I've gotten an opportunity to combine my interest in computing with my interest in building design. And, lacking anyone to talk about it with, I've decided to start a series on computing and building design here. I'm going to begin with a discussion of current work, and move on to both software specfics and broader issues.
Some background first: the University of Oregon's Environmental Control Systems series covers HVAC (heating, ventilation, air-conditioning), lighting, and acoustics. As part of the two-course series we are expected to do two Vital Signs studies as term projects. For my first study, I did something about air-flow in a stairwell. For the second, I am now doing a project involving daylight and artificial light in a fine old library reading room; you can see some of the preliminary results here. Part of the project involves predicting lighting.
Now the usual means of doing this is to build a physical scale model and photograph it in actual daylight. But this cannot predict the combination of artificial light and daylight with any reliability; it is very difficult to build accurate working scale models of luminaires (lamps + fixtures + controls) and even more difficult to match the light that a luminaire actually puts out.
An ideal situation for computer modelling, not so? And when, back in the '70s, daylighting experienced a revival, a research effort was begun which bore fruit in a program called Radiance. Now Radiance is a strange and wonderful beast. It is an extra-ordinarily good lighting modeller; it combines radiosity and ray-tracing approaches to the prediction of radiance (which is the technical name for electromagnetic radiation coming from a point) in the visible spectrum and produces models accurate enough to use for lighting design and handsome enough to use architectural presentations.
However, Radiance is exceptionally difficult to use. I would guess there are at most 500 regular users of the system; perhaps as few as 100. These users, however, include some of the foremost engineering firms and lighting designers in the world. The difficulties with Radiance lie in two areas: compatibility with existing building modelling software and user interface.
And I think I'm going to break this off here...I will eventually get to such fun things as setting up Radiance on a Beowulf cluster and the need for an open framework for computer architectural modelling.