"New" primality test/check

[Nick]

@gophne:
You appear to be experiencing culture shock.

Imagine spending some time at one of the world's top universities.
You would find spaces devoted to socializing, spaces for education, and other spaces dedicated to research.
In the research spaces, the focus would be on ideas and results, at a very high level, and not on people.
Everyone is extremely critical of new ideas because that way we make progress faster.

This forum is organized in much the same way.
To socialize, you can browse people's posts and react in the Lounge or the Soap Box.
To learn the basics, head for the Number Theory Discussion Group.
But in the parts of the forum dedicated to new ideas and results, you can expect us to be highly critical and focus on the concepts, not the people.
You say you arrived at your results independently and worked very hard to do so. We do not doubt it.
But once it became clear that your result was equivalent to something already well known, we lost interest.
That is not personal - our reaction would have been the same whoever presented the idea. It is the normal culture in research.

[CRGreathouse]

Originally Posted by gophne

Danaj, did you read the link/paper that you posted. If you have, which i doubt, I think you simply "googled" a paper which you thought might support your argument.

Danaj is an expert, the author of Math::Prime::Util among others. He certainly did not just Google up a random paper.

Originally Posted by gophne

Honestly, if you really think that is EASY math, equivalent to what I have introduced to you, then you need to change your medication.

Wheel sieves and splitting consecutive integers into columns representing congruence classes mod the width are very easy math, yes. The Quesada & Van Pelt paper contains more advanced mathematics as well (but nothing terribly difficult, on a glance).

Originally Posted by gophne

I am in the process of preparing to submit my algorithms to other sites/forums, as well, SO THAT ANY FRAUDULENCY COULD BE EXPOSED FAR AND WIDE, but not by an apparent biased grouping.

That sounds like a plan. With any luck you'll get useful feedback elsewhere that will improve your exposition etc.

Originally Posted by gophne

Look I might be proved to be the world's greatest fraud in the end, but I believe in my own integrity.

Rediscovery is very common in mathematics, it's nothing to be ashamed of. (I don't think anyone here is impugning your integrity.) Ramanujan rediscovered a tremendous number of results and no one thought any less of him as a result. But if you want to publish you have a scholarly obligation to search for prior art and attribute original ideas to their respective discoverers. (There are plenty of papers out there reproving known theorems, just because the proofs are easier, shorter, or more enlightening -- but they are expected to cite the original work, even if their discovery was independent.)

Originally Posted by gophne

But what I do know that soon very soon at least one reader/contributor would appraciate what I have "discovered", much of which I have not posted, fearing a witch-hunt of the order of the Inquisition of the Dark Ages. The more times change....the more they remain the same.

That's a bit melodramatic, don't you think? Thousands were tortured or executed in the Inquisition, whereas here not only is no one hurting or silencing you, but you're even being provided a platform, for free, on which to express your thoughts.

Originally Posted by gophne

However, very soon (about a month or two) if I would be able to still post, I would like to attempt a Proof for the Twin Prime Conjecture.

I'm not surprised -- there have been dozens if not hundreds of claimed proofs of the twin prime conjecture, it's one of the most popular targets of cranks. Dudley has two or three books on the general subject, though I don't recall if he included examples of twin primes proofs.

Originally Posted by gophne

Not sure if it would be a valid proof.

That modesty might be your saving grace. Good luck, and may your proofs always hold.

[CRGreathouse]

Originally Posted by Nick

Imagine spending some time at one of the world's top universities.
You would find spaces devoted to socializing, spaces for education, and other spaces dedicated to research.
In the research spaces, the focus would be on ideas and results, at a very high level, and not on people.
Everyone is extremely critical of new ideas because that way we make progress faster.

This forum is organized in much the same way.
To socialize, you can browse people's posts and react in the Lounge or the Soap Box.
To learn the basics, head for the Number Theory Discussion Group.
But in the parts of the forum dedicated to new ideas and results, you can expect us to be highly critical and focus on the concepts, not the people.
You say you arrived at your results independently and worked very hard to do so. We do not doubt it.
But once it became clear that your result was equivalent to something already well known, we lost interest.
That is not personal - our reaction would have been the same whoever presented the idea. It is the normal culture in research.

Well said.

[gophne]

Originally Posted by Nick

@gophne:
You appear to be experiencing culture shock.

Imagine spending some time at one of the world's top universities.
You would find spaces devoted to socializing, spaces for education, and other spaces dedicated to research.
In the research spaces, the focus would be on ideas and results, at a very high level, and not on people.
Everyone is extremely critical of new ideas because that way we make progress faster.

This forum is organized in much the same way.
To socialize, you can browse people's posts and react in the Lounge or the Soap Box.
To learn the basics, head for the Number Theory Discussion Group.
But in the parts of the forum dedicated to new ideas and results, you can expect us to be highly critical and focus on the concepts, not the people.
You say you arrived at your results independently and worked very hard to do so. We do not doubt it.
But once it became clear that your result was equivalent to something already well known, we lost interest.
That is not personal - our reaction would have been the same whoever presented the idea. It is the normal culture in research.

Hi Nick

Good post, I like it very much (I do not know yet how to add smileys to post like some of the other contributors, but I would have liked to add a "good post" smiley to your post, because eventhough you are very thorough and firm in your stance (and clearly not impressed with time wasting and spiels), you never-the-less criticise in such a way that it actually feels good! Seriously.

