Well, the author himself says, "Finally we are reaching something which is non-stupid and non-sexist enough that it might actually exist somewhere in reality"

Well, the author himself says, "Finally we are reaching something which is non-stupid and non-sexist enough that it might actually exist somewhere in reality" I think the male/female tables were meant to be a model of how people actually think, as is the whole of the article, not how any even incompetent data modeler would model the problem (except in the second half, where acceptable mental models and acceptable data models converge).