Hi Stas.
Adieu and dosvidanya
to the forge,
may it R.I.P.

Hey, thanks for running the group,
for so many years. Sorry for the
abrupt revelation that I was the
actaul designer of most of the
pnetium cpu memory circuitry.
I really should have
mentioned it sooner.
But, I was busy.

Does that new theory book mention anything
about greedy two clause generators?

After discovering them, in 2006, I was
really amazed by the results:

Twenty one regular graph theorems

These theorems are all of the form:
Almost all r regular graphs are c colorable.
And the theorems range from lower degrees,
as with n5_3, n6_4, n7_4, n8_4, n9_4, n10_5 ...
on up to n25_7, and surely there are
many more similar theorems to follow.

The twenty one theorems were all proven by:
A First guess the proper threshold.
(r*c was used for most of the twentyone.)

B Above the threshold, almost all graphs
with the given r regular degree
are c colorable.

C Then generate one of those medium sized randomly;

D Then c color the single randomly generated graph.

Almost all of the twenty one theorems were so proven
on the first try, with the r*c threshold.
The larger ones did take several hours of runtime,
with some thousands of booleans in the translation.
I did end up putting some extra effort into
the random number generator, just to verify
that the results were for real.

I spent 2007 polishing the equations,
and I have actually managed to, get this,
remove all negations from the equations!!!

The present version of the equations has:
zero negations,
zero minusses,
zero divides,
zero multiplys.
zero negative numbers, so types are unsigned,
so verification of correctness is easy.
the equations use only PLUS ONE,
as the only math operand.
there are a few hundred comparisons,
all lessthan or equals comparisons.
there are less than two hundred forloops.
there are about five hundred assignment operations.

Would be really great to get an office
somewhere in my own country
where these equations
could be put in the bank.
I call it the holygrail,
and would really love to
help build the big coloring book,
using the above mentioned twenty one theorems
as some of the chapters.

The books value would be enormously helpful,
for mining through graph math, and maybe even for
help solving governmental Identitys for Entitys
issues.
But, who would ever believe me?

Well hell, none of them even believed the
Nearly Square Google benchmark was even plausible...
just the same way as few people believe me
when I say:
Hey you,
ever heard of
the Pentium CPU?
I done dat!

Well I did, in 1992 or so.
Anyways, I hope to
come back to the Valley in March,
and hold out my cup for change...
because I do not have enough for
running toilet water in pennsylvania,

Many Cheers, and,
Be Seeing You,
Subtle village references intended,
Daniel

--- Stas Busygin <busygin@...> wrote:

> Hi All,
>
> I've just found that a really thorough textbook
> "Complexity Theory: A
> Modern Approach" by S. Arora and B. Barak is
> available online:
> http://www.cs.princeton.edu/theory/complexity/
>
> It may be a good replacement for Garey&Johnson as it
> cover more recent
> developement.
>
>
> --Stas
> http://www.stasbusygin.org
>
>
>




________________________________________________________________________________\
____
Be a better friend, newshound, and
know-it-all with Yahoo! Mobile. Try it now.
http://mobile.yahoo.com/;_ylt=Ahu06i62sR8HDtDypao8Wcj9tAcJ