User Profile — FixpointFeline
About FixpointFeline
"Both fixpoints are valid."

Lattice theory, abstract interpretation, model checking. Will defend Tarski at all costs. Proof assistant enthusiast. I spend most of my time chasing fixed points through complete lattices and shouting at people who think μ is the only one that matters (ν-types are people too). Ask me anything about abstract interpretation, CTL/LTL model checking, or how to torment undergrads with inductive/coinductive definitions.

Currently: formalising Knaster–Tarski in Lean 4. It's going fine. [citation needed]
Pinned Theorem
Knaster–Tarski Fixed-Point Theorem (Tarski 1955)
Let (L, ≤) be a complete lattice and f : L → L be an order-preserving (monotone) function. Then the set of fixed points of f in L forms a complete lattice under . In particular, f has both a least fixed point μ = ⊓Post(f) and a greatest fixed point ν = ⊔Pre(f).

Tarski, A. "A Lattice-Theoretical Fixpoint Theorem and Its Applications." Pacific J. Math. 5 (1955), 285–309.
Research Interests
Tarski's Theorem Knaster–Tarski Abstract Interpretation Lattice Theory Model Checking μ-Calculus ν-Calculus Coinduction Proof Assistants Lean 4 Coq Fixpoint Induction CTL / LTL Complete Lattices Domain Theory Denotational Semantics Static Analysis
Reputation +2,105
Top 0.3% of all members by reputation · Consistently rated as one of the most helpful contributors in Lattice Theory & Abstract Interp.
Awards & Badges
🏆 Top Contributor
📐 Proof Master
🔁 Fixpoint Fan
🌐 Lattice Lord
🐈 Cat Girl Core
✍️ 3k+ Posts
🧠 Mentor
5-Year Vet
Recent Posts
Signature
— FixpointFeline's signature —
μX. F(X)  =  lfp(F)  =  ⊓{ x | F(x) ⊑ x }
νX. G(X)  =  gfp(G)  =  ⊔{ x | x ⊑ G(x) }
Both are valid. Both are beautiful. Don't @ me.

Proof assistants used: Lean 4 · Coq · Isabelle/HOL
« Back to Member List  ·  « Board Index  ·  All posts by FixpointFeline