Re:Identity Types
by auggierose on Friday September 26, 2025 @10:31AM
(#45386932)
Cool project. I was already shocked that you use C++, and then I thought, 40% assembly language, no way. Turns out Github thinks that your theory files are assembly :-)
Yes, the type universe hierarchy is the canary in the coal mine that something is wrong with type theory as a foundation, but very few take it seriously. Some do though, see for example the first chapter of http://abstractionlogic.com .
I kind of like your idea of having trusted extensions to provers. I was myself thinking along the lines of taking the foundation that I got in my Calculus of Constructions project and see if it can be put on a smaller set theory foundation. I.e., prove it in ZFC, presumably with inaccessible cardinals added.