While the language used makes understanding some of the examples difficult during the first time reading the article, the use of identical variable names (b = bottom = bowl) is utterly confusing. Another example of this is "...relate height change to height difference, but there is no way to achieve this directly." While this makes sense with careful study of the rest of the example description, it is not helpful when the bulk of the material presented is new to the reader. One more before we move on, "...substituting equals for equals..."

      Some of the details of this system were surely missed by me, but I did find the path tracing between physical structure, existing interactions and potential interactions quite clever. The graphical representation of this used in the article was very clear and gave an excellent summary of the main concepts involved in interaction-based design.

      I enjoyed the discussion notes at the end of this article. It was very interesting to see other people's questions and the author's answers. It did, however, use up some of the material I intended for this critique, but such is life. There are worse fates than having your questions answered. It did present another question though. I am slightly unsure of what was meant by "performance of the system" in the discussion. It is vague enough to be referring to the speed at which the system runs or the quality of the results produced.