Talk:Legendre's conjecture

Related Conjectures

If any one of the following is proven true, then Legendre's is proven true too.

Andrica's: p_n+1^1/2-p_n^1/2<1 for all n => 1.

Opperman's: pi(n²+n) > pi(n²) > pi(n²-n) for n > 1.

Brocard's: pi(p_n+1²) - pi(p_n²) => 4 for n => 2.

Grimm's: if n+1, ..., n+k are consecutive composite numbers, then we can find distinct prime numbers p_i so that p_i divides n+i.

and Sierpinski: For every integer n > 1, let the n² integers 1,2, ... ,n² be written in an array with n rows, each with n integers, like an n x n matrix, then there exist a prime number in each row.

Reddwarf2956 (talk) 21:41, 15 July 2008 (UTC)

Not that it matters much, but the last conjecture is due to Schinzel, not Sierpinski, although the latter was a coauthor of the paper in which the result appeared. - CRGreathouse (t | c) 01:23, 30 July 2022 (UTC)

Regardless, shouldn't these be added? Supyovalk (talk) 18:04, 14 November 2022 (UTC)

Not without a reliable source. JBL (talk) 19:30, 14 November 2022 (UTC)

Proof

I've proven that Legendre's conjecture is true. Please visit... www.oddperfectnumbers.com to see my solution; I've included a proof for Legendre's {stronger} conjecture as well. Enjoy! Bill 99.118.131.124 (talk) 02:21, 26 February 2012 (UTC)

Thank you for posting this to the talk page and not the article. As you may know, Wikipedia has policies against original research and conflict of interest edits. If your work is published by a reliable source like a peer-reviewed journal then you can suggest addition to the article. PrimeHunter (talk) 03:03, 26 February 2012 (UTC)

I've proven that Legendre's conjecture is true. Please visit Yu, W. (2023) The Proofs of Legendre’s Conjecture and Three Related Conjectures. Journal of Applied Mathematics and Physics, 11, 1319-1336. doi: 10.4236/jamp.2023.115085 — Preceding unsigned comment added by Wky2023 (talk • contribs) 18:59, 1 June 2023 (UTC)

Oh, lovely, a SCIRP journal. --JBL (talk) 19:12, 1 June 2023 (UTC)

It would be very unprofessional to blame the journal if you can't find any errors in the paper. — Preceding unsigned comment added by Wky2023 (talk • contribs) 03:36, 2 June 2023 (UTC)

It would be very unprofessional to publish high-quality mathematics with such a low-quality publisher, and force readers to sift through all of the other bad content from that publisher to figure out whether it's any good. Nobody has time for that. That would be a worse strategy for getting your work read and used than just putting up a preprint and leaving it at that. The purpose of a journal is to act as something of a filter so that readers can have some expectation of the reliability and significance of the published content. —David Eppstein (talk) 06:24, 2 June 2023 (UTC)

The value of a paper lies in the content rather than the carrier, and the magazine is only the carrier of the paper. Just as gold always shines, whether it's in the dirt or in a jewelry box. Wikipedia is a platform for sharing and exchanging knowledge. The author of the paper shares his research results here. No one is forcing you to read his paper. It is your choice. Those who accuse others of knowing how to prove Legendre's conjecture but themselves not knowing how to do it can only prove their ignorance. I am not interested in discussing which journal is better, I will only answer questions about the discussion of the paper. — Preceding unsigned comment added by Wky2023 (talk • contribs) 16:04, 2 June 2023 (UTC)

When

When did this conjecture get first published? John W. Nicholson (talk) 00:48, 4 June 2014 (UTC)

"Most historical accounts of the Prime Number Theorem mention Legendre's experimental conjecture (made in 1798 and again in 1808)...." http://www.mathpages.com/home/kmath032.htm 2601:18C:C601:7D90:C8BB:88CA:8966:7BAB (talk) 15:58, 1 February 2016 (UTC)

Unsolved ?

