While this article was extremely interesting, I found it cluttered with the various lists, things that are recognized, categories of things, etc. I just started to not care and would skip over these lists. Sometimes they were useful for understanding a certain section or point but most often not.

      I found the idea of physics simulator that did not deal with numeric, physical characteristic parameters very interesting. Usually when one thinks of simulators, very complicated and exact representations and descriptions are used as the basis for manipulating whatever is being simulated. The one used in EDISON seems to be more general, dealing with relations rather than exact values for size, motion, etc. This is very consistent with the näive approach to design. While it may not be concerned with exact values for length, weight and other such physical characteristics, a substantial body of knowledge must have been needed to make it work. It might have been even a more difficult task than if it had used actual numeric values since with an exact representation, manipulation and other such things can be calculated rather than reasoned about. As we've seen in many systems, it is far easy to calculate than intelligently reason.

      I thought it was very clever to use the observation that people can often figure out what an object is for and how to use it with only a basic understanding of the physical world and turn it into a paradigm for design. The description of the the EDISON system does illustrate well how complicated and knowledge rich even a näive representation can be.