杏吧原创

A POPULAR song in the 1960s famously dismissed newly constructed Californian housing as 鈥渓ittle boxes made out of ticky-tacky鈥� that all looked just the same. And social critics have complained endlessly about the drab conformity of London鈥檚 suburbs鈥攕uggesting occasionally, in despair, that people who chose to live in identical houses could never be expected to have the revolution they so obviously needed.

Now a different revolution could be coming. Software developers, engineers and architects are promising to transform the construction industry with a mixture of computer-aided design and the latest in industrial manufacturing technology. Their ideas could turn the industry鈥檚 traditional hierarchy of house buyer, builder and architect on its head. In its place will come personalised architecture for all. Welcome to a world of infinite variety.

It could hardly be more different from traditional architectural design, in which sophisticated use of repetition is often a key feature. Take the Georgian squares of Dublin and the crescents in the English city of Bath, where houses in rows and arcs create grand urban elements. There鈥檚 a sense of civic decorum at work here: the sober exteriors of the individual units are like discreet business suits around a boardroom table. Only the doors with their fanlights鈥攖he architectural equivalents of neckties鈥攑rovide the opportunity to express some individuality.

More often, this uniformity arises simply from the cost and time required to custom design each dwelling. Right now, the world of residential design mirrors that of the fashion industry: the affluent few can afford expensive, bespoke products, while the majority get cheaper, standardised, off-the shelf designs. To minimise design and construction costs, most new housing throughout the industrialised world is built as blocks or estates in which each unit looks similar to its neighbours.

The dream is that new houses could be designed and built in a completely different way鈥攁 way that puts the responsibility for design not just on architects but on the people who will buy the dwellings and live in them. Rather than simply purchasing one of the many faceless boxes on offer, every buyer of a new home would sit down with an architect and design it exactly the way they wanted it.

To make this happen you need ways to simplify, or even partially automate, the design process; there simply aren鈥檛 enough architects around for it to be possible with conventional design techniques. In practical terms, this could mean creating architectural expert systems, suites of software that can apply a 鈥渧ocabulary鈥� of design elements, together with rules for combining these elements, to produce complete buildings. Just as important, there鈥檚 a need for cheap ways to translate these designs into bricks and mortar, or metal and glass鈥攁 requirement for new manufacturing techniques that the construction industry can use to produce custom-shaped elements in whatever size and material the designs call for.

It turns out that the first requirement鈥攖he idea of creating a formalised vocabulary to aid construction鈥攊sn鈥檛 new. It was a strategy originally pursued by the Italian Renaissance architect Andrea Palladio, for example, who designed numerous villas and palaces in and around Venice in the 16th century. Palladio employed a set of standard room shapes and proportions, together with standard construction elements such as cylindrical vaults and classical columns, and combined these elements according to strict rules of repetition, alignment, proportion and symmetry. Once he had made a few key decisions about dimensions, materials and motifs, much of the rest of the design followed logically from the layout and his rules of construction. He had created a system with which he could progress rapidly from simple plan sketches to a complete design.

The same strategy can work with modern architecture too. For example, the distinguished Portuguese architect Alvaro Siza adopted the idea in his housing project at Malagueira in Elvora which he began in 1977. He provided a kit of crisp, modernist housing parts鈥攔oom shapes, wall units, doors, windows, stairs of various types, and so on鈥攖hat could be assembled in all kinds of ways using simple rules to produce house designs that responded to the varying needs and budgets of home buyers.

The rules encoded a great deal of knowledge about domestic layout. So the high-priced senior architect didn鈥檛 have to design all the individual units himself; he could leave it to assistants, or even to clients themselves, to work out the details of the units and still be sure they wouldn鈥檛 go far wrong. They could decide on things like sizes and shapes of rooms, positions of doors and windows, and the external layout on a case-by-case basis. The technique also provides the sense of stylistic consistency and urban coherence that you feel so powerfully in places like Georgian Dublin, but without the inexorable repetition. Instead, you get engaging variations on a theme, as in a clever musical composition.

This idea has been further
developed by researcher Jos茅 Duarte at MIT. He has computerised Siza鈥檚 vocabulary to produce the prototype of an interactive computer system for the design of customised mass housing. Duarte formalised Siza鈥檚 system by expressing it as a 鈥渟hape grammar鈥�, a structured language much like the sort that has become familiar in computer programming. An architectural shape grammar has its own vocabulary and syntax rules, including rotations, translations, reflections, additions, subtractions and scalings. By applying these rules to a simple shape such as an L-shaped box, it鈥檚 possible to quickly create a variety of two or three-dimensional forms that retain desirable design qualities
(see Diagram). And it鈥檚 a system that is simple to adapt for use on a computer.

Duarte has demonstrated his shape grammar by building an online software package that can be used to design Malagueira-style housing. Visitors to the website can quickly create designs that meet their needs, in the time that would be required, in a more traditional marketing process, to visit one or two standard display homes. In effect, Duarte has provided an efficient Malagueira house server, or a Siza amplifier.

Eventually such a program might take the form of an interactive computer-aided design system that takes customers through the design process from start to finish. It could use artificial intelligence to search through the housing designs specified by the grammar to find one that meets specific functional requirements, site constraints and budget limits. The result could be an expert system which offers a user the skill and experience traditionally provided by an architect or designer, but as a software package with an easy-to-use interface.

