Re: JSH: Latest factoring idea is crap
- From: "Tim Peters" <tim.one@xxxxxxxxxxx>
- Date: Sun, 16 Jul 2006 05:12:12 -0400
[cut sci.crypt & alt.math]
[Proginoskes, to JSH]
...
A problem does not get to be known as "hard" until it has been attacked
by someone and refuses analysis.
The upshot is this:
If professionals --- and amateurs --- have tried the same approaches as
you and gotten nothing, what makes you think you will succeed?
[gjedwards@xxxxxxxxx]
Unfortunately, he thinks he already *has* succeeded. This factoring
thing is an interesting situation for him though, I wonder how he'll
eventually deal with the mess he's got himself into.
You're seeing it right now. He's done nothing more embarrassing this year
than last wrt factoring, but has picked up a marvelous new defense
mechanism: never, ever try to factor an integer anymore. That allows
delaying recognition of the truth much longer than he could back when he
still tried his methods.
With his earlier 'work', e.g. FLT, he seems to have eventually resorted
to simply believing it is correct, 'so there!' and deciding that he's
above the usual requirement of presenting your proof in a way that
others eventually accept is correct.
Actually, he believes that "top mathematicians" /know/ his FLT proof is
correct. He doesn't care what anyone on Usenet thinks, and since his proof
is good enough for top mathematicians, he's got no reason to try to make it
intelligible to us lesser beings. Hell, we're probably too dumb to
understand it anyway (as he's noted more than once).
Thus, he's the man who found a short proof of FLT, a fantasy that he
can hold together with ease.
Well, I'm not sure it's all _that_ easy to maintain the elaborate web of
other strange beliefs needed to support this; e.g., see the paragraph above
;-)
Because the factoring thing is applied, it's a much more difficult
fantasy to hold on to.
I used to think that too. But he makes ways. If he tried it himself, he'd
quickly learn that a method doesn't actually work worth beans, but he won't
make that minimal effort anymore. He has a pile of goofy rationalizations
for refusing to try, and adds to that pile over time (which I expect means
he's still sometimes having trouble convincing himself that it's not just an
avoidance mechanism).
So what's to pop the bubble? If other people report testing results, he
dismisses them for various reasons (some quite valid, some not).
Since he's attracted to messy methods involving multiple free choices, it's
often quite hard to rigorously prove that /no/ way of making those choices
can possibly lead to an efficient method -- and who's going to make that
effort? Not me, and I'm one of the few people left who will still gives
some serious consideration to his methods.
If someone did make that effort, the very fact that it's going to be a
tedious, involved proof means James can dismiss it as just another attempt
to wear him down with "long posts". What do you think the chances are that
he could /follow/ a proof even if he were serious about it? Before
answering ;-), note that it would almost certainly need to invoke results in
number theory that are far beyond his knowledge.
What /used/ to work every time was to find a specific example using small
integers (and I mean tiny -- two-digit factors at worst) demonstrating that
no possible way of applying his method leads to a factor. I actually hoped
it would be enough this time around to complete his partial attempt at
factoring 35 and show him that it didn't find a factor after all. He
doesn't want another "layers and layers of searching" method (to his credit,
I'll add).
But no. Now the failure to factor has been turned into "an unappreciated
feature": the method can factor integers other than the one you give it.
While it's become untenable to claim that it /is/ an efficient factoring
method, he can cling to the hope that there's some undiscovered efficient
way to make the free choices that /will/ lead to factors. And it's a safe
bet that nobody will ever care enough about a fundamentally broken method
with no apparent promise to /prove/ that's not possible.
In the meantime, he can enjoy believing that the answer is just sitting
there waiting to be found, and having given the world the solution he's
under no obligation to finish working out every little detail too.
Which would be fine by me. So long as he continues to believe this "is it",
I don't have to look at another new method ;-)
He's presenting it to non-mathematicians too, and by his own admission
its just simple algebra, so even if we grant him the temporary fantasy
of the Corrupt Society of Mathematicians, *somebody* should be factoring
numbers with it by now, if only to pick up the prize monies.
That does /seem/ like a hard one to get around. But I do believe he's
planted the seeds for rationalizing away that too. This kind of post shows
up rarely, but more than once:
From: jstevh@xxxxxxx
Newsgroups: sci.crypt,sci.math
Subject: JSH: Now what?
Date: 11 Jul 2006 17:53:43 -0700
Message-ID: <1152665623.853702.143950@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>
...
I've been looking for federal people all day, expecting the
FBI or even maybe someone from the NSA to show up, but nothing...
See the beauty of this? If nobody reports a factorization using the method
for a few months, it could very well be because Shady Government Agencies
got wind of it and put them on ice. That's just cool. The beauty part is
that it's another "reason" for James to avoid finishing the method, or
trying to factor an RSA challenge number, himself. Hell, he's /already/
risked government wrath by revealing as much as he has.
The sad fact is that in his heart or hearts, he KNOWS nobody will do
so, because he knows it's wrong.
Of course I expect that's the real reason he stopped trying to factor
anything himself.
If he doesn't give up on the current method, I expect it can live forever in
the same bin as his prime-counting formula. There always are x and y such
that gcd(x +/- y, T) reveal non-trivial factors of composite T, and that's
what his method uses at the end, so of course it has to work. If other
people are too dumb to find them efficiently, BFD -- they were also too dumb
to figure out how to make his prime-counting formula the most efficient one
known (despite that it _must_ be).
.
- Prev by Date: Re: power series
- Next by Date: Re: Probability to find a substring in a string
- Previous by thread: power series
- Next by thread: *unique* prime factorizations; the fundamental theorem of arithmetic
- Index(es):
Relevant Pages
|
