Re: Prime magnitude and keys...a ?

[Jim choate <ravage@bga.com>]

> 
> And you think no one would have noticed such a thing before.
>
Is a possibility...especially since I can find no reference to it
or why it won't work.

> I can pretty much hint to you that such a thing can't really be done
> in log base 2 of n time in the sense that I believe I can prove that
>
This is a joke right? Why in the world should the base have a damn thing
to do with the algorithm? A number is a number last time I checked.

any algorithm that did that would have to involve none of the basic
> four arithmetic operations on the numbers in question. (Algorithms
> involving no arithmetic on the numbers are still possible, but
> intuitively quite unlikely.)
>
Sorry, I don't follow your reasoning here at all. Could you clarify?

As far as I am concerned if it could be done w/ a neural network, or
boolean algebra (course if no arithmetic ops no logic I guess), or
even a fuzzy algorithm (the original impetus to this line, I was looking
at "close enough" algorithms for a robot project I am in the middle of.
) would be ok by me. 

Seems to me though that if one looks at the results of the operation one 
could glean some sort of magnitude info out of the errors...

> Perry
>