From Computer Aid to Computer Dependency

The phrase 'knowledge is power' long ago became rather trite as sayings go, but where the knowledge is in computer systems, and the power is enjoyed by architects, there may still be something to learn from it.

One of the key areas of anxiety in commercial architecture today is the projected future cost of computing. At present there is hardly a sizeable architecture firm in the world that is not deeply enmeshed in a programme of updating, reorganising, expanding, restaffing, outsourcing or desperately trying to bring under control its computer operations.

In the United States, at those collegiate weekend retreats beloved of senior corporate architects, the hot topic is no longer a possible exit strategy from China, but the cost of keeping pace with computerisation at home. All this has come about because, over the last six months, there has been a revolution of expectations in the computer world now that is clear that falling prices will not be the same thing as cheaper operations for very much longer.

Design firms may still exult over the increasing number of plotters, monitors, scanners and knickknacks that can be bought or hired for less than they cost last year, but the real computer challenge is still to come. Peering over the purchasing horizon is the cheap supercomputer, the biggest onslaught of artificial intelligence on the architectural profession since the 1950s when firms first bought primitive Bendix G-15s to handle their accounts - and found out they could do away with their accounts departments as a result. The difference this time is that artificial intelligence has it in mind to convert computer aided design into computer dependent design - a trick it has long since pulled in fields as disparate as defence and high finance.

Like many other perturbations in today's technological universe, cheap supercomputing is part of the peace dividend. During the Cold War the United States government spent an estimated US$ 4 trillion on intelligence gathering, code-breaking and computer controlled weapon systems - far more money than any consortium of corporations, let alone professions, could have mustered for any civilian task. But when the Cold War combatants stood down, the knowledge embodied in defence activities lost its protected status overnight and it went civilian. In 1996, as part of this process, Cray Research, whose founder Seymour Cray had designed the Control Data 6600 and 7600 supercomputers to simulate nuclear explosions and crack enemy codes, was bought out by Silicon Graphics, a firm that announced its intention of turning the Cray computer, a 'bespoke' product tailored to the needs of individual clients, into the basis of a saleable range of supercomputers available to anyone.

Until then, supercomputing had been a slow-growing phenomenon. The outcome of years of the pursuit of computing power regardless of cost. Supercomputers not only performed astronomical numbers of calculations per second, but related different sets of data from dozens of different disciplines so as to produce an immediate synthesis that would be totally uneconomic, or too time consuming, to obtain in any other way. Supercomputers tested hypotheses about the nature of universe, designed supersonic aircraft, and carried out advanced studies in meteorology.

Understandably the market for the kind of work they could do was confined to government departments, defence establishments, universities and major industrial corporations. That is why, even today, there are so few supercomputers in the world. According to Oak Ridge Laboratories in the United States there are no more than 500 at present. But rare though they are, they are valuable too. Knowledge of the outcome of any project before it is implemented - whether it be the stretching of an airliner, or the verification of another milestone in the progress of global warming - is power indeed. The prospect of a free market and lower costs devolving this extraordinary synthesising power down to relatively small organisations, like those responsible for the design of buildings, is what is of burning interest now.

Architecture on its own is regarded as a small sector of the computer market that, until a year or two ago, thought six PCs in a LAN was the height of sophistication. Now it is about to confront the equivalent of the computational master of the universe -- a machine that can handle all its supposedly mysterious variables, all its complexities, all its aesthetics, and still have enough calculating power left to take care of structural design, building services, aerodynamics, specifications, costs, codes, claims, contract management and work out the payroll too. So who will stump up the millions to be the possessors of this synchronised knowledge?

It would be nice to say 'architects will', but it may not be true. Most architects are not only relatively impoverished, they make less coordinated use of their computing power than other profession in the construction industry. Structural, civil and environmental engineers spend millions on coordinated computer equipment. Several of them have their own software houses and all of them have better and more advanced computational capacity than firms of architects of equivalent size. In the same way large international construction firms invest more heavily in purchasing, contract and cost management computing than anybody else. Both these areas of expertise are compatible with each other and architects routinely rely on both. Very few architects have their own 3-D CADD -- with clash detection -- for production drawings. Most spend what money they have, not on computation but on visualisation. On achieving dramatic effects and arresting images. They rarely model entire projects in 3-D and rarely use the same software as other consultants.

So who will buy the new supercomputation capability that could turn any office into a total building design facility? In all probability not many architects. Instead, if they are not careful, they may find themselves confronted by construction firms and engineering consultancies looking for architectural services - for processing on their own supercomputers, of course.