David, as usual your message is thought provoking and skirts the limit of our understanding. Here, however, I am concerned about knowing what we don't know, to coin a phrase ;-)
I agree, and I see others who seem to concur, that human behavior is of legitimate concern in systems design, and therefore in protocol design. When we consider typical, individual human behavior it is often called Human Factors Design, or more generally user-friendly design. Consideration of human behavior in the aggregate, i.e. social issues, is different.
I have more than once been in a meeting in which someone who has a PhD in the subject as well as successful experience has had to say, "I know you know computer science, I know you know networking, and I know you think you know what's intuitive for the average person, but you know NOTHING about human factors." And she was absolutely right.
Conveniently at the protocol level it often doesn't matter too much. But it might matter in areas like Spam abatement and certificate verification and internationalized domain support where we have trained people to assume certain things, and we have limited real estate in which to deliver information.
I guess what I'm saying is that who develops the running code and who actually knows how the code should be developed to interact with humans can be very different, and while we at the IETF have the expertise in writing code, we probably have very little on the CHI side, and probably cannot, with the rare exception of the person who has transitioned careers.
That can present a problem at times, but the simplest of resolutions is to make sure that those who spot these issues are adequately reflecting the views of those who are experts in CHI, when it comes to design.
Eliot _______________________________________________ Ietf@xxxxxxxx https://www1.ietf.org/mailman/listinfo/ietf