Why Does this Algorithm Generate Primes?

แชร์
ฝัง
  • เผยแพร่เมื่อ 31 พ.ค. 2024
  • The recurrence R(n) = R(n - 1) + gcd(n, R(n - 1)) generates primes. But why? It turns out it's essentially implementing trial division in disguise.
    Previous video on this sequence: • In 2003 We Discovered ...
    ----------------
    Reference:
    Eric Rowland, A natural prime-generating recurrence, Journal of Integer Sequences 11 (2008) 08.2.8 (13 pages).
    cs.uwaterloo.ca/journals/JIS/...
    ----------------
    0:00 Generating primes
    1:26 Shortcut
    4:42 What if 2 n - 1 is prime?
    9:31 What if 2 n - 1 isn't prime?
    14:46 Trial division
    ----------------
    Animated with Manim. www.manim.community
    Music by Callistio.
    Audio recorded at the Lawrence Herbert School of Communication at Hofstra University. www.hofstra.edu/communication/
    Web site: ericrowland.github.io
    Twitter: / ericrowland

ความคิดเห็น • 40

  • @RSLT
    @RSLT 3 หลายเดือนก่อน +29

    Nice to see a new video after quite some time.

    • @EricRowland
      @EricRowland  3 หลายเดือนก่อน +10

      Thanks! This one almost killed me 😂

  • @tcaDNAp
    @tcaDNAp 3 หลายเดือนก่อน +4

    I'm sure it's equally cathartic for the viewers and Eric to end without a cliffhanger this time! I never would've learned the meaning of this paper without the animations and explanation 👏

    • @tcaDNAp
      @tcaDNAp 3 หลายเดือนก่อน

      One of the only math research papers I have read was in the back of the Prime Suspects graphic novel, so thank you to all artists connecting math and media!

  • @_Wombat
    @_Wombat 3 หลายเดือนก่อน +2

    This style of video and explanation is really good. I appreciate how you constantly pause to run an example rather than always talking in terms of n, p and i. My brain needs examples to understand the algebra.

    • @EricRowland
      @EricRowland  3 หลายเดือนก่อน

      Thanks! I agree… Examples are essential!

  • @achrafidou537
    @achrafidou537 3 หลายเดือนก่อน +7

    It's how unlikely that i rewatched the first part yesterday and now i find this

  • @gabitheancient7664
    @gabitheancient7664 3 หลายเดือนก่อน +2

    oh my god I love how elementary this all is, must have been really satisfying to figure this all out

  • @emanuellandeholm5657
    @emanuellandeholm5657 3 หลายเดือนก่อน +6

    You made me wait an entire year. Totally worth it!

  • @mebamme
    @mebamme 3 หลายเดือนก่อน +9

    Dangit, right before I meant to call it a day. :D This will be the first video I'll watch tomorrow!

  • @lynxfl
    @lynxfl 3 หลายเดือนก่อน +13

    He's back!

  • @burnstjamp
    @burnstjamp 3 หลายเดือนก่อน +2

    Mind-boggling. The information here is presented beautifully!

  • @firozabegum4373
    @firozabegum4373 2 หลายเดือนก่อน

    Primes are my favourite. This video is really really great, I like it! Waiting for the next one.

  • @Ryloon
    @Ryloon 3 หลายเดือนก่อน

    Im so happy! Thank you for uploading

  • @gabitheancient7664
    @gabitheancient7664 3 หลายเดือนก่อน

    oh damn I need to rewatch the previous video but damn great this video came out

  • @johnchessant3012
    @johnchessant3012 3 หลายเดือนก่อน +1

    Great proof!

  • @sccur
    @sccur 3 หลายเดือนก่อน +1

    I am probably not understanding something, but it seems obvious to me that this sequence would generate primes in this manner having GCD as one of the operations and the rest basic arithmetic. And you can probably make a million different formulas with GCD that will have patterns generating primes. I am sure I just don't understand because I'm just finishing calculus, but what makes this interesting?

    • @mebamme
      @mebamme 3 หลายเดือนก่อน +1

      You can sometimes get composite numbers if you start the sequence with a number other than 7. (the previous video explains it a little more.)

