My Math Forum (http://mymathforum.com/math-forums.php)
-   Number Theory (http://mymathforum.com/number-theory/)

 Bogauss February 15th, 2012 10:38 AM

Maybe maybe maybe Mister Greathouse remember what I told him long time ago.
He rebuked my assertion without even trying to check.
I will find the post.

 CRGreathouse February 16th, 2012 06:01 AM

I'm not sure what you're talking about, in particular, but thanks for the article.

 Bogauss February 16th, 2012 08:51 AM

Quote:
 Originally Posted by CRGreathouse I'm not sure what you're talking about, in particular, but thanks for the article.
Here is a quote of Momo my previous nick.
I do not know how to give the exact post but try in search ' Transparent semi-primes" then you will find all my thoughts at that time.

Two things you have to keep in mind :
1. All the keys to encrypt are PUBLIC that means that you could build a database of those numbers (a billion or more semi-prime numbers to factorize).
2. Every day I discover a new idea to add anoher algo to my "batterie" (in french I do not know how to translate it in english) of tests.

If each algo require 1 second to test any number and it allllows to factorize thousanda of numbers imagine what I can do with a hundreds.
I have calculated that 1 out 422 is solved for less than 100.000.000 using only one algo.

I'm sure that there is some very complex link between some values (and/or digits) that we can extract from any number and its factors.

So if I can factor one out of even 1.000.000 I can still hope to find the factors of the billion or more public numbers.
RSA works if you have only a few numbers known by the public but when you have billions known it will not work.

 Bogauss February 16th, 2012 08:54 AM

Here is the core of my idea.
It is a quote of my post
RSA works if you have only a few numbers known by the public but when you have billions known it will not work.

Today I can say that I was right!

 Bogauss February 16th, 2012 08:57 AM

Here is the more important post

viewtopic.php?f=40&t=2376&p=9681&hilit=transparent +semi+primes#p9681

There are others where I defended the same idea.

 CRGreathouse February 16th, 2012 09:06 AM

Quote:
 Originally Posted by Bogauss 1. All the keys to encrypt are PUBLIC that means that you could build a database of those numbers (a billion or more semi-prime numbers to factorize). 2. Every day I discover a new idea to add anoher algo to my "batterie" (in french I do not know how to translate it in english) of tests. If each algo require 1 second to test any number and it allllows to factorize thousanda of numbers imagine what I can do with a hundreds. I have calculated that 1 out 422 is solved for less than 100.000.000 using only one algo.
Ah. No, I still disagree with that, for reasons I probably outlined elsewhere. That someone has found an unrelated flaw does not change my feelings here. But on the other hand, I think you should still experiment with new methods -- I just don't think that the one outlined in that post is useful. (The chance that it would work on a given 2048-bit RSA key (or even an old 1024-bit key) is much less than one in a trillion trillion, so there are probably exactly zero RSA keys which have ever been used to which your method would apply. But even if it did apply, the expected time to factor the number would be something like 10^300 seconds, much longer than the life of the universe.) So keep looking! I wish you luck.

 Bogauss February 16th, 2012 09:26 AM

Quote:

Originally Posted by CRGreathouse
Quote:
 Originally Posted by Bogauss 1. All the keys to encrypt are PUBLIC that means that you could build a database of those numbers (a billion or more semi-prime numbers to factorize). 2. Every day I discover a new idea to add anoher algo to my "batterie" (in french I do not know how to translate it in english) of tests. If each algo require 1 second to test any number and it allllows to factorize thousanda of numbers imagine what I can do with a hundreds. I have calculated that 1 out 422 is solved for less than 100.000.000 using only one algo.
Ah. No, I still disagree with that, for reasons I probably outlined elsewhere. That someone has found an unrelated flaw does not change my feelings here. But on the other hand, I think you should still experiment with new methods -- I just don't think that the one outlined in that post is useful. (The chance that it would work on a given 2048-bit RSA key (or even an old 1024-bit key) is much less than one in a trillion trillion, so there are probably exactly zero RSA keys which have ever been used to which your method would apply. But even if it did apply, the expected time to factor the number would be something like 10^300 seconds, much longer than the life of the universe.) So keep looking! I wish you luck.

Factorization of large semi primes is over at least for me. I'm not kidding. Just try to reread me. I gave you the solution.
Keep in mind that those guy who have discovered the flaw I discovered more than 4 years ago have a large team of programmers very powerful computers money and so on.
I have an old computer. I'm not programmer. I use only Excel to check my ideas.
More than that it seems to me that you have short memory.
I can give other posts it is going to take me time. Going through 100"s of posts to find other arguments it is really hard even with the help of search program.

 CRGreathouse February 16th, 2012 02:26 PM