For your information I am contesting the verdict that the algorithm in question was in fact a copy or clone (of Fermat's Primality Test or another)...BUT NOT ON THIS SITE OT THREAD, I promise!!!!!!

I am contesting this finding on five grounds (I am working with friends who are having another closer look at this matter and the supporting premises for the current accepted status, much of it which is due to my own acceptance of the suggestions of sameness), but the conundrum is this as well....I am such a dunce, why is it that when I accept suggestions of the "sameness" of the posted algorith, I suddenly become a guruji.

For you information, should you be interested at all, which I doubt and don't blame you for, and won't bore you with it as well! the grounds for the contestation are the follow, just for the information of other readers;

1) Fermat's Primality test works with congruancy, whilst the accused algorithm works with equality (changing remainders for changing inputs)

2) Reducing the formula's of the two algorithms to standard form, LHS mod x, furnishes different dividends.

3) Changing the base of the dividend from base 2, does not furnish the same remainders. For Fermat's Primality the remainder/congruancy is always one(1) for prime witness.

4) Fermat's Primality Test does not utilize mersenne numbers at all! the accused algorithm has the mersenne form as the dividend of the algorithm!

5) Running the suspect algorithm for false negatives up to 1,000,000 none was found -This has been confirmed unintentionally by one of the senior contributors on the site that was running the algorithm in Pari, I think -check some of the earlier posts.
Fermat's Primality check has a problem with false negatives (Carmichael Numbers).

It is only over once the fat lady sings, but this is not for this Site or Forum.

Thanx and regards.

[jnml]

Originally Posted by gophne

5) Running the suspect algorithm for false negatives up to 1,000,000 none was found -This has been confirmed unintentionally by one of the senior contributors on the site that was running the algorithm in Pari, I think -check some of the earlier posts.
Fermat's Primality check has a problem with false negatives (Carmichael Numbers).

Please link that post you're talking about, I'm not sure which you mean, thanks.

[gophne]

Originally Posted by jnml

Sure.

Hi jnml

Don't be so harsh on yourself. I know the feeling too.

[gophne]

Originally Posted by CRGreathouse

Danaj is an expert, the author of Math::Prime::Util among others. He certainly did not just Google up a random paper.



Wheel sieves and splitting consecutive integers into columns representing congruence classes mod the width are very easy math, yes. The Quesada & Van Pelt paper contains more advanced mathematics as well (but nothing terribly difficult, on a glance).



That sounds like a plan. With any luck you'll get useful feedback elsewhere that will improve your exposition etc.



Rediscovery is very common in mathematics, it's nothing to be ashamed of. (I don't think anyone here is impugning your integrity.) Ramanujan rediscovered a tremendous number of results and no one thought any less of him as a result. But if you want to publish you have a scholarly obligation to search for prior art and attribute original ideas to their respective discoverers. (There are plenty of papers out there reproving known theorems, just because the proofs are easier, shorter, or more enlightening -- but they are expected to cite the original work, even if their discovery was independent.)



That's a bit melodramatic, don't you think? Thousands were tortured or executed in the Inquisition, whereas here not only is no one hurting or silencing you, but you're even being provided a platform, for free, on which to express your thoughts.



I'm not surprised -- there have been dozens if not hundreds of claimed proofs of the twin prime conjecture, it's one of the most popular targets of cranks. Dudley has two or three books on the general subject, though I don't recall if he included examples of twin primes proofs.



That modesty might be your saving grace. Good luck, and may your proofs always hold.

Hi CRGreathouse

I like everything you say, and finds your comments very uplifting, yet firm yet balanced.

I shall take to heart all you say and model my responses making use of your counsel.

What can I say, ever since my first post on this thread, making outlandish claims, you have been very positive yet authoritive in demanding a professional standard w.r.t claims made, research done, supporting evidence required, etc. You came/come accross as a Sage who can see the folly in other mortals, but do not stamp on them like they are insects because of their ignorance/stubbornness/foolhardiness. For this I thank you, and I think I speak for many others, especially the inexperienced contributors on the site.

Whatever happens on the site/thread in days to come, I will only have gratitude for your approach, even to the unwise, as I admit I could be, or possibly am.

[jnml]

Originally Posted by gophne

Hi jnml

Don't be so harsh on yourself. I know the feeling too.

> Quote:
> Originally Posted by gophne
> 5) Running the suspect algorithm for false negatives up to 1,000,000 none was found
> This has been confirmed unintentionally by one of the senior contributors on the site that
> was running the algorithm in Pari, I think -check some of the earlier posts.
> Fermat's Primality check has a problem with false negatives (Carmichael Numbers).

Please link that post you're talking about. Making claims that no one can verify is a typical crackpot tactic. Don't be a crackpot.

[science_man_88]

http://www.mersenneforum.org/showthread.php?t=14901

[danaj]

Originally Posted by gophne

Danaj, did you read the link/paper that you posted. If you have, which i doubt, I think you simply "googled" a paper which you thought might support your argument.

As my text indicated, I did read it. I spent many nights after work looking at the paper and a white board, trying to figure out how to speed up my SoE.

I put a reference to the paper in my project documentation in 2012. Github backs me up on this.

I was mainly pointing you to their matrices looking at arrangements of primes/composites/squares. I did a crapton of that on my whiteboard trying to figure out how to write code to exploit it. Which makes me really impressed with some of the people who internalize this more.

[gophne]

Originally Posted by jnml

Please link that post you're talking about, I'm not sure which you mean, thanks.

Hi jnml

I was referring to your post #60 it seems. - 0% negatives