Information, the new material

smart-things-cover.jpg The mundane things around us are quietly getting a lot smarter, fundamentally changing how everyday objects are designed. Anoto pens, Roombas and Tickle Me Elmos are all computers at heart, but little resemble their word processor-running/DOOM-playing ancestors.

Thanks to Moore's Law it is now nearly as easy (and cheap) to incorporate information processing into a mass-produced object as it is to create a custom injection-molded plastic part. The capability to manipulate information and create behavior has become a component instead of the goal of digital product design. Soon, Internet-connectivity for bathroom scales and wireless sensors in running shoes will be expected options, like color and finish. We will take for granted parking meters that send us messages before they run out, just as we take for granted Bluetooth headsets, MP3 players and Netflix streaming.

This movement towards computational objects that don't look like what we currently describe as computers represents a fundamental shift in the design of technology. It blurs the edges between industrial design, product design, architecture and interaction design. It's been called many terms ("pervasive computing," "ambient intelligence," "The Internet of Things", etc.). I use "ubiquitous computing," the name Xerox PARC gave the trend in the late 1980s.

Ubicomp has been a long time coming. Xerox PARC made good progress in the early 90s (see this 1991 video), and momentum to develop useful embedded computing devices was starting to build, but then the Web hit. Attention and investment shifted to the Web, which only required traditional computers with a screen, keyboard and mouse. Digital hardware came to mean lighter laptops and mobile phones. Ubicomp disappeared from the public eye for a decade.

However, it didn't actually vanish. At the edges, outside of the mainstream consumer electronics and personal computer worlds, using small processors to create interesting behaviors became increasingly popular. Adding a cheap processor and some clever mechanics in a $15 plush toy let retailers sell Tickle Me Elmos for $35. Toy designers did not think of themselves as being in the computer business, because they (and other groups, such as car designers) used information processing tactically as a way to differentiate in a competitive market.

Now, mainstream ubicomp is coming back. The success of Internet services on mobile phones demonstrates that networked products can stretch beyond a laptop browser. The prices for CPUs have fallen below a threshold where incorporating them becomes a competitively viable business decision. Research labs have developed new technologies for embedding information processing in virtually anything. New businesses, such as FitBit, and Green Goose are based on the fact that processing is cheap, and you can include it in anything.

The idea of a single general-purpose "computation" device is fading into the same historical background as having a single steam engine to power a whole factory, or a single electric motor to power every appliance in a house. As it fades, designers and developers have to learn to design smart things that serve the interests, abilities, and needs of people. We must create a practice of ubiquitous computing user experience design.



  1. Yeah, because I’m going to be happier if my easy chair, my end table, and the beer bottle I’m drinking from are all networked.

    Look, computing is fun, but it has its place. The reason cell phones are popular computing devices is the same reason transistor radios were popular. Take anything that provides some utility and make it smaller and more portable, and people will love you for it. But the whole point is that you can take one little radio with you everywhere, and it’s no longer necessary to build radios into things. Everything else can be just… what it is.

    The smaller and more efficient computers get, the less ubiquitous they need to be.

  2. How does your vision of ubicomp match, and how does it differ from, Bruce Sterling’s notions of an infrastructure for (his term) “spimes”?

    1. I agree with Bruce (most of the time ;-), and cite him a fair bit in the book. I think that where we differ, if we differ at all, is that I believe that ubiquitous computing will come primarily in the form of familiar objects that are increasingly augmented with information processing until they become dependent on information processing to do their jobs (this, for example, is what’s happened with cars), rather than new classes of objects that are wholly unfamiliar. That’s why I don’t describe things as spimes all that often. Many everyday digital products today are spimes, but their users call them by other names–“running shoe”, “camera”, “toy”, etc. I wanted to focus in the book on the challenge of designing one of these things, so I didn’t discuss them as a class quite as much as Bruce (or Adam Greenfield in “Everyware”) has.

  3. Once upon a time, there were nerds into “wearable computing” — some of the most outré, like Steve Mann, strapped 15-pound video cameras to their heads.

    Sometime in the last decade these people shrugged and said, “Buy a cell phone.” This is a good thing.

    Now we have Camera Barbie, who lacks only facebook integration, and that only because of concerns about the target market.

    We will always need the cellphone type device — communication is fundamental to human nature, and a small device that can facilitate all forms is best: audio, video, text, push or pull, confidential or broadcast.

Comments are closed.