Last week, I had the good fortune of representing the SPM at the 2010 Advances in Architectural Geometry (AAG) Conference in Vienna.
Following the successful 2008 conference, the 2010 AAG was significantly larger and broader in scope. With an opening talk by Hugh Whitehead (Foster and Partners), a closing talk by David Rutten (McNeel, Grasshopper), and many extremely fascinating talks in between, the conference tackled issues of complex geometry, robotics, acoustic optimization, and genetic algorithms. Even the numerous coffee breaks were packed with interesting conversations on a plethora of issues associated to the AEC industry.
Generally speaking, I found this conference to be mainly about implementation and software development – which in a world of value engineering, cost cutting, and budget meetings is good, since tangible objects are much easier to price and evaluate than intangible speculation. Many of the top buzz words of the AEC industry (dynamic relaxation, genetic algorithms, sustainable optimization, planarization, etc…) had all been implemented in a variety of ways, on variety of platforms, and, it should be said, to varying degrees of success. It would seem that all that is left is to navigate the vast ocean of plugins and scripts.
I think that these developments are good for the industry in that they provide a much more rational framework for design exploration – a freeform surface paneled with planar quadrilateral panels is cheaper that one with curved panels (see Evolute’s interesting new plugin). At the same time, though, it’s misleading to assume that any one of these plugins tell anything close to the whole story.
For instance, a paper was presented wherein a planar quadrilateral version of the famous British Musuem roof was presented (which is interesting in and of itself) and then modified to be “influenced” by the principal stress lines computed from the roof’s dead load. Under these circumstance, the authors of the paper managed to reduce the deflection of the roof by 8 percent. It turns out, though, that the authors didn’t compare the performance of the original (built) design to the solution they proposed in their paper.
This sort of approach is likely to produce, to use the parlance of our times, locally optimal solutions rather than globally optimal ones. Although influencing panelization schemes along stress lines and other such investigations are extremely interesting, they should constantly be checked by real world situations and considerations.
But typically, real innovation happens when we think outside of what believe are our boundaries – would adding too many “real world” constraints to the mix simply paralyze the burgeoning field of architectural geometry?
I think if we knew the answer to that, we would know a lot of other things, too.