Learning curves

Why is it so hard to make buildings with curves? I'm talking about serious curves, here. That is: curves in multiple axes, complex curves and non-linear surfaces. It isn't as if there is anything revolutionary about curvy buildings; organic forms had a golden age in architecture from the 1950s through to the late 1960s. Classics include the Sydney Opera House, engineered by Arup, and the TWA terminal at JFK airport by Eero Saarinen. But times changed, and a more right-angled architecture dominated subsequent decades. Until recently.

There is a renewed enthusiasm for curved designs, partly because of the freedom offered by CAD systems. This year's degree show at the Bartlett School of Architecture in London was a case in point: complex curved designs outnumbered flat and angled designs by about 10 to one. From memory, it was half-and-half last year. As a top international school, its graduates will inevitably be running serious projects over the next decade or so.

So curves are in. The famous are now all at it: Frank Gehry, Richard Rogers, Norman Foster, Will Alsop, Zaha Hadid, Future Systems … But it has become apparent that the cost is often prohibitive. The process of making complex curved cladding or casting formwork for building structures is time consuming and labour intensive – and construction costs in the past decade have risen well ahead of inflation. The result is that curves are now mostly out of the reach of buildings without sky-high budgets.

Now it may be that the problem will be fixed by fashion, and the shapes will get simpler again. But I don't think so this time. The recent history of architecture has some distinct trends when it comes to design: the best of the pioneering work invariably resurfaces a few decades later. After a period of evolution, the good ideas are taken up by a new generation, reinterpreted and integrated into their lexicon. This is true with both technology and forms. We see this pattern in areas from high-rise living through to heat reclamation, from fibreglass cladding to the once dodgy fibre-reinforced cement. Often, the prototype designs had technical problems, and the second coming is catalysed by advances in engineering and materials.

But if curvy designs are struggling against economics this time round, then my natural instinct is to look at construction. If we can't make labour cheaper and quicker, then the curves must be made another way. Other industries are my first port of call for pinching good ideas. I notice that consumer products – boat hulls, cars and so on – are pretty well made and certainly curvy. But on closer inspection, these are all mass produced and repeated shapes, based on hugely expensive tools and formwork – apart from monocoque boat hulls, which are probably not a cheap way to go.

This year’s degree show at the Bartlett was a case in point: complex curved designs outnumbered flat designs by 10 to one

More interesting is the notion of self-forming surfaces. These don't need the formwork; they droop or get pulled into shape. Catenary curves form naturally, although loading is not always correct, as Gaudí discovered. Tension fabric and air-inflated solutions are well established, but I see these as niche applications because they dictate the building form.

So, we're stuck.

Well … maybe not. A blindingly simple idea has come along to give curves a chance. Working initially with DLM Architects (a young practice in London), a bright engineer has been studying the possibility of using composite stretchy materials that can be "set" to create structural shells. It may be a year or two before the engineer, Richard Harrison, has completed his development and launched a product with all the certification that is needed. But I hope it works. In my experience, simple ideas usually have a higher chance of success.