Digital Cities

The forms that cities take, the ways they function, and the mixes and distributions of activities within them have always been influenced very strongly by the capabilities of their underlying network infrastructures. Furthermore, cities have often been transformed by the introduction of new infrastructures. It is impossible to imagine Rotterdam without its canals and connection to the North Sea, Chicago without its railroads, Los Angeles without its freeways, or any large modern city without water supply, sewage, electrical, and telephone networks. Today, a new type of network infrastructure — high speed digital telecommunications — is being overlaid on cities everywhere. Its effects will be at least as revolutionary as those of the new network infrastructures of the past. It is already causing traditional building types and neighborhood patterns to fragment, recombine, and form startling new arrangements. This process will continue and accelerate. In this paper I describe the new digital infrastructure, analyze its major spatial effects, consider some illustrative examples of the resulting fragmentation and recombination, and discuss possible design responses with particular attention to social equity and long-term sustainability. Digital cities are being developed all over the world. Digital cities integrate urban Let us ask some fundamental, practical questions about the role of digital telecommunications and ubiquitous computation in shaping our future cities. What new opportunities do these technologies provide to produce cities that are attractive, equitable, and sustainable? What unwanted side-effects must we contend with? And what strategies should architects, urban designers, and urban planners pursue to take maximum positive advantage of the potential benefits while avoiding the possible downsides?1 Network Infrastructures We can best approach these questions, I believe, by reflecting upon the roles of earlier network infrastructures — water supply and sewer systems, streets and roads, canals, railroads, electrical grids, telegraph, telephone, and broadcast systems — in forming urban structures and patterns. Most obviously, these networks augment the 1 These strategies are discussed in more detail, with extensive supporting documentation, in William J. Mitchell, City of Bits: Space, Place, and the Infobahn (MIT Press, 1994), and William J. Mitchell, E-topia: Urban Life, Jim — But Not As We Know It (MIT Press, 1999). 2 William J. Mitchell affordances of the places they serve, and so support activities that would not otherwise be possible there. Furthermore, they allow greater concentration of human activities by connecting urban locations to distant hinterlands. Thus, for example, a piped water supply system allows habitation of sites that would otherwise be too barren to support life. And, by drawing upon a far-flung catchment area, it can support a much greater population than local water resources would otherwise allow. Similarly, sewer systems disperse waste that would otherwise accumulate locally, road systems allow trade with the food-producing countryside and other cities, electrical supply systems mitigate the effects of darkness and climatic extremes, and so on. We have to consider not only the immediate effects of particular networks, but also the ways in which multiple networks interact to produce joint effects. An irrigation network might allow a desert location to produce crops, for instance, but it is also necessary to have road or rail access to get those crops to market. Water supply and transportation networks may allow population to concentrate at a location, but this population will not be sustainable unless there is also effective waste removal. As geographers and planners have long-since discovered, the interactions of these networks with patterns of land use are complex. On the one hand, construction of networks creates the possibility of new land uses at the locations served. On the other, existing land uses generate demands for network service. Urban spatial development is best understood, then, as a recursive process, unfolding over lengthy periods of time, in which network infrastructures and land-use patterns evolve by continually responding to one another. When a new type of network infrastructure emerges, it is not deployed across homogeneous terrain; it is overlaid on a spatial pattern that has developed in response to its predecessors. Typically, by creating new relationships among existing activities and introducing new activities, new network infrastructures produce significant transformations of such existing patterns. Since the industrial revolution, in particular, we have seen the effects of overlaying modern transportation, electrical supply, and telephone systems on older urban fabrics. Conversely, we have also seen the effects of removing network infrastructure or getting bypassed by it; there are numerous sad examples of towns and cities that have declined when railway or riverboat service ceased, or when interstate highways passed them by. What, then, are the effects of overlaying digital telecommunications on existing urban patterns? How will this new type of network interact with existing ones? And what sorts of transformations can we expect to result? Fragmentation and Recombination The basic function of digital telecommunications, of course, is to allow human interaction at a distance. To the extent that remote interaction successfully substitutes Designing the Digital City 3 for face-to-face, traditional requirements for adjacency among activities — the bonds that have always held buildings, neighborhoods, and cities together — are eliminated. However, only some — by no means all — such bonds are loosened or removed. You may now do your banking remotely, for example, thus eliminating the need for faceto-face interaction with a teller at a local branch bank. But, if you want to get your hair cut, you still need to go to the hairdresser’s for a face-to-face interaction. The net effect is neither decentralization of everything nor rampant centralization (as some early commentators had suggested), but a complex process of fragmentation and recombination of familiar building types and urban patterns. It is much like a chemical reaction in which some bonds are broken, others remain, new ones form, and a new compound — with interesting new properties — results. Consider, for example, the now-familiar effects of online bookselling. By making use of Web sites instead of traditional sales floors, online book retailers such as Amazon.com radically decentralize the activities of browsing and purchasing; instead of taking place in a few retail establishments at central locations, these activities are distributed to Internet-serviced desktops in huge numbers of homes and offices. The space to accommodate these activities, which had once been grouped with other retail space, now fragments and recombines with domestic space and work space. At the same time, book storage and distribution functions equally dramatically centralize. They no longer need to be clustered with browsing and purchasing, as in a traditional bookstore, but are now performed at a few national centers located at convenient air transportation hubs; this allows both economies of scale and maintenance of much larger stocks than are possible in scarce and expensive urban retail space. And backoffice functions such as billing and stock control, which reduce to manipulation of digital data, no longer need to near either the books or the customers, and can float free to wherever teleworkers are available at a price the management is willing to pay. There are implications not only for location of activities, but also for transportation demand. With old-fashioned bookstores, books were delivered in bulk to the stores (that is, to intermediate storage points), then carried away by purchasers. With online bookstores, the emphasis shifts to express package delivery from the national distribution center to widely scattered homes and workplaces. You can carry out this sort of analysis for just about any of the emerging online retailing or service industries, and the results vary according to the natures of the particular products or services offered. Books are small, high value, imperishable, and a delivery time of a day or two is generally acceptable, so national distribution centers make sense. The same goes for music CDs, videos, consumer electronics, and many drugstore items. But groceries are bulkier, less valuable, and more perishable, so they demand regional distribution centers rather than national ones, and fleets of specialized local delivery vans rather than national package express systems. Hot pizzas are even more perishable, and require nearby local production and distribution centers. On the other hand, computer software, digital music recordings, and digital videos can be delivered online (provided that the bandwidth is sufficient), so the distribution centers can be located just about anywhere there’s good network service, and telecommunication does not just restructure transportation requirements but completely substitutes for transportation. 4 William J. Mitchell Finally, we should not forget that delivery points for products and services — homes and workplaces in particular — are likely to change in response to their new roles within these systems. At the very least, they need network connections for placing orders; today, these typically take the form of PCs running Web browsers, but we are likely to see increasing use of smart appliances and closets that electronically order their own supplies, and sensor-equipped spaces that can summon medical and security services. The humble mailbox is likely to evolve into a larger and more sophisticated repository that can keep perishable goods from deteriorating and keep high-value goods secure. And home TVs, VCRs, CD players, radios, and videogame conso