Re: Surrogate factoring, reasons for my concerns

From: Will Twentyman (wtwentyman_at_read.my.sig)
Date: 07/13/04 

Date: Tue, 13 Jul 2004 15:27:56 -0400

James Harris wrote:

> Jeroen Boschma <boschma@fel.tno.nl> wrote in message news:<40F39740.590EB00B@fel.tno.nl>...
>
>>James Harris wrote:
>>
>><SNIP>
>>
>>>You have to completely factor T^2 - 1 and consider ALL the possible
>>>values for j that result from ALL possible combinations of those
>>>factors before you can claim a particular conclusion.
>>>
>>
>>So let's do that. In fact I already did. I wrote a little piece of Matlab code (took me about 20
>>minutes) which generates 100 random pairs of primes, each prime less than 1000. Those 100 pairs of
>>primes give me 100 composite T's which I fed to the algorithm. And yes, for each T I calculated f1
>>and f2 (thus j) for all possible combinations of factors of T^2-1. At the end, it turns out that
>>only 3 T's out of 100 are factored.
>
>
> Thanks!!! That's interesting. What size numbers?
>
>
>>Conclusion: the algorithm is worthless in its present form.
>
> That's not surprising, but your conclusion is weird.
>
> 3% is actually rather good, and definitely good enough to cause
> worries if your results hold with large numbers, though I doubt they
> will.

His conclusion makes sense. 3% success rate is not something that would
inspire me to code your method into an algorithm first, hoping it works,
then do something else more reliable. The added computation time for
the failures when factoring a collection of numbers is not likely to pay
off. 

-- 
Will Twentyman
email: wtwentyman at copper dot net

Relevant Pages

  • Re: Surrogate factoring, reasons for my concerns
    ... James Harris wrote: ... > That's not surprising, ... the failures when factoring a collection of numbers is not likely to pay ...
    (sci.math)
  • Re: Surrogate factoring, reasons for my concerns
    ... > James Harris wrote: ... >> That's not surprising, ... factoring a huge composite. ... communications and we're all interested I'm sure. ...
    (sci.math)
  • Re: Surrogate factoring, reasons for my concerns
    ... > James Harris wrote: ... >> That's not surprising, ... factoring a huge composite. ... communications and we're all interested I'm sure. ...
    (sci.crypt)
  • Re: Surrogate factoring demonstrated
    ... I expect that everyone seriously working on factoring has wondered whether ... common in several factoring algorithms. ... That's not surprising, because ... Nora was explicitly trying to show how simple the idea _could_ be. ...
    (sci.crypt)
  • Re: I dont understand why my phone system does what it does.
    ... be a regulatory violation is doubtful in my mind, ... maintenance to their infrastructure and timely response to trouble reports, ... surprising those that do, however. ... I wish EVERYONE would pay for out service but not insist on ...
    (alt.home.repair)