Dear Victor,
You may be happy but still I am sorry that there is a mistke in the link which I share with you before on FLT : ttp://noticingnumbers.net/FLTsummary.htm[/URL]
I understand already from your web-site that your work on FLT is base on splitting the numbers a, b, c into pairs of sums, grouping them in two sums U' and U'' and multiplying the equation a^n + b^n – c^n = 0 by 11^n and then looking on the (k+3)-th digit from the right in the number a^n + b^n – c^n (where k is the number of zeroes at the end of the number a + b – c) . when you look on it on base n we can see that is not equal to 0 .
This is very nice [B]organic[/B] looking and I really like it !
But please notice to Ramsey 2897 note :
[QUOTE]It appears that Victor's current position is that a,b and c should be multiplied first by d to give u = ZZZZZZ00000 where Z=n-1, the number of Z's is "s" and the number of 0's is k. Secondly, multiply the new a,b,c by (n-1) to give u=YZZZZZ10000 where Y=n-2 the number of Z's is s-1 and the number of zeros remains k. He has presented this position piecemeal and his posts are only in outline form and contain typographical errors that are significant to an understanding of the proof. We have given plenty of time to respond to Victor in a meaningful way. He owes it to us (and especially to you given your willingness to verify his proof) to explain his present position in a clear manner in a single post which has been checked for errors and which can be followed without undue effort.[/QUOTE]
So, I understand , that you did not correct yet in your web-site of FLT
the mistaks that you admit in this thread.
Could you do that for me and I will continue to study your intersting work ?
Your sincerely
Moshe
[QUOTE]
The idea of the elementary proof of the Fermat’s Last Theorem (FLT) proposed to the reader, is extremely simple:
After splitting the numbers a, b, c into pairs of sums, then grouping them in two sums U' and U'' and multiplying the equation a^n + b^n – c^n = 0 by 11^n (i.e. 11 power n, and the numbers a, b, c by 11), the (k+3)-th digit from the right in the number a^n + b^n – c^n (where k is the number of zeroes at the end of the number a + b – c) is not equal to 0 (the numbers U' and U'' are multiplied in a different way!). In order to understand the proof, you only need to know the Newton Binomial, the simplest formulation of the Fermat’s Little Theorem (included here), the definition of the prime number, how to add numbers and multiplication of a two-digit number by 11. That’s ALL! The main (and the toughest) task is not to mess up with a dozen of numbers symbolized by letters. The formal account of the history of the Theorem and the Bibliography are not included in the Russian version
An elementary proof of Fermat’s Last Theorem
VICTOR SOROKINE
TOOLS: [Square brackets are used for additional explanation.]
Notations used:
All the numbers are written using a base n with n being a prime number and n > 10.
[All the cases where n is not prime, except n = 2k (which can be reduced to the case n = 4),
are reduced to a case with a prime n using a simple substitution.
We don’t study the cases with n = 3, 5 and 7.]
ak – the digit at the place k from the end, in the number a (thus a1 is the last digit).
[Example for a = 1043 in the base 5: 1043 = 1x53 + 0x52 + 4x51 + 3x50; a1 = 3, a2 = 4, a3 = 0, a4 = 1.]
a(k) – is the k digits’ ending (it is a number) of the number a (a(1) = a1; 1043(3) = 043).
Everywhere in the text a1 ≠ 0.
[If all three numbers a, b and c end by a zero, we need to divide the equation 1° by nn.]
(ain)1 = ai and (ain - 1)1 = 1 (cf. Fermat’s Little Theorem for ai ≠ 0). (0.1°)
(n + 1)n = (10 + 1)n = 11n = …101 (cf. Newton Binomial for a prime n).
A simple corollary from the Newton Binomial and the Fermat’s Little Theorem for s ≠ 1 [a1 ≠ 0]:
If the digit as undergoes a rise or is reduced by d (0 < d < n),
then the digit ans+1 undergoes a rise or is reduced by d (or d + n, or d – n). (0.2°)
[Digits in negative numbers are « negative ».]
***
(1°) Let us assume that an + bn – cn = 0 .
Case 1: (bc)1 ≠ 0.
(2°) Let u = a + b – c, where u(k) = 0, uk+1 ≠ 0, k > 0 [we know that both u > 0 and k > 0 in 1°].
(3°) We multiply the equality 1° by a number d1n (cf. §§2 and 2a in the Appendix) in order to transform
the digit uk+1 into 5. After that operation the numbers’ labeling is not changed
and the equality still keeps its index (1°).
It is clear that also in the new equality (1°) u = a + b – c, u(k) = 0, uk+1 = 5.
(1*°) Then let a*n + b*n – c*n = 0, where the sign “*” designates numbers of the equation (1°) written in a canonical way, after the multiplication of the equation (1°) by 11n .
