Graduate mathematics, and what Lean has of it

Sixteen standard graduate mathematics textbooks — Hartshorne, Weibel, Lee, Neukirch, Serre, Matsumura and others — read cover to cover. Every definition, theorem, lemma, example and remark in them is written down with its statement and the page it appears on: 2,701 definitions, 2,706 theorems and propositions, 1,514 lemmas and corollaries, and 2,659 examples, constructions, remarks and conventions besides.

For 6,959 of those statements someone has then gone looking for the same statement in Lean, and recorded what they found:

Already in Mathlib 1,504
In some other Lean development 114
In this repository 72
Nowhere in Lean, as far as a documented search could tell 5,269

That last row is what the site is for. Each of those entries names the pieces Lean does have and the gap that stops them adding up to the statement, so it reads as a description of work to be done rather than a blank.

Open a chapter and you get its mathematics in order, each statement with its location in the book and, where it has been checked, its standing in Lean. Every declaration name links to its Mathlib documentation; every commit and pull request links to the code.

Where to start

What the statuses mean

Status Statements Share What it means
not formalized 5269 76% A documented search found no Lean statement of it anywhere.
in Mathlib 1504 22% Mathlib states and proves it.
in another Lean project 110 2% Some other Lean development has it; using it means porting that code.
in lean-categories 72 1% This repository states and proves it.
in a Lean package 4 0% A library available as a Lake dependency has it.
checked in total 6959

What this is not

Checking is a search, not a proof. “Nowhere in Lean” means a documented search of Mathlib at a named commit, of open pull requests, and of the Lean developments its entry names, turned nothing up — not that nothing exists. Mathlib moves, so an entry is only as current as the commits its chapter records.

The statements here are compressed restatements, not quotations, and they carry the compressor’s mistakes. Where one of them and the Lean disagree, the Lean is right; where one of them and the book disagree, the book is right.