here and elsewhere) held up as a gold standard for the Web, and for the other devices and artifacts that increasingly litter our days: that if they are everywhere, like the air, we should at least only be as aware of them as we are of the air.

Asking the same thing of a ubiquitous computational environment seems only reasonable, following Weiser’s beliefs that "people live through their practices and tacit knowledge" and that therefore "the most powerful things are those that are effectively invisible in use."

There are other points of view, though, just as valid and just as respecting of the depth of the user’s humanity.

A key concept I picked up at the conference, thanks to the University of Glasgow’s brilliant Matthew Chalmers, was "seamfulness": the idea that, instead of pretending that we’re surrounded by and immersed in one perfect envelope of digital empowerment, the joints and moments of connection should be exposed to user view.

Matthew gave an example framed in terms that most of us will understand: the variation in signal strength from one to another cell of a mobile phone network. Most cellphones represent signal strength with a simple, rather uninformative bar graphic, but one or two open this decision to user input: such-and-such cell offers a stronger uplink than the one you’re currently in, would you care to move in that direction?

We’re familiar with this tension from the desktop: it’s the difference between a UNIX command-line and an OS X bauble. Obviously I think both approaches are valid - for different types of users, under divergent circumstances. The difficult part is in anticipating what options are appropriate to offer - what seams are reasonable to expose - and when.

Too, none of this should be imposed in a developmental vacuum. We’ve cumulatively learned enough in the last half-century of software development to understand that "you are not your user," that design that takes place in the absence of deep, ongoing user research is bad design.

At one point I asked a panel of the community’s thought leaders what other measures could be taken to ensure that the usability of ubicomp applications exceeds the average found on the Web - a standard which I described as "atrocious." Gregory Abowd, one of the field’s prime movers, replied that he actually thought of the Web as being quite successful, that it is impressive how much of the time it actually enables us to do, with relative ease, things that were impossible to achieve for any amount of love or money just a few years ago.

To which I can only reply: indisputably (and wonderfully) true, but that’s just looking at things through one particular set of lenses. There’s also the user experience to consider, and one of the things I get sensitized to as a consequence of my work is not merely how frequently Web sites fail to meet their users’ expectations, but how often those users then go on to blame themselves.

They internalize the blame for not being able to find their way around poorly-imagined sites, and they thereafter approach new technological challenges with anxiety and disdain. Not everyone, by a long stretch, but enough to matter.

more human than human

And if this is unacceptable on the Web, where at least you can switch the damn thing off, it’s still more worrisome when the system under discussion is by definition one that will insinuate itself into every available aspect and moment of a human life: nobody wants to have to confront incomprehensible alert boxes or the Blue Screen of Death when all they simply set out to do was water their plants, order a pizza, or administer anesthesia.

My hope is that, with care and attention to the needs of human beings in all the texture and mystery of their real, everyday lives, we can design systems that will support us when we need support, operate imperceptibly on our behalf when required, and just as imperceptibly fade from the picture during those valuable moments when we choose to go it alone. (Or, alternately, in the seamful version, shift in and out of our awareness in a way that affords experiences of beauty.)

And when I say human beings, I mean just that, and not the abstractions that we allow to stand in for same when we develop Web sites: the just-so personae we gin up and fool ourselves into thinking that they stand for anything real.

Nurri had given me a book for the long flight, a volume edited by Paul Auster called "True Tales of American Life." It’s a collection of narratives sent in to a show Auster hosted for the American National Public Radio network, with the only requirements being that the submissions be short, and that they be true. The collection is beautiful, precisely because it’s not literature. The stories are sentimental, haunted by the disproportionate weight of coincidence, occasionally clumsy. At every step, the authors reveal themselves to be in search of attention, praise, validation: in short, reveal themselves to be human.

Thankfully, real ubiquitous computation arrives at a moment when we have the formal and methodological ability to account for this variability and imperfection. We’ve learned, the hard way, that ethnographic studies and contextual inquiry and repeated rounds of user testing improve our designs immeasurably, by forcing us out of the position of godlike arrogance we tend to adopt as designers (frequently without noticing) and into a more collegial relationship with the people we build systems for.

My hope is that ubicomp, as it moves toward practical realization, incorporates these perspectives. And it had better, because the alternative is too depressing to contemplate: innumerable daily interactions - intimate and subtly all-encompassing - each with the interaction quality of reformatting copy in Microsoft Word. To me, the choice is clear.