(4°) Let’s introduce following numbers, in that order: u, u' = a(k) + b(k) – c(k),
u'' = u – u' = (a – a(k)) + (b – b(k)) – (c – c(k)), v = (ak+2 + bk+2 – ck+2)1, u*' = a*(k) + b*(k) – c*(k),
u*'' = u* – u*' = (a* – a*(k)) + (b* – b*(k)) – (c* – c*(k)), 11u', 11u'', v* = (a*k+2 + b*k+2 – c*k+2)1.
Let’s then calculate the two last significant digits in these numbers:
(3a°) uk+1 = (u'k+1 + u''k+1)1 = 5;
(5°) u'k+1 = (–1, 0 or 1) – because – nk < a'(k) < nk, – nk < b'(k) < nk, – nk < c'(k) < nk
and the numbers a, b, c have different signs;
(6°) u''k+1 = (4, 5 or 6) (cf. 3a° and 5°) [it is important: 1 < u''k+1 < n – 1];
(7°) u'k+2 = 0 [always!] – because \u'\ < 2nk ;
(8°) u''k+2 = uk+2 [always!];
(9°) u''k+2 = [v + (ak+1 + bk+1 – ck+1)2]1, where (ak+1 + bk+1 – ck+1)2 = (–1, 0 or 1);
(10°) v = [uk+2 – (a(k+1) + b(k+1) – c(k+1))k+2]1 [where (a(k+1) + b(k+1) – c(k+1))k+2 = (–1, 0 or 1)] =
= [uk+2 – (–1, 0 or 1)]1;
(11°) u*k+1 = uk+1 = 5 – because u*k+1 and uk+1 are the last significant digits in the numbers u* and u;
(12°) u*'k+1 = u'k+1 – because u*'k+1 and u'k+1 are the last significant digits in the numbers u*' and u';
(13°) u*''k+1 = (u*k+1 – u*'k+1)1 = (3 – u*'k+1)1 = (4, 5 or 6) [it is important: 1 < u*''k+1 < n – 1];
(14°) (11u')k+2 = (u'k+2 + u'k+1)1 (then, reducing the numbers into a canonical form –
the value u'k+1 «goes» into u*''k+2, because u*'k+2 = 0);
(14a°) it is important: the numbers (11u')(k+2) and u*'(k+2) differ from each other
only by k+2-th digits, more exactly: u*'k+2 = 0, but (11u')k+2 ≠ 0 generally;
(15°) (11u'')k+2 = (u''k+2 + u''k+1)1;
(16°) u*k+2 = (uk+2 + uk+1)1 = (u''k+2 + uk+1)1 = (u''k+2 + 5)1;
(16а°) Let’s note that: u*'k+2 = 0 (cf. 7°);
(17°) u*''k+2 = (u*k+2 +1, u*k+2 or u*k+2 – 1)1 = (cf. 9°) = (u''k+2 + 4, u''k+2 + 5 or u''k+2 + 6)1;
(18°) v* = [u*k+2 – (a*(k+1) + b*(k+1) – c*(k+1))k+2]1
[where u*k+2 = (uk+2 + uk+1)1 (cf. 16°), but (a*(k+1) + b*(k+1) – c*(k+1))k+2 =
= (–1, 0 or 1) – cf. 10°] = [(uk+2 + uk+1)1 – (–1, 0 or 1)]1.
(19°) Let’s introduce the numbers
U' = (ak+1)n + (bk+1)n – (ck+1)n, U'' = (an + bn – cn) – U', U = U' + U'',
U*' = (a*k+1)n + (b*k+1)n – (c*k+1)n, U*'' = (a*n + b*n – c*n) – U*', U* = U*' + U*'';
(19а°) Let’s note that: U'(k+1) = U*'(k+1) = 0.
(20°) Lemma: U(k+2) = U'(k+2) = U''(k+2) = U*(k+2) = U*'(k+2) = U*''(k+2) = 0 [always!].
Indeed, from 1° we can find the following:
U = an + bn – cn =
= (a(k+1) + nk+1ak+2 + nk+2Pa)n + (b(k+1) + nk+1bk+2 + nk+2Pb)n – (c(k+1) + nk+1ck+2 + nk+2Pc)n =
= (a(k+1)n + b(k+1)n – c(k+1)n) + nk+2(ak+2a(k+1)n - 1 + bk+2b(k+1)n - 1 – ck+2c(k+1)n - 1) + nk+3P =
= U' + U'' = 0, where
U' = a(k+1)n + b(k+1)n – c(k+1)n,
(20a°) U'' = nk+2(ak+2a(k+1)n -1 + bk+2b(k+1)n -1 – ck+2c(k+1)n -1) + nk+3P,
where (ak+2a(k+1)n -1 + bk+2b(k+1)n -1 – ck+2c(k+1)n -1)1 = (cf. 0.1°)=
(20b°) = (ak+2 + bk+2 – ck+2)1 = U''k+3 = v (cf. 4°).
(21°) Corollary: (U'k+3 + U''k+3)1 = (U*'k+3 + U*''k+3)1 = 0.
