Thanks for initiating this interesting discussion, Janet. We had considered making Diversity the theme of this year's SIGCIS workshop, and maybe we dropped the idea too soon.
I agree that McPherson's article is not fully developed, but then again, it's not a journal article. She conspicuously labels sections fragments and is writing to explore. It's heartening to hear that several people are using this as a conversation starter in their classrooms.
McPherson cites Winner and other scholars who border on a deterministic analysis, and her writing walks that line too. However, we could also fault her article for going the other way. Cultural determinism is as much a fallacy as technological determinism, and we could fault her question that way as well: just because Unix was developed in an era when racism was keenly felt, does Unix have to somehow bear the mark of that culture?
However, the keenly expressed antagonism to this question seems disproportionate. As a profession, we're willing to say that personal computing carries traces of cold war culture (thanks to Edwards and others), or that the economic and political system of the USSR made it unlikely that a distributed communication system would develop (thanks to Gerovitch). It's clear that we don't have the full picture of how the interaction between the predominantly white, male culture at the time of Unix's development would result in particular design decisions. At the very least, though, we could assert that the particular kind of time-sharing we got worked well with a community that was lacking in diversity.
Asking these kinds of questions is important because we are still dealing with a community that is finding it difficult to diversify. The medical profession, which also was largely white and male in the 1960s, made it a professional imperative to diversify with the result that medical school admissions are much different than they were 50 years ago. Engineering and computer science have not been successful in that regard, despite noble attempts in various corners. SIGCIS and other groups could make an effort to explore the reasons for this failure.
Chris Leslie