The process could eventually become similar to the now commonplace one of configuring computers online, as pioneered by computer maker Dell. Automobile manufacturers are moving towards similar processes for cars. The general idea is to replace mass-produced, standard models with efficiently mass-customised ones.

The efficiencies of mass-customisation really begin to emerge when you take automation beyond the design stage. The decisive step is to integrate a customer鈥檚 online design with flexible, computer-controlled manufacturing. In other words, the information generated online should drive the house production line. To make a timber house, for example, the designs might be downloaded direct from the house buyer by computer-controlled machinery that would cut the studs, joists and side panels needed to build the house, ready for delivery to the construction site. This approach is already being used by some conservatory manufacturers and in the custom packaging industry. The Austrian start-up Aie.tech, based near Salzburg, is developing technology for computer-controlled, automated assembly of masonry houses, in which entire walls, floor and roof elements are constructed on a robotic production line before being packed off for rapid construction on site.

Once computer-aided design and automated assembly plants are integrated, the possibility of radically new architectural languages opens up. Traditionally, buildings have mostly been made up from shapes that are simple to draw, fabricate and assemble鈥攕traight lines, planar surfaces and arcs of circles, for example. But a computer-aided design and fabrication system can handle complicated curves as readily as straight lines or arcs, and a computer controlled cutter鈥攁 laser or a water jet, for example鈥攃an cut them out swiftly and accurately. Suddenly non-standard shapes do not carry the cost penalties that they did in the past, making it reasonable to develop architectural languages based on more complex surfaces.

Repetition is no longer necessary. Classical architecture depended upon vocabularies of standard design elements that could be produced by craftsmen from templates. And modern architecture has relied heavily upon industrially mass-produced elements such as bricks and tiles. In an extreme form this has given us industrialised component building systems such as the precast concrete post, beam and panel systems that were popular in post-war Europe, and which produced some remarkably grim housing. Computer-controlled mass-customisation can free us, opening up the possibility of creating buildings from an unlimited number of individualised parts. The elements of architectural composition need no longer be translations, rotations and reflections of simple shapes. Elements can morph their shapes in non-linear ways as designers manipulate the values of key variables.

So far, architectural designs that feature curved surfaces and non-repeating parts have mostly appeared in expensive public and institutional buildings. Take a look at Norman Foster鈥檚 elegant new roof for the Great Court at the British Museum in London, for example, and you鈥檒l notice that it is made from computer-designed and fabricated structural ribs that vary smoothly in shape, rather than rigidly repeating units. This allows the roof to adjust gracefully to the unusual shape of its surroundings, and to the unique requirements for structural support this creates. Even better known is Frank Gehry鈥檚 Guggenheim Museum at Bilbao in Spain, which relied heavily upon computer-aided design and fabrication technology in its design and construction. Gehry also applied these techniques to the Neuer Zollhof building in D眉sseldorf. Its curved facades are made from individual plates, each cast from polystyrene foam moulds shaped by a computer-controlled cutting machine.

Now some architects such as Los Angeles-based Greg Lynn have begun to experiment with the possibilities of using similar techniques for relatively low-cost housing. Lynn has used a computer-controlled mill to create a prototype for mass-customised housing called the Embryological House, a structure shaped like a cell. And Finnish architect Jarmo Suominen is developing a Web-based software system called VIRAPS, the Virtual Apartment Server, that will allow house buyers to participate in the planning and design of their homes.

Keeping costs low will be a crucial factor in making customised housing a reality, and engineering companies are already devising low-cost, flexible housing systems. You can already buy homes in Europe and the Caribbean built from mass-produced fibre-reinforced composite panels which can be put together to create a wide range of layouts and styles. Another approach involves modular panels made from materials such as concrete that slot together, and come with electrical and plumbing services already fitted. Although they are designed primarily for industrial and commercial use, there is little reason why they can鈥檛 be used to create customised housing too. And at MIT, the Chassis and Infill project is taking technology developed for the automobile industry and adapting it to make low-cost buildings. The 鈥渃hassis鈥� takes the form of fibre-reinforced composite beams and columns, which can quickly and cheaply be turned into a complete house by fixing together mass-customised units for the walls and floors.

The effect of the digital revolution on architectural form will be the reverse of the effect of the Industrial Revolution. Industrialisation provided new and vastly improved materials, notably steel, glass and reinforced concrete, along with repetition, standardisation and mass-production. The digital revolution is supplying new methods of design and construction. The characteristic housing type of the industrial era was the high-rise residential tower, with its repeating unit plans and elevation details. That of the post-industrial age may be the computer-customised dwelling that can vary to reflect the diversity of urban populations and their needs.

Not everyone will like this. Whenever new technologies take intellectual and aesthetic creation out of the hands of an elite and open it to a wider public, the threatened elite scream about reduced quality. It occurred when printing and mass literacy took reading and writing out of the hands of a few priests and scribes, when the Macintosh computer let anyone dabble in typography and graphic design, and most recently when digital video technology allowed anyone to shoot and edit their own films. Certainly the possibilities for bad and amateurish housing design will multiply. But I would argue that putting powerful design customisation tools in the hands of the public is generally a good thing. The results might not always be of uniformly high quality, but that鈥檚 a price well worth paying.

However, we鈥檙e left with a more difficult question. Given the choice, will we really want to make all our houses different? Will homebuyers choose to express individual difference, or conform to some norm? Will they choose the excitement of innovation or the comforts of tradition? Where, in the end, is uniformity bred鈥攊n your technology or in your head?