I would like to start this talk with a little story about the cult of the cargo plane in Melanesia. The natives of Melanesia believe that planes come from paradise and that their ancestors send them, but that the white man, who is a crafty pirate, manages to get his hands on them by attracting them into a big trap - the airport.
So what do you do? You build your own plane, and you wait, and you have faith, because sooner or later, your ancestors will discover the white mans trap and will guide the planes to your landing strip, and then you will be rich and happy. So these people build things that look very much like planes, essentially from bamboo and wood and whatever they can find; but you could say that their mental models of airplanes are somewhat problematic.
With this in mind, Id like to tell you a bit about what ubiquitous computing is. I call it ubiquitous computing (the conference calls it pervasive computing, but its basically the same thing), and why it might be the next computing paradigm. And Ill try to point out why designing for ubiquitous computing is very hard, and what we can maybe do about it.
At the Future Applications Lab, we do what we call applied research in the future of information technology. In other words, we think, we invent, we build and we test cool new computer stuff. In doing that, we are following a grand tradition of previous success stories in this area. Artificial intelligence taught computers to think, so that we no longer need to think ourselves, this was in the 1950s already. The paperless office, youll remember, was in the 1970s, and now we only use electronic documents. I know thats true in my office.
Virtual reality was even better, because we took away not only the paper, but also the office, so you could spend your whole time inside a virtual world. And then we have things like nanotechnology, so you can turn water into Champagne; intelligent agents as you know are now roaming the internet, gathering information for you; and the latest success story in this tradition is the mobile internet, which we all know is having such a huge economic impact in the early 2000s. And now were all using those fantastic services that people invented at that time.
So here I am talking about ubiquitous computing. Now, if we take a look at what we think of today as a computer, that was really also based on research. The Xerox Star was released as a commercial product in 1980. It had essentially everything we think of as a computer interface: windows, icons, mouse, pointers, it had an ethernet connection, a laser printer, PostScript, all of that. And it even has some interesting random coincidences that still define how we think about computing.
For instance, the terms cut and paste come from the fact that the Star was based on ethnographic studies of book publishers, who would sit with long sheets of paper and use big scissors to cut paper and then paste it into other documents. And we still do that today, because of how book publishers worked in the 1970s.
Also at Xerox Parc, where this research was done, Mark Weisers group started work on what they called ubiquitous computing in the late 1980s and early 1990s. Rather than basing the design of a computer on a typewriter as was the previous paradigm, they started basing it on human measurements and they made computers of different sizes, inch size computers, foot size computers, yard size computers, big board computers and they also tried to use the fact that handwriting is a very natural way of interaction.
In the ubiquitous computing Parc they had a lot of things which we still almost have, or would like to have, like hand-held computers, wireless networking everywhere, interactive wipe boards, notepad interfaces, location awareness. Were still not quite where wed like to be, because the ubiquitous computing vision is much more than that, with computation seamlessly available everywhere we need it, supporting and augmenting human activities. Digital information then flows freely between people, responsive environments, and everyday computational things.
So how come I cant even connect my pocket computer to my mobile phone, you ask?
Because ubiquitous computing is really hard, there are a lot of really deep technical problems that you really need to solve before you have a chance of getting anywhere near this vision. So, we have networking infrastructure, we have software we have power management, sensing and perception, communication standards, miniaturisation, and so on. And thats why we have a whole conference, the Ubicomp Conference, which is dedicated to figuring out how we can make useful applications while still solving some of these things.
Now Id like to tell you another story, which is the story of the magic flower stick. And this is a true story, but Im not identifying anyone because it could be about any one of us really.
I was once shown a prototype of an intelligent flower stick. This was a stick that you would put in your flower pot, just like an ordinary flower stick, but it would do all kinds of wondrous things. It would start monitoring the plants well-being, so it would know if I was watering it, if I was talking to it. It could communicate the plants feelings to me, it could communicate with other plants, it would run forever by getting its energy from the sun and the soil, like the flower did, and so on.
So this person was waving this stick around and talking about how fantastic it was. And Im going, Hey, thats just an ordinary wooden flower stick that you have painted silver and glued some electronic components on.. Wait a minute, whats going on here?
Then I started thinking, what does this really mean? And this is a generic example, we see hundreds of these in design schools and in real research labs and concept laboratories. Its what I would call the cargo cult design.
You essentially build a fetish object that has the appearance of a real artefact, but it doesnt actually do what you have promised it will do (it doesnt have to be an artefact, it could be an interactive surface on the internet or on the mobile internet, or it could be a business plan). What youre doing in effect, is praying to the gods of technology to come and sort out all your real problems, while youre waving this flower stick around and promising it will do things which it cant yet do.
What about that paperless office again? Of course, Im one of those that do these cargo cult designs, frequently because the technology that I build doesnt necessarily work in the way that I would hope. Its important to think about this, because if you want to invent an amazing new product, its really easy if you dont actually have to make it work. And its easy to make good technology, without making people use it or sell it.
This is dangerous, because we ignore the real problems, or we pray to the gods to sort them out. I think that the only way that we can explore the design problems of ubiquitous computing is to build and test real stuff. And in testing, I include making people actually use it, and asking them questions, and observing what they do with it.
