Intelligent readers often complain that they don’t get what I mean by “meta-rational,” and want examples.
Here are some nice ones from @JohnDCook's blog:
in my CS courses I learned a lot about algorithms and not a whole lot about how to make large, complex, multi-layered software/hardware systems work let alone how to structure and maintain them and collaborate with others to do this work.
-
-
It’s hard to teach that stuff, partly because it is partly meta-rational, and we don’t yet have good methods for talking about (or teaching) meta-rationality
-
a sympathetic grad student told me early on that the best project collaborators are the ones that reduce the SLOC count and don't check in code without tests written for it that advice was worth more than the rest of the curriculum combined
End of conversation
New conversation -
-
-
we were taught a crash course in a development methodology that was considered a critique of a broken process when it was first published! (waterfall) then we were told to summarily ignore the contents of the course and instead we would be graded on a short written response test
-
Because university is forced to function as a credentialing machine, it mostly can only teach material that is easy to put into a standardized test
-
well, that's what was an interesting exception in this case. they actually threw out the standardized test because it was so useless. the short response written test they gave was idiosyncratic to the department.
-
Ah, I see—I misread your previous tweet!
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.
“Bring out your differential equations!” is a failure of naive rationalism.
“A feedback loop of provisional problem formulation, attempted solution, revised formulation” is meta-rational.
Intro stats books create the rationalist misunderstanding that science gets results by pushing data through a formula.
Meta-rationality requires asking what your data *mean*; and only then asking which statistical methods are relevant and why.
The CS curriculum teaches methods for solving small, well-formulated hard problems: the essence of rationality.
Mostly irrelevant to software engineering practice, which is about managing vast, amorphous messes: a major theme in meta-rationality.