(22°) Let’s calculate the digit (11nU')k+3:
[as the numbers (11u')(k+2) and u*'(k+2) differ only by k+2 digits, with the difference being equal
to (11u')k+2), then this will also be the difference between the digits (11nU')k+3 and U*'k+3, which
means that the digit (11nU')k+3 will be greater than the digit U*'k+3 by (11u')k+2 (cf. 0.2°)]
(11nU')k+3 = U'k+3 = (U*'k+3 + (11u')k+2)1 = (U*'k+3 + u'k+1)1.
(23°) From there we come to: U*'k+3 = U' k+3 – u'k+1.
(24°) Let’s calculate the digit U*'' k+3 :
U*'' k+3 = v* = (uk+2 + uk+1)1 – (–1, 0 or 1) – cf. (18°);
(25°) Finally, let’s calculate the digit (U*'k+3 + U*''k+3)1:
(U*'k+3 + U*''k+3)1 = (U*'k+3 + U*''k+3 – U'k+3 – U''k+3)1 = (U*'k+3 – U'k+3 + U*''k+3 – U''k+3)1 =
(cf. 23° and 24°) = (– u'k+1 + v* – v) = (cf. 18° and 10°) =
= (– u'k+1 + [uk+2 + uk+1 – (–1, 0 or 1)] – [uk+2 – (–1, 0 or 1)])1 =
= (– u'k+1 + uk+1 + (–2, –1, 0, 1, or 2))1 = (cf. 3a°) =
( u''k+1 + (–2, –1, 0, 1, or 2))1 = (cf. 6°) = (2, 3, 4, 5, 6, 7 or 8) ≠ 0,
which contradicts to 21° ; therefore the expression 1° is an inequality.
Case 2 [is proven in a similar way, however it is much more simple]: b (or c) = ntb',
where b1 = 0 and bt+1 = b'1 ≠ 0.
(26°) Let’s introduce number u = c – a > 0, where u(nt – 1) = 0, but unt ≠ 0 (cf. §1 in the Addendum).
(27°) After multiplying the equality 1° by a number d1n (in order to transform the digit unt into 5)
(cf. §§2 and 2a in the Addendum), we will keep the numbers’ notations.
(28°) Let’s: u' = a(nt – 1) – c(nt – 1), u'' = (a – a(nt – 1)) – (c – c(nt – 1)) (where, evidently, u''nt = (ant – cnt)1);
U' = a(nt)n + bn – c(nt)n (where U'(nt + 1) = 0 – cf. 1° and 26°), U'' = (an – a(nt)n) – (cn – c(nt)n),
U*' = a*(nt)n + b*n – c*(nt)n (where U*'(nt + 1) = 0), U*'' = (a*n – a*(nt)n) – (c*n – c*(nt)n),
v = ant+1 – cnt+1.
The calculations, in all ways analogical to the case 1, show that the digit at the nt+2-th place in the Fermat’s equality is not zero. The number b in all calculations (except the very last operation and also in paragraph 27°) can be ignored, because the digits bnnt+1 and bnnt+2 , after multiplying the equality 1° by 11n, won’t change (this is because 11n(3) = 101).
Therefore, for any prime number n > 10, the theorem is proven.
==================
ADDENDUM
§1. If the numbers a, b, c don’t have common factors and b1 = (c – a)1 = 0,
then, from the number R = (cn – an)/(c – a) =
= cn –1 + cn –2a + cn –3a2 + … c2an - 3 + can - 2 + an - 1 =
= (cn –1 + an –1) + ca(cn –3 + an –3) + … + c(n –1)/2a(n –1)/2 =
= (cn –1 – 2c(n –1)/2a(n –1)/2 + an –1 + 2c(n –1)/2a(n –1)/2) + ca(cn –3 – 2c(n –3)/2a(n –3)/2 + an –3 + 2c(n –3)/2a(n –3)/2) +
+ … + c(n –1)/2a(n –1)/2 = (c – a)2P + nc(n –1)/2a(n –1)/2 comes that:
c – a is divisible by n2, therefore R is divisible by n but is not by n2;
as R > n, then the number R has a prime factor r not equal to n;
c – a is not divisible by r;
if b = ntb', where b'1 ≠ 0, then the number c – a is divisible by ntn – 1 but is not by ntn.
§2. Lemma. All n digits (a1di)1, where di = 0, 1, … n – 1, are different.
Indeed, admitting that (a1d1*)1 = (a1d1**)1, we come to the following conclusion: ((d1* – d1**)a1)1 = 0.
And then d1* = d1**. Therefore, the sets of digits a1 (here along with a1 = 0) and d1 are the same.
[Example for a1 = 2: 0: 2x0 = 0; 1: 2x3 = 11; 2: 2x1 = 2; 3: 2x4 = 13; 4: 2x2 = 4.
If n is not prime, then the Lemma is not true: in base 10 both (2х2)1 = 4, and (2х7)1 = 4.]
§2a. Corollary. For any digit a1 ≠ 0 there is such a digit di, that (a1di)1 = 1.
[Example for a1 = 1, 2, 3, 4: 1x1 = 1; 2x3 = 11; 3x2 = 11; 4x4 = 31.]
[/QUOTE]