Note: I am working on a rewrite at
josephmckinsey/flean2.
Flean is an attempt at formalizing floating point numbers in Lean 4.
Personally: Learn Lean in a serious but not hard way. I've already learned a lot.
In general, the idea is to (1) model the floating point numbers, (2) characterize their properties via a convenient set of rules,
and (3) apply that logic to prove things about Lean's Float by declaring an interface axiomatically.
- Be capable of matching operations of normal floating points as exactly as possible IEEE 754
- Extensible to different precisions
- Theoretically capable of supporting lots of bounds on Lean's
Float
