Mies's constraint of matter by ideal geometry is based on an essentialist notion: that matter is formless and geometry regulates it, that geometry is transcendetal and in some sense indifferent to the material that substantiates it.
When freed from such essentializing conceptions, matter proves to have its own capacity for self-organizations. As an analog computer, it can perform optimizing computations that have been shown to be trans-scalar; as for example, when stretched stockings are used to calculate the geometry and form of full-scale tensile structures.
This logic, however, can be expanded to more complex situations, as, for example, when a magnetized ferrofluid calculates in a dynamic balance the forces of magnetism, gravity, and surface tension. In this state, matter is much more dynamic and pregnant with the possibility of its own constraints and leniencies than any projected geometry or simple optimizing principle could render. It becomes a model not only for dealing with structure but for dealing with the feedback that occurs between multiple forces at work on a building, encompassing program, use, organiztion, and form.
（上网查了一下Analog Computer的定义：An analog computer is a form of computer that uses continuous physical phenomena such as electrical, mechanical, or hydraulic quantities to model the problem being solved. - From Wikipedia)
这段话的重点是在物质自身的"self-organization"。物质内在的这种逻辑和自我组织形成的模型不是外部的简单的几何形式可以描述的，这种模型可以用来对应在结构/功用/组织/形式之间的多种张力和反馈的。Analog Computer和Magnetized Ferrofluid (铁磁流体) 的比喻当然很精彩很后结构(post-structuralism)，但是把这些和建筑扯在一起稍微觉得有些远 - 建筑毕竟不是高级物理。而且总觉得后结构主义的一些追随者从"数字化"，"流体"，"复杂物理"来生成建筑形式的做法缺少了一些诗意。
铁磁流体毕竟不是大家经常能见到的东西。后结构也可以不必是形式上的模拟。觉得把在不同层面 - 结构/力，功能/空间，形式/几何 - 对建筑的思考和操作结合起来会更后结构一点。看老库的一些建筑，不就是在结构和空间上都玩上一把，就自然形成了很有张力很丰富的形式？比起密斯对几何的专注，康对材料和空间的偏爱，赖斯对流动空间/有机建筑的执著，这样好像也会更好玩一些更加responsive一点。