The most important and the most difficult thing is to keep a balance. On the one hand is the fun blue-sky innovation part and on the other is the technical, or human feasibility, of what we can actually do, and you dont want to lose either of those.
Now I want to mention three key ubiquitous computing problems, and offer three solutions that I and my colleagues around Europe have been working on.
THREE KEY PROBLEMS:
1) Interfaces and interaction design
Were saying that computation should be seamlessly available everywhere, but what do we mean by that, and how can we integrate digital information into the human environment?
2) Power and networking
How can we make objects communicate, and how can we tell them to communicate; and - even more importantly, but often ignored - communication takes a lot of power, so where can we get that power from?
3) Sensing and context awareness
If we want to build proactive applications that actually sense what the user is doing, how can we do that? How can we make applications and objects that are aware of their surroundings?
THREE EMERGING SOLUTIONS:
1) Informative art as interface
Informative art takes the appearance and function of ordinary wall-hung art, but it is not art at all, but flat panel displays showing dynamic computer programs. These display information, but make the artefact fit into the human environment. You dont see any figures or icons on that, you just see something that hopefully looks like art.
Some examples (indicates screen): the top one is an informative Mondrian. It can show the weather, in this case, in six cities around the world, and its doing it by letting each of these squares represent a city. The colour of the square shows the weather condition and the size shows the temperature, so from this I can see thats its raining in Rio de Janeiro, while its sunny and cold in Gothenburg, and so on.
The bottom one is an informative Andy Warhol with soup cans. It shows the passage of time and other quantitative information through changes in the appearance and number of soup cans. So what you have is an information display that still is an attractive everyday object.
2) Networking with pin and play
The Pin and Play project at the University of Lancaster consists of a surface to which you can pin different smart artefacts. You can get your power from the surface and you can also let the artefacts communicate with each other, because theyre also networking on the surface.
Heres an example of how you can use this. Here are some pins with a deadline, so if you a pin a number of things on this surface and you make it into a notice board, the pins can get their power through the surface and communicate and find out which pin is currently holding the most important note. Then the most urgent or important pin lights up.
You can pin other stuff. If you have this material covering your whole wall, you can put your light switch in one place and your light any other place you want. And since they get their power through the surface and they talk through networking you can flick your light switch and your light will turn on and off. This offers new ways of thinking about interior decorating and amateur electricians would like this.
3) Context awareness with Smart-Its
For a European Union project we have built something called Smart-Its, which is basically a stick on computer that you can stick onto any object that you want to make smart and intelligent (well, not intelligent, but at least aware). Its a sensor board with (currently) five sensors, and a communication board for wireless communication.
One good example of this is proactive IKEA furniture. You put sensors on the different parts of IKEA kit furniture, and the sensors tell you if youre doing the right or wrong thing when you assemble it, so they can figure out the rotation, which pieces fit together, and if youre screwing in the right or wrong direction with the screw driver, and so on.
A second example: if you leave your sensors on after youve assembled your table, you can make things like the weight sensor table at Lancaster University. This has a sensor on each corner, so when you put an object on the table, the table knows where it is. It can also make some assumptions about what it is, because it knows the weight. And it can track objects on the table. You can even use the table as a sort of remote control bug, just by placing your hand and dragging it you can control a mouse pointer or something else, so in essence the whole piece of furniture becomes a sensor in itself.
I would like to leave you with this thought: ubiquitous computing is really hard, so dont put too much faith in the gods of technology. We might get there eventually, but theres still a lot of work to do in all the areas Ive discussed. As a designer you should really try to keep it real and be aware of the technical and human limitations of what youre doing. If at all possible, its very important to make working prototypes for all designs, and actually try them out so real people use them.
So with that I thank you and I say, Keep watching the skies!
QUESTION TIME WITH JOHN THACKARA
John: You did a great job, because youre in a flow of research and papers from all around the world on the technology of ubiquitous computing, so I gave you the impossible task of telling the whole story in 20 minutes.
Lars: The brief was basically a two-day lecture.
John: But weve heard in the previous session about the beginnings of connections between people with different bodies of knowledge. You mentioned that some designers or product developers just make assumptions about technology. Is it your impression as an engineer that youre getting better communication with developers at an earlier stage?
Lars: What kind of developers do you mean?
John: Whoever is going to use this stuff.
Lars: You mean the end customer?
John: You made a comment about not putting your trust in the gods of technology: is that situation getting worse or better?
Lars: I think its getting better. I think its like in the early internet bubble, people were doing web pages that could only run on the local computer because they had such high bandwidth and then youre putting your faith in bandwidth going up and up, which it didnt really do; so you shouldnt assume, because the only thing that really increases is the capacity of the transistors and so the speed and memory of computers. Things like power and interfaces are changing much more slowly, if at all. So thats one thing, dont assume that it is going to get better, and whatever youre going to do now is going to work in five years.
The other thing that I think is important is that we, who work on the technology side, must actually think much more about the design on the human side, which is now happening. The Ubicomp Conference is considered by many of the people who build the technology as a human-computer interaction conference because we actually do applications, we dont do networks and algorhythms. So I think it clearly needs to be a meeting from both sides, from the technology side and from the interface and design side.
John: Well follow it up in the panel discussion in a few minutes, but thank you very much Lars.