Co-operative Buildings, the title of a recent conference in Darmstadt*, covers a variety of developments in integrating buildings and computing. One of the inspirers of much of this work is Mark Weiser from Xerox parc at Palo Alto in the us, with his concept of ubiquitous computing. In his vision the desktop computer disappears. There are chips in everything, with the building fabric acting as interfaces. In a sense every surface becomes a screen, except it addresses other modalities as well as vision - sound, touch, gesture. And it is smart, aware of who we are and what we are involved in, providing context-related information and facilities. For example, it might offer information from a database half-way round the world. It might offer an electronic white-board facility to support a group meeting.
It is Weiser's job to dream. Others must build something. Norbert Streitz and his colleagues at the German National Research Centre for it are developing i-Land, a project to develop a work and collaboration environment for new work practices and organisational change. A meeting room supports whichever individuals are there - it is active, attentive and adaptive. For example, while you pause in a corridor the electronic wallpaper might be used as a computer interface.
The current experimental focus is Roomware, a set of facilities for supporting individual and team working without a fixed desk. There are three visible components so far:
CommChairs - chairs with a built-in power supply and wireless networking that either provide a docking station for a portable computer or have similar computing facilities built in
DynaWall - a large wall display surface which can handle text and written input, including diagrams. These can be moved about in response to simple gestures. There are supporting computing and database facilities. Potentially, video conferencing could be included
InteracTable - a touch-sensitive tabletop screen for group discussion. A displayed document can be turned round to face you with a simple rotation gesture.
All these devices are linked by wireless network. So, for example, the wall display can be annotated from the chair screen. While the devices exist as working prototypes, the operating system software to orchestrate their interaction is still under development.
What people may find difficult is conceptualising virtual information spaces across a building. One approach, presented by Michael Soor from the University of Karlsruhe, was to provide an existing three-storey building with a four-storey 3D computer graphics model. The first three storeys were a typical facilities presentation. The fourth floor was a virtual floor, trying to map out the information environment using a spatial metaphor of the relationships between databases and facilities.
One implication of treating the building as a set of interfaces is that we will use many of them and so need to transfer data readily from one interface to another. It is the equivalent of cutting and pasting within our pc screens. The approach by Jun Rekimoto of Sony Computer Science Lab is called 'pick-and-drop'.
All interfaces need to be touch-sensitive. A simple pen is used to pick the file, image, etc, from one screen by touching it with the pen and then dropping it sometime later on another screen by touching that. The neat trick is to give the pen a unique identification code. So the pen does not need to hold any data. As soon as the pen touches the screen it stores the data; when it touches the next screen the computer recognises it and drops there the data previously picked. (The video of people passing each other in the corridor and dropping data on to each other's palmtop computers en passant had a surreal air.)
Sony's most recent prototype of pick-and-drop goes a step further. The Data Picking Wand, no bigger than your hand, embodies a digital camera and infra-red sensors. It supports data transfer not only between digital devices, like computers and digital video, but also from paper.
All this is some distance from Weiser's image of ubiquitous computing, but a practical way of moving in that direction.
The intelligent room
Not everyone takes Weiser's line. Michael Coen of mit's Artificial Intelligence Lab is suspicious of embedding such fast-changing technology in the building fabric. He prefers minimal hardware modification to ordinary spaces. It is a view shared by Nicholas Kohler of the University of Karlsruhe, from a sustainability point of view. Kohler also argued that, despite some of the images of special future buildings offered at the conference, the focus should be on embedding future it in existing buildings.
From Coen's background in artificial intelligence, from the simpler world of industrial robotics, he is hoping to transfer capabilities of vision, speech and gesture recognition to the unpredictable world of everyday life in buildings.
His Intelligent Room can be, say, a conference room. The input devices are cameras and microphones (see diagram). Voice is used to interact with the room/computer to call up specific work facilities and environmental conditions. Cameras are used in monitoring who is where. 'Pointing cameras' monitor people's pointing gestures at displays, which are then interpreted by computer - an alternative to cursor control of the display using a mouse. Potentially, a more general understanding of a language of gesture might be developed.
Coen may be right about the risks of embedding it too intimately in the building fabric. But, as he acknowledges himself, artificial intelligence is not yet at the stage where general interpretation of sound, vision and gesture is robust enough for broad everyday use. So the nearer future may well come by evolving existing technologies, as Streitz and others are attempting, rather than making the big leap to usably smart computers.
Another thread of Weiser's argument is that ubiquitous computing makes it possible to make use of peripheral awareness and communication - like the 'cocktail party effect' in which, among the noise of many conversations, you suddenly attune to someone mentioning your name.
Sometimes we want to be alerted to something happening - the existence of a message, rain outside - while focused on a central task. Weiser suggests that such information can be communicated via peripheral vision or other peripheral sensing, a near-unconscious mode of monitoring not currently addressed by computing, though part of our everyday experience. Ubiquitous computing provides an infrastructure for this through its many interfaces.
The Tangible Media Group at mit's Media Lab has been trying to make examples. Active Wallpaper can be made to brighten. The Water Lamp has three legs moved by servos and a tray of water above the bulb to create light ripples on the ceiling to order. A child-size windmill is driven by a small motor. Such devices could be used to convey messages in peripheral vision. The message may be directly analogous, eg the water movement indicates seismic activity at some location round the world or the windmill represents local wind speed. Or the meaning could be more tangential, such as movement indicating rising share values or the level of activity somewhere in the building.
So far many of these devices have been found to be not peripheral enough, too noticeable, after a while distinctly annoying.
Beyond the building
One of the more immediately practical projects is work at Darmstadt University of Technology and the Fraunhofer Institute on setting up a web site for trading in spare space. With moves towards space-use intensification and nomadic working, a market may develop in office space available outside the Monday-Friday, 9-5 slot. If such a market becomes well developed it could even become the foundation for funding mixed-use buildings.
While the focus of the conference was on individual buildings, John Worthington of degw ran a final session looking at the urban and regional scale. He suggested that the nomadic working facilitated by it has significant implications for development at this scale, that information-focused businesses will gravitate to movement nodes. 'As the office becomes the meeting point . . . points of attraction are going to be multi-functional, 16-hour usage, accessible by both public and private transport and with a distinctive personality.'
Worthington is currently involved in a study of Utrecht in the Netherlands which he sees as one such emerging node. Adjacent to its historic core, he plans dense living, working and commercial areas, directly related to the station with over 45 million passengers a year. These are brownfield sites, developing around the core. They will be 22 minutes by rail from Schiphol Airport, adjacent to world-quality exhibition grounds.
It may appear ironic that a conference featuring communication technologies, some of which could save people the need to meet physically at all, should finish on the built-format effect of people moving, and that facilitated by it. But, as so often with it, the new does not replace the old. It is not 'either-or' but 'both-and'. New sets of options are created.
* Co-operative Buildings. Norbert A Streitz, Shin'ichi Konomi and Heinz- Jurgen Burkhardt (editors). Springer. Lecture Notes in Computer Science 1370. 267pp. £25.50.