Jason (jcreed) wrote,
Jason
jcreed

Got back into town. I realize that come September I am going to miss the fuck out of Pittsburgh, you guys. The warm fuzzy feeling of familiarity as the 61B passed the Forbes Ave. Arby's --- which is to Oakland as the Colossus is to Rhodes --- was unmistakable.

I taught myself the rudiments of Coq on the train back. Addition of natural numbers is commutative, in case you were unsure. I have to say, I do see why people like its interactive proof-building mode. I am not sure how I feel about tactics, tacticals, etc. It makes for a lot of different languages to learn, most of which don't feel very principled. Obviously the advantage is that sometimes your proof just gets proved faster, but eh. It feels similar to arguments about laziness in programming languages --- how it can make your program run faster, but can make for less predictable performance.

There are some things that confuse me a bit about what the difference between Set and Prop really is, and the difference between parametrized induction type families and inductive definitions of things arrowing into types, the latter which seemed reminiscent of a similar potentially confusing point in Agda. But mostly it seemed like an ok system except for the glaring lack of direct Twelfish support of HOAS. Some day I'll have to figure out how I feel about the whole Washburn/Weirich/Chlipala/Atkey world of PHOAS.
Tags: coq, travel
Subscribe
  • Post a new comment

    Error

    Anonymous comments are disabled in this journal

    default userpic

    Your reply will be screened

    Your IP address will be recorded 

  • 9 comments