Peter Boettke at Coordination Problem also comments on the Journal of Economic Perspectives symposium on econometrics. These thoughts are especially good, so I wanted to highlight them in an entirely new post.
Generally, I think Boettke significantly understates the importance of mathematical models and empirical work, and he significantly overstates the dependability of logical deduction and praxeology. He overcompensates and therefore makes his own mistakes, in other words, but he's on the look out for the right things.
This was especially interesting:
"I have tried to tell the narrative of 20th century economics in a variety of ways --- the metaphor of an hour glass (which is derivative from the story that David Warsh tells) is one of my favorites, but the transformation in the classical period of a discipline that sought to derive universal propositions through logical deductions in natural language and empirical observations to a highly formal discipline with the goal of still deriving universal propositions to the emergence in the 1990s of a highly technical discipline that abandoned the search for universal propositions and instead resulted in a form of "formalistic historicism"."
I like his term "formalistic historicism", especially. The old Historical School is an approach that I think a lot of us would like to guard against, regardless of other differences we might have.