Is it correct to say that a statement is unsolved? Personally, I would say that only problems, questions, puzzles and the like can be unsolved. What I think is meant is that the statement is neither proven nor disproven. Can I change to that? --Ettrig (talk) 15:03, 27 August 2015 (UTC)

A conjecture is a question which has not been solved, so no don't change it. Yes, proving/disproving it will solve the question at hand, but the language is correct. John W. Nicholson (talk) 15:33, 27 August 2015 (UTC)

It is your personal view that A conjecture is a question. This view is not supported by the articles in Wikipedia or Wictionary on Conjecture. I think you are referring to a special meaning which is common among mathematicians (I can see why it is) but cannot be assumed in a Wikipedia article. --Ettrig (talk) 15:47, 27 August 2015 (UTC)

I will put it like this, you can state the conjecture, I will state the question that the author of the conjecture is asking. In this case it is Legendre asking 'is there a prime number between each natural square?' John W. Nicholson (talk) 20:11, 27 August 2015 (UTC)

I have no objection to the proposed alternative wording -- it seems at least as clear as the current wording. --JBL (talk) 23:07, 28 August 2015 (UTC)

Proof of Legendre Conjecture

I think that the conjecture is very close to proof, as I myself have proved it, Or probably have made it very close to proof. Sandeep Kumar Kandi (talk) 13:49, 12 June 2020 (UTC)

And once it is published as a WP:RS, we could add something about it to the article. But not before then (per WP:OR). --JBL (talk) 14:33, 12 June 2020 (UTC)

conjecture subject matter

(n²+1)² -minus- n² = equality = numer. this is a prime number and the only one: which, most likely, is exactly what he was talking about the "Legendre hypothesis" with an incomplete explanation of such words. Nean Like 16:38, 20 November 2020 (UTC) — Preceding unsigned comment added by Nean Like (talk • contribs)

consistence

et the таким орбразом: (n+1)²-минус-n²=is a prime number between n² and (n + 1)² for every positive integer n.

that is, this number is their distance between these prime numbers in square multiplication. Nean Like 14:04, 21 November 2020 (UTC)

consistence 2

Is it really not trivial ? Make the Square Root of all terms. Then you get back that an Irrational (a Prime is never a square) is packed between two following integers... that is not trivial ? — Preceding unsigned comment added by StefanoMaruelli (talk • contribs) 18:00, 24 November 2020 (UTC)

It is trivial that every prime number lies between two consecutive squares, but that is not what Legendre's conjecture asserts.
This page is for discussing improvements to the article, not for discussion of the article subject; your question would be better placed at the reference desk. --JBL (talk) 18:06, 24 November 2020 (UTC)

Thanks, that's clear, and was just a suggestion for a note into the article... — Preceding unsigned comment added by 87.26.160.199 (talk) 07:26, 25 November 2020 (UTC)

Suppressed vague and unsourced assertion

I have suppressed the vague and unsourced assertion according to which : "The prime number theorem suggests that (*) the actual number of primes between n² and (n + 1)² OEIS: A014085 is asymptotic to n/ln(n'’). Since this number is large for large n, this lends credence to Legendre's conjecture." I left, however, the reference to the sequence in the external links.

The original assertion, in which one had the unreasonably optimistic verb « implies » in place of « suggests », was unsourceable in addition of being unsourced, as it would imply the truth of Legendre’s conjecture for all integers surpassing some large integer N (which, as of today, remains of course unproved as well). But replacing « implies » by « suggests » does not make the sentence much more convincing. Without any reliable source provided all I can think the prime number theorem would be able to provide is the existence of an infinite series of unspecified integers for which (*) is true, through some elaborate version of the pigeonhole principle. I am a number theorist, and this does not in any way « suggest » to me that (*) remains true for every n: a reliable source, by a reliable author at least stating that assertion, is definitely needed. (Moreover, one might add that this type of arguments (using the pigeonhole principle) does not need, by far, an estimate as strong as the prime number theorem in order to ensure that Legendre’s conjecture is true for an infinity of integers n: clumsy as well as unsourced). Sapphorain (talk) 14:23, 25 July 2022 (UTC)

...On second thought, it appears a use of the prime number theorem does not even yield the existence of an infinite series of unspecified integers for which (*) is true, but only the existence of an infinite series of unspecified integers for which the number of primes between n^2 and (n + 1)^2 is at least n/logn(1+o(1)). So if kept this dubious assertion would need to be (1) made more precise, and (2) first and foremost backed by a sound and reliable source. --Sapphorain (talk) 22:00, 25 July 2022 (UTC)

Added a source for this obvious line of heuristic reasoning. Sapphorain's objections are unreasonable: of course it is not a proof, because if it were a proof then the conjecture would not be a conjecture. That does not make it dubious, unsourceable, or worth removing. —David Eppstein (talk) 23:16, 25 July 2022 (UTC)

Thank you for providing a reference. With a source confirming it, it is now justified to include something short like « the prime number theorem suggests Legendre’s conjecture is true ».

