View Single Post
Playdo
#13
Jul3-06, 01:09 PM
P: 92
[QUOTE=shmoe]

This is not an equality[\quote]

The caveat is [itex]n+1=(m+1)(s+1)[/itex] in which case it definitely is an equality as [itex]a^ib^i = (ab)^i[/itex] so technically it simply starts to ask the implications of the base in (1) of the first post being composite. A general useful goal in factoring is finding a way to exchange factorization of a large number for the factorization of a smaller one. If you keep that in mind you can chain together any number of apparently unrelated algorithms and in whatever order, so long as at the link you can get back and forth (think in pictures). I admit its a bunch of scratch paper so far. It will improve over time.

Your issue with the totient, your are right I made a mistake. I will fix that shortly. In fact primes have maximal totients compared to all numbers less than them. The result applies when [itex]\tau(k)>1[/itex]. That will be corrected. In fact it can be easily seen to be true by examining the formula for the totient function (You have to rewrite it, most books have it in a form where the terms are [itex](1-\frac{1}{p_k})[/itex]. I will go back and clarify that when I get a chance.


Over time I will work out examples. My understanding is that the two most popular sieve forms are modular sieves and quadratic sieves, the last two RSA numbers to be had fell to these and three months worth of massive parallel computing.

The story line here is simple. It is an adventure, nothing more. But there is a specific problem that is being looked at using a few basic concepts. The issue is how are composites and primes related? Clearly no number can be prime unless the primes below its root fail to divide it. Why there would exist algorithms that are faster than this is of interest. Behind all of this is a question of complexity as Chaitin defines it. Eventually I will flesh this thread out, but you may have to wait for a while. I don't get paid to do this, it is a hobby that I started enjoying when I first became intersted in mathematics. Although I do know some of the latest in cryptography and have a good backgorund in number theory and math in general, I am not pursuing this in the interst of learning what 'they' say is possible. This is experimental mathematics done in the spirit of the early number theorists. We have a toolbox of mathematical tricks and a set of objects to be investigated. I will take pleasure in revealing a picture that is eventually understandable to people outside the ivory tower. The important thing for me is that this is experimental mathematics and it is recreational. I do not expect to win a fields medal here. I do it because I like it, and therefore I do it the way I want to do it. That said I know the difference between a meaningful result and triviality. Sometimes even things that appear trivial require something more then 'it's obvious that' which is so popular among mathematicians and have started to pervade our textbooks to the point that many proofs cannot be followed confidently even by those with great skill. IMHO you might as well not even include a proof if you are only going to give half of it.