    • @keagangrahamcallis7857
      @keagangrahamcallis7857 3 หลายเดือนก่อน +2

      You're using smaller primes in a neat way to find bigger primes.
      Which is kinda what you always do; like q is a prime if all primes less than q don't divide into it.
      But that standard way requires you to know all the primes less than q. This way doesn't.
      I think...

  • @debmalyalodh1
    @debmalyalodh1 2 หลายเดือนก่อน

    Welcome back

  • @johndoyle2347
    @johndoyle2347 3 หลายเดือนก่อน

    Willans' Formula for primes:
    2 to the n part = vertical asymptote and p-adic numbers. 1/n part = vertical tangent. Factorial part = vertical line. These tensors from differential calculus determine singularities in stable matter as represented as primes.

    • @drdca8263
      @drdca8263 3 หลายเดือนก่อน +2

      Sorry? This isn’t particularly clear

    • @johndoyle2347
      @johndoyle2347 3 หลายเดือนก่อน

      @@drdca8263 Yes. I am referring to "functions" in differential calculus that are continuous, yet not differentiable at points. There are 5 cases: a corner/cusp, which fits with dark matter singularities. A ring/cylinder/horn, which fits with singularities in baryonic matter. A vertical asymptote, a vertical tangent, and a vertical line, which are tensors that are involved in both keeping matter stable and are involved in Big Bounce events.

    • @drdca8263
      @drdca8263 3 หลายเดือนก่อน

      @@johndoyle2347 Vertical asymptotes aren’t continuous (unless, I guess, if you compactify the codomain?). They also are not tensors.

  • @DOTvCROSS
    @DOTvCROSS 3 หลายเดือนก่อน

    @11:49 the reason for 'circular logic'
    2n-1=2n-1
    2n-1=(3-1)n-1+(i-i)
    2n-1=3n-n-1+i-i
    2n-1=3n+i-1-n-i
    2n-1=(3n+i-1)-(n+i)
    Basic Algebra trick of adding and subtracting. Then put LHS and RHS into the same function, of course it is equal. Don't get lost in basic algebra.
    i is an ~'eigenvalue'~ on a 2n-1 plane maybe 'parameter' is a better word.

    • @disonaroaurelo
      @disonaroaurelo 3 หลายเดือนก่อน

      Very simple but very useful content in number theory.

  • @disonaroaurelo
    @disonaroaurelo 3 หลายเดือนก่อน

    I have a technique to obtain semi-random primes but in an increasing manner. Which is multiplying a number by 6.67-> 0.67 for 1 to 10; 6.7 from 10 to 100; 6.67 for 100 to 999; 6.667 to 1000 to 9999.The problem is that the numbers grow and you have to factor them in a common way. '--' Multiply only by multiples of 3. And it's okay, they come kind of randomly but increasingly, and they all come. My sieve is thick, it's a bad thing. Take everything, just the bulk! Skdkdndjndndj

  • @hylens5111
    @hylens5111 3 หลายเดือนก่อน

    Did I miss it? Why does the sequence start with 7?

    • @EricRowland
      @EricRowland  3 หลายเดือนก่อน +2

      No great reason to start with 7, other than it's not too small. If you start with a number other than 7, you get similar behavior. I explored this a little in my other video on the topic: th-cam.com/video/OpaKpzMFOpg/w-d-xo.html

    • @hylens5111
      @hylens5111 3 หลายเดือนก่อน

      Thank you!@@EricRowland

  • @kaininjago6161
    @kaininjago6161 3 หลายเดือนก่อน

    YAY!

  • @DeathSugar
    @DeathSugar 3 หลายเดือนก่อน

    I wonder if this could be displayed as some kind of L function

  • @pizzarickk333
    @pizzarickk333 3 หลายเดือนก่อน

    Finally

  • @SpinnyDisk
    @SpinnyDisk 2 หลายเดือนก่อน

    Manim! (Or whatever it's called)!

  • @kristofferpaulssonmisc2195
    @kristofferpaulssonmisc2195 3 หลายเดือนก่อน

    Could you write an example program in Java using normal integers and BigInteger class?

  • @_eagle_299
    @_eagle_299 3 หลายเดือนก่อน

    DAMN THIS MUSIC IS SO FIREEEE

  • @j.21
    @j.21 3 หลายเดือนก่อน +1

    a

  • @marti7716
    @marti7716 2 หลายเดือนก่อน

    *promosm*