News for hackers. Stuff that matters.
Re:Kind: A Modern Proof Language by c-cube on Wednesday August 11, 2021 @01:43PM (#28145174)
Not all proof assistants are based on dependent type theory. The HOL family is based on simple type theory + a classical axioms: you get choice, excluded middle, extensionality etc. in exchange for a less powerful type system. Maybe not coincidentally, Isabelle/HOL also has arguably the most powerful automation thanks to its ability to call automatic provers, almost all of which are based on non-dependent classical logic.

Other non-dependent type theory systems are Mizar, the B method, TLAps…

Re:Kind: A Modern Proof Language by sadfev on Wednesday August 11, 2021 @02:32PM (#28145841)
Very true. But the most popular ones are based on DTT, GP only mentioned Lean/Coq/Idris.

Indeed by sacrificing the power of type system we can augment classical axioms consistently as shown by HOL family.

These come with their own disadvantages.