But certainly not the present assertion, which «  suggests » that a much more precise property holds, namely (once more) that the number of primes between consecutive squares n^2 and (n+1)^2 is asymptotic to n/logn. This is not mentioned in your source. And this is very dubious.

You apparently didn’t read carefully the source you provided, nor what my objection was (and still is) about.--Sapphorain (talk) 07:50, 26 July 2022 (UTC)

The average number of primes between consecutive primes is exactly of the form shown. This is a trivial calculation under WP:CALC. Both our article and the source suggest that the number of primes cannot be distributed far from its average, reasonable and I think even likely but far from provable. —David Eppstein (talk) 16:39, 26 July 2022 (UTC)

Thank you again, for providing more convincing sources. The first source should not be given right after the suggested asymptotic equivalence, as it doesn’t even address this issue. If kept, it should be put elsewhere. --Sapphorain (talk) 22:23, 26 July 2022 (UTC)

It does address the issue, if you actually read it. The quote given from the source is exactly on-topic. It is about the fact that the average number of primes over intervals of this length is as stated, that it seems unlikely for the primes to be distributed so non-uniformly over intervals of this length, and therefore that the prime number theorem strongly suggests the truth of Legendre's conjecture. It doesn't contain the exact wording used here, but we don't want sources to contain the exact wording we use (that generally indicates a copyvio problem). You have to use some understanding to interpret the sources, something you appear to be stubbornly refusing to do. —David Eppstein (talk) 23:25, 26 July 2022 (UTC)

The quote given from the source just says it is unlikely the primes are clustered in such a way they completely avoid a whole interval between consecutive squares, it does not in any way even hint there could be an asymptotic equivalence for the number of primes in such an interval. The author just goes on with « a random scattering of these primes strongly suggest at least one prime per slot » (referring to this as the "Prime-Square Betweenness Conjecture"). The stubborn person is you, not me.--Sapphorain (talk) 07:59, 27 July 2022 (UTC)

The claim in the article "this lends credence to Legendre's conjecture" is exactly the same thing, that it is unlikely the primes are clustered in such a way they completely avoid a whole interval between consecutive squares. —David Eppstein (talk) 17:13, 27 July 2022 (UTC)

You now provided a good source (Bazzanella) for the expected number of primes between consecutive squares n^2 and (n+1)^2, and this is very satisfactory. Unfortunately you persist in keeping the Francis source for backing the sentence that follows, « Since this number is large for large n, this lends credence to Legendre’s conjecture », although Francis says nothing concerning the said expected number (and does not refer to Bazzanella’s paper; incidentally, Francis' three only references are three papers by himself). This is inaccurate and clumsy. --Sapphorain (talk) 06:51, 28 July 2022 (UTC)

I think that David Eppstein has more than made his point; the sourcing is solid for both sentences. Now, if you think the wording is clumsy, feel free to suggest a replacement. Otherwise I think this discussion has run its course. - CRGreathouse (t | c) 01:02, 30 July 2022 (UTC)

The wording is not clumsy. Sourcing the second sentence with Francis' article is clumsy, because this article doesn’t address the more specific topic of the first sentence. Any of the other 4 sources of this paragraph (or even no source at all) would be more appropriate. --Sapphorain (talk) 07:34, 30 July 2022 (UTC)

Here are my two cents

For 1<=n<=6, we'll have to check each one on its own:
from 1 to 4 (n=1): 2 and 3.
from 4 to 9 (n=2): 5 and 7.
from 9 to 16 (n=3): 11 and 13
from 16 to 25 (n=4): 17, 19, and 23.
from 25 to 36 (n=5): 29 and 31.
from 36 to 49 (n=6): 37, 41, and 47.
Now for n>=7, since every n since have an entire set of 10s between n^2 and (n+1)^2, we can use Dirichlet's theorem to prove that for every m in N, there exists k in N cap [0,9] such that 10m+k is prime, so there is always a prime then as well. Serouj2000 (talk) 09:05, 21 June 2023 (UTC)

It's unclear what you mean but Dirichlet's theorem on arithmetic progressions does not say anything about the size of primes, only their form. It cannot be used to prove Legendre's conjecture. Also, Wikipedia content must be based on published reliable sources, not original research. PrimeHunter (talk) 23:22, 21 June 2023 (UTC)

PrimeHunter is correct. Further, your statement is false (take m = 20; 201, 203, 207, and 209 are all composite). - CRGreathouse (t | c) 12:34, 27 June 2023 (UTC)