Re: JSH: Contradictory behavior, issue of math fraud


"rossum" <rossum48@xxxxxxxxxxxx> wrote in message
news:g2lqd3l9b22hig0ql2tj52ab9bhte7kf90@xxxxxxxxxx
On Mon, 03 Sep 2007 17:28:34 -0700, JSH <jstevh@xxxxxxxxx> wrote:

The terrible thing is that the potential for surrogate factoring may
be with very large numbers where that number of probes is significant
and far better than random, but with small numbers and an inability to
pick k and n properly it can vastly inflate that number of checks to
support opinions against surrogate factoring.
Easy enough to show - tell us how to pick k and n for large numbers.
Unless there is a quick way to get to the right choices for k and n
then your method is likely to be slower than existing methods for
large numbers.

The problem with your method is that it cannot get quickly to the
right choices. Currently all we seem to be able to do is to guess
starting points and go from there. I have seen versions where you
pick n and vary k as well as vice-versa. I have seen you use k=1,
k=2, k=30, k=floor(T/30) and others.

You need to concentrate on how to pick k and n from a much narrower
range than currently - it is the wide range of possibilities to be
tried that is slowing your method.

rossum


Surrogate Factoring is DEAD.


.

Quantcast