Permanently Deleted

  • asaharyev [he/him]
    hexagon
    ·
    edit-2
    4 years ago

    Think about what happens if x, y, z are all equal to 1....how would you make the denominator go away without that knowledge, though?

    it may make sense just to brute force it by removing all denominators from the left side of the equation.

    sometimes number theory means doing a bunch of stupid multiplications and then simplifying.

    also, sometimes i purposefully make it painful, because discovering the solution is far more important, I am a Freirean educator...

    • Puffin [any, they/them]
      ·
      4 years ago

      I really don't see what you're saying. Like 1/2+1/3+1/6 = 1, why can't something similar happen here?

      Clearing the denominators and doing some algebra I get x3+y3+z3−5xyz-3(x2y+xy2+x2z+xz2+y2z+yz^2) = 0 . This doesn't exactly look easier... I tried a parity argument, but it doesn't work. I don't think anything like that can work since multiplied out like this we introduced (0, 0, 0) as a solution and I wrote a python program to brute force this and I found solutions such as (-1, 4, 11) if we allow for negative integers.

      • asaharyev [he/him]
        hexagon
        ·
        edit-2
        4 years ago

        A lot of it here is just trying a bunch of what we called "numericals" in college. Just plugging in numbers for x, y, z and looking for the structure when we simplified.

        Once you get to the expanded algebraic equation, which I'm pretty sure you've got right, you should be able to show that the solution isn't possible for positive integers.

        What must be true about that equation for it all to equal zero? I see a couple negatives, it may make sense to move them to the other side of the equation to find a statement equating two cubic relations. Pay close attention to the structure of the equation when you work with it. What happens with the odd numbers, prime