You are not logged in. Login Now
 0-24   25-49   50-74   75-81       
 
Author Message
7 new of 81 responses total.
brighn
response 75 of 81: Mark Unseen   Jun 19 18:41 UTC 2002

Aruba, if I'm reading correctly, you do have an error. You say that if p does
not divide a and doesn't divide n, it doesn't divide na. What about p = 6,
n = 4, a = 3? Am I getting "p divides a" backwards?
aruba
response 76 of 81: Mark Unseen   Jun 19 19:15 UTC 2002

There's an interesting application to repeating decimals.  (Try this if you
really want to amaze people at parties - I've tried it - it works!)

You can generalize the fact that 1, 2, ..., p-1 form a multiplicative group
for any prime p to say that for any integer n, the positive integers which
are less than n and relatively prime to n form a group (mod n).  The size
of that group is called phi(n), the "Euler phi function".  For instance,

 n  numbers m < n with gcd(m,n) = 1            phi(n)
 -  -------------------------------            ------
 2  1                                               1
 3  1,2                                             2
 4  1,3                                             2
 5  1,2,3,4                                         4
 8  1,3,5,7                                         4
 9  1,2,4,5,7,8                                     6
12  1,5,7,11                                        4
18  1,5,7,11,13,17                                  6
21  1,2,4,5,8,10,11,13,16,17,19,20                 12
60  1,7,11,13,17,19,23,29,31,37,41,43,47,49,53,59  16

Question: how long is the repeating decimal for 1/n?  Well, of course, it
goes on forever, but what I mean is, how many digits repeat?  For
instance, I'll say that:

length(.111111 ...) = 1
length(.373737 ...) = 2
length(.024024 ...) = 3,

etc.  We've all seen this method of writing a repeating decimal as a
fraction:

    X = .579579579...
1000X = 579.579579...
-----   -------------  (subtract top from bottom)
 999X = 579

==> X = 579/999.

More generally, if the decimal representation of X is a string S of k
digits repeated, then X = S/(10^k-1).

So suppose you're going the other direction; you know X=1/n (for the
moment assume n is relatively prime to 10), and you want to know
the length k of its repeating decimal.  Well, k will be the smallest
number such that you can write 1/n as S/(10^k-1), for some S.  But that
implies that S*n = 10^k-1, so actually k is the smallest integer such that
n divides 10^k-1, i.e., k is "the order of 10 (mod n)".

Now 10 is a member of the group above, so we know that its order must
divide the order of the group, namely, phi(n).  So:

/---------------------------------------------------------------------\
| If n is relatively prime to 10 and k is the length of the repeating |
| decimal for 1/n, then k divides phi(n).                             |
\---------------------------------------------------------------------/

There's actually a pretty simple way of calculating phi, if you know the
prime factorization of n.  If n = p1^k1 * p2^k2 *** pr^kr, then

phi(n) = n(1 - 1/p1)(1 - 1/p2)...(1-1/pr).

So, for instance, the prime factors of 21 are 3 and 7, so
  phi(21) = 21 * (2/3) * (6/7) = 12.
That means that the length of the repeating decimal for 1/21 is a factor
of 12 - i.e., 1, 2, 3, 4, 6, or 12.
(In actuality, the answer is 6:  1/21 = .047619 repeating).

This is as far as I've ever been able to take the analysis; maybe someone
can take it farther.  (It's not hard to dispense with the case where n has
some facors in common with 10; basically, you throw out those factors
before finding phi(n)).

Note also that 10 hardly appeared in the answer; you could replace it by
any base b.
aruba
response 77 of 81: Mark Unseen   Jun 19 19:17 UTC 2002

Re #75: p is prime, so p can't be 6.  If a prime divides ab, then it must
divide either a or b.
brighn
response 78 of 81: Mark Unseen   Jun 19 21:59 UTC 2002

Ah. My mistake.
orinoco
response 79 of 81: Mark Unseen   Jun 20 01:00 UTC 2002

Does it work as a way of amazing people at parties, or only in a strictly
mathematical sense?
aruba
response 80 of 81: Mark Unseen   Jun 20 03:28 UTC 2002

Well, I guess it depends on which parties you go to.
orinoco
response 81 of 81: Mark Unseen   Jun 21 01:08 UTC 2002

Mark, if you ever throw the sort of party that trick will work at, I'm there.
;)
 0-24   25-49   50-74   75-81       
Response Not Possible: You are Not Logged In
 

- Backtalk version 1.3.30 - Copyright 1996-2006, Jan Wolter and Steve Weiss