Usability starts here …
I think both Shumin Zhai and Henry Lieberman are right. The most playful aspect of their arguments, http://www.almaden.ibm.com/cs/people/zhai/papers/EvaluationDemocracy.htm and http://web.media.mit.edu/~lieber/Misc/Tyranny-Evaluation.html, respectively, is that, though they ostensibly take different sides, they are actually describing the same intent, from different vantages. Shumin argues that HCI researches need to reconcile intuition with reality, using evaluation, to produce better (usable) results; Henry argues that quantifying an interface, or evaluating it against a narrow set of controllable criteria, as many evaluatistas do, avoids addressing the truly important aspects of design, and thus produces shallow results.
Do HCI researchers have too many ideas? Too little? What is really interesting?
When I read Henry’s prediction of a coming implosion of the HCI community (which is driven by producing quick results that fit with what most other researchers hold important), I half believe him. Research projects might be be better oriented if their metric for success was solving human problems, not just advancing the politics of the day.