src/HOL/ex/Lagrange.thy
 author hoelzl Fri Oct 24 15:07:51 2014 +0200 (2014-10-24) changeset 58776 95e58e04e534 parent 37885 c43805c80eb6 child 58889 5b7a9633cfa8 permissions -rw-r--r--
use NO_MATCH-simproc for distribution rules in field_simps, otherwise field_simps on '(a / (c + d)) * (e + f)' can be non-terminating
 paulson@11375 ` 1` ```(* Title: HOL/ex/Lagrange.thy ``` paulson@5078 ` 2` ``` Author: Tobias Nipkow ``` paulson@5078 ` 3` ``` Copyright 1996 TU Muenchen ``` paulson@5078 ` 4` ```*) ``` paulson@5078 ` 5` wenzelm@17388 ` 6` ```header {* A lemma for Lagrange's theorem *} ``` wenzelm@17388 ` 7` haftmann@16417 ` 8` ```theory Lagrange imports Main begin ``` nipkow@14603 ` 9` wenzelm@17388 ` 10` ```text {* This theory only contains a single theorem, which is a lemma ``` wenzelm@17388 ` 11` ```in Lagrange's proof that every natural number is the sum of 4 squares. ``` wenzelm@17388 ` 12` ```Its sole purpose is to demonstrate ordered rewriting for commutative ``` wenzelm@17388 ` 13` ```rings. ``` wenzelm@17388 ` 14` wenzelm@17388 ` 15` ```The enterprising reader might consider proving all of Lagrange's ``` wenzelm@17388 ` 16` ```theorem. *} ``` wenzelm@17388 ` 17` nipkow@23477 ` 18` ```definition sq :: "'a::times => 'a" where "sq x == x*x" ``` paulson@5078 ` 19` wenzelm@17388 ` 20` ```text {* The following lemma essentially shows that every natural ``` wenzelm@17388 ` 21` ```number is the sum of four squares, provided all prime numbers are. ``` wenzelm@17388 ` 22` ```However, this is an abstract theorem about commutative rings. It has, ``` wenzelm@17388 ` 23` ```a priori, nothing to do with nat. *} ``` nipkow@14603 ` 24` nipkow@23477 ` 25` ```lemma Lagrange_lemma: fixes x1 :: "'a::comm_ring" shows ``` wenzelm@20807 ` 26` ``` "(sq x1 + sq x2 + sq x3 + sq x4) * (sq y1 + sq y2 + sq y3 + sq y4) = ``` nipkow@23477 ` 27` ``` sq (x1*y1 - x2*y2 - x3*y3 - x4*y4) + ``` nipkow@23477 ` 28` ``` sq (x1*y2 + x2*y1 + x3*y4 - x4*y3) + ``` nipkow@23477 ` 29` ``` sq (x1*y3 - x2*y4 + x3*y1 + x4*y2) + ``` nipkow@23477 ` 30` ``` sq (x1*y4 + x2*y3 - x3*y2 + x4*y1)" ``` hoelzl@58776 ` 31` ```by (simp only: sq_def algebra_simps) ``` nipkow@14603 ` 32` nipkow@14603 ` 33` wenzelm@25475 ` 34` ```text {* A challenge by John Harrison. Takes about 12s on a 1.6GHz machine. *} ``` nipkow@14603 ` 35` nipkow@23477 ` 36` ```lemma fixes p1 :: "'a::comm_ring" shows ``` wenzelm@20807 ` 37` ``` "(sq p1 + sq q1 + sq r1 + sq s1 + sq t1 + sq u1 + sq v1 + sq w1) * ``` wenzelm@20807 ` 38` ``` (sq p2 + sq q2 + sq r2 + sq s2 + sq t2 + sq u2 + sq v2 + sq w2) ``` wenzelm@20807 ` 39` ``` = sq (p1*p2 - q1*q2 - r1*r2 - s1*s2 - t1*t2 - u1*u2 - v1*v2 - w1*w2) + ``` wenzelm@20807 ` 40` ``` sq (p1*q2 + q1*p2 + r1*s2 - s1*r2 + t1*u2 - u1*t2 - v1*w2 + w1*v2) + ``` wenzelm@20807 ` 41` ``` sq (p1*r2 - q1*s2 + r1*p2 + s1*q2 + t1*v2 + u1*w2 - v1*t2 - w1*u2) + ``` wenzelm@20807 ` 42` ``` sq (p1*s2 + q1*r2 - r1*q2 + s1*p2 + t1*w2 - u1*v2 + v1*u2 - w1*t2) + ``` wenzelm@20807 ` 43` ``` sq (p1*t2 - q1*u2 - r1*v2 - s1*w2 + t1*p2 + u1*q2 + v1*r2 + w1*s2) + ``` wenzelm@20807 ` 44` ``` sq (p1*u2 + q1*t2 - r1*w2 + s1*v2 - t1*q2 + u1*p2 - v1*s2 + w1*r2) + ``` wenzelm@20807 ` 45` ``` sq (p1*v2 + q1*w2 + r1*t2 - s1*u2 - t1*r2 + u1*s2 + v1*p2 - w1*q2) + ``` wenzelm@20807 ` 46` ``` sq (p1*w2 - q1*v2 + r1*u2 + s1*t2 - t1*s2 - u1*r2 + v1*q2 + w1*p2)" ``` hoelzl@58776 ` 47` ```by (simp only: sq_def algebra_simps) ``` nipkow@14603 ` 48` paulson@5078 ` 49` ```end ```