6 ms·
The Lenstra–Pomerance–Wagstaff conjecture seems obvious, and it's hard to prove. What did I miss? Given that there are an infinity of primes, then it seems obv
by lutusp 14y ago
The Lenstra–Pomerance–Wagstaff conjecture seems obvious, and it's hard to prove. What did I miss?
Given that there are an infinity of primes, then it seems obvious that an infinite subset will be Mersenne primes. To dispute this, one would need to imagine a mechanism that would prevent an infinite set of primes from possessing an infinite subset of Mersenne primes.
But proving this is much harder than stating it.
- ColinWright 14y ago> What did I miss? They're asking for theorems, which means things that have been proven. > ... it seems obvious that an infinite subset > will be Mersenne primes. Not to me - I can easily picture these things petering out after a time. For example, it is suggested that there are only finitely many Fermat primes.
- lutusp 14y ago> They're asking for theorems, which means things that have been proven. I managed to miss that. My bad.
- ColinWright 14y agoNP. I, personally, also don't think it's obvious there are infinitely many Mersenne primes.
- lutusp 14y ago> I, personally, also don't think it's obvious there are infinitely many Mersenne primes. If I were a better mathematician, I might agree with you. But I can't think of a compelling reason why some percentage of the infinity of primes would not coincide with the infinity of members of (2^n)-1 , 1 <= n <= ∞, thereby creating an infinite set.
- ColinWright 14y agoThen why not an infinite intersection between the primes and the numbers of the form 2^(2^n)-1? And yet we only know 5 of them. Let me help your intuition. Take these numbers: k(1) = 0 k(2) = 2 k(3) = 3 k(n) = k(n-2)+k(n-3) For a given n, ask if n divides k(n) You'll find up to 10^5 or so this works if and only if n is prime. Does it work forever? Consider the primes of the form 4k+1. For the first 10^5 or so of these they turn out to the be sum of two squares. Does it work forever? Heuristic arguments are all well and good, but at the end of the day, these things need to be proven or disproven before we can be sure.
- lutusp 14y ago> ... at the end of the day, these things need to be proven or disproven before we can be sure. Yes, and in the meantime, we're free to conjecture -- within reason.