home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: sci.math
- Path: sparky!uunet!cs.utexas.edu!qt.cs.utexas.edu!yale.edu!yale!gumby!destroyer!sol.ctr.columbia.edu!usenet.ucs.indiana.edu!newshost.cs.rose-hulman.edu!news
- From: goddard@NeXTwork.Rose-Hulman.Edu (Bart Goddard)
- Subject: n doesn't divide .......
- Message-ID: <1992Aug28.230054.21269@cs.rose-hulman.edu>
- Sender: news@cs.rose-hulman.edu (The News Administrator)
- Nntp-Posting-Host: g214-1.nextwork.rose-hulman.edu
- Organization: Rose-Hulman Institute of Technology
- Date: Fri, 28 Aug 1992 23:00:54 GMT
- Lines: 61
-
-
- Allow me to nurse my ego a bit. I have a Ph.D, in number theory. I
- have taken on a project this summer, which is to produce a solution
- manual for someone else's number theory book. The book has 1235
- problems. I have worked 1214 of them so far, and TeXed them up. I
- posted one of the remaining 21 problems (to prove that n doesn't divide
- 2^n-1) which I spent about 2 hours on and thought someone might help me
- out a bit. Thanks to Adam Logan and Keith Ramsay who gave concise
- proofs. It's a big help for someone is has worked number theory
- problems until he is nearly cross-eyed. I intend to post a couple more
- of these problems during the next week or so as I finish this project.
- When I do, I would like to request that ALL ARROGANT ASSHOLES KEEP
- THEIR BULLSHIT OUT OF MY MAILBOX.
- Here a nice solution, appropriate to the level of the problem, given by
- Adam Logan:
- >If n is even, the result is trivial.
-
- >Otherwise, letn be an exception, and let p be the least prime
- >dividing n. Since n > 1, such a prime exists. By hypothesis, n | 2^n
- >- 1, so p | 2^n - 1, i.e. 2^n == 1 (mod p). Fermat proved that
- >2^(p-1) == 1 (mod p). If gcd(n, p-1) = 1, we can choose a and b with
- >an + b(p-1) = 1; then, raising the first equation to the power a, and
- >the second to the power b, and multiplying, we get 2 == 1 (mod p),
- >which is absurd. Hence gcd(n, p-1) > 1, contradicting the minimality
- >of p. (== means congruent to.)
-
- >Similar but harder problem: prove that n^2 | 2^n + 1 iff n = 1 or 3.
- >(The description of numbers n s.t. n | 2^n + 1 is not so simple.)
-
-
- Here's a sample of the drivel I've been getting:
-
- > Sure. Proof by extended example:
- >
- > n=2, 2^n-1=3 -> n obviously does not divide 2^n-1 for n=2
- > n=3, 2^n-1=8 -> n obviously does not divide 2^n-1 for n=3
- > n=4, 2^n-1=15 -> n obviously does not divide 2^n-1 for n=4
- > ...
- > QED.
- >
- > I'm sure that you can fill in the dots with all the time you have
- > from not doing homework...
-
- (har har har...You seem to have plenty of free time.)
-
- > Look up "Fermat's theorem" and the "Chinese remainder theorem"
- > in any book on number theory.
-
- (Big help. Actually, one would have to strain oneself to use the CRT
- on this problem.)
-
- I have a head cold just now, and may regret my tirade later when I'm
- not so grouchy, but I doubt it.
-
- Very, very sincerely,
- Bart
-
- P.S. If you meant your comments all in fun, and are surprised I took
- offence, I sorry, I'm just not in the mood. Further, keep any
- apologies to yourself, I don't want to read it, just consider it
- accepted.
-