The book taught act one (getting a kernel to accept a proof) across twelve
chapters and never taught act two: the protocol that makes a green light mean
something to someone who was not present. ch11 armed the reader to interrogate a
CERTIFICATE and left the script that interrogates certificates — software the
author wrote, verified by nothing — entirely unexamined. That was the gap eight
rounds of external review found in the companion projects, and it is a didactic
failure of this book too.
ch13 teaches it from the war stories: the two acts and why act one is the easy
one (eleven theorems in two days, never disputed; eight review rounds and
eighteen defect classes to make the button over them credible); the single shape
every failure had (something load-bearing sat outside the binding), with the
demonstrated exploits including ALL GREEN over a repository proving False, a
certificate reduced to 'the loop equals the loop' with every fingerprint
byte-identical, and a stubbed compiler wrapper going green in 3.6 seconds over
destroyed proofs; completeness of binding and its four rules (derive the
population, fail closed on absence, exact not subset, a stranger must re-derive);
and the meta-defect of assertions that pass for the wrong reason, including the
tautological assert that appeared twice — the second time inside its own repair.
Closes with the habit to carry: ask both questions, and invite someone to attack
your button early, because none of the eighteen was found by the author.
ch11 now forward-references it at the point where the old blind spot sat.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- pen-and-paper worked examples in all 12 chapters, using the REAL
constants throughout: 2^-64 waiting-time arithmetic, headroom budgets,
hand type-checking, rfl traces, full goal-state boards, the column-sum
audit at 2^54, inverting 19 mod p via Euclid, the x19 fold at real
weights, denoting p itself (telescope), the 16p audit (8 fails by 151),
the 254+11 inversion-chain bookkeeping, the substitution test, sizing
the 28-vs-1000 extraction, cofactor/torsion arithmetic, and the full
Bernstein-Lange completeness derivation
- CORRECTNESS FIX: ch7 asserted a false factorization of p-1; replaced
with the computationally verified p-1 = 2^2 * 3 * 65147 * Q (Q 71-digit
prime), witness w=2 verified for all four Pratt conditions
- every chapter's exercises now followed immediately by 'Solutions and
pathways' (pathway first, then answer), incl. new exercises
- NEW Interlude: a complete two-clause verification done entirely by
hand, then mapped line-by-line onto the compiled Lean proof
- NEW appendices: A pen-and-paper toolkit (8 recipe cards + drills +
answers), B guided walkthroughs of every exercise-file hole, C tour of
the real repositories; plus glossary, instructor notes, 13-week plan
- preamble: worked-example box, solution macros, math-safe inline code
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- 53-page LaTeX/TikZ book (main.pdf + full sources): from zero background
to reading the real Ed25519/Pasta verification projects
- runnable exercises with sorry-holes + complete solutions for chapters
2-7, 9, 12; every solution file compiles clean (zero errors, no sorry)
against Lean v4.30.0-rc2 + Mathlib 5450b53e
- lake project pinned to the same toolchain/Mathlib the solutions were
verified with; students fetch the Mathlib cache, never build it
- honesty ledger in README: what was machine-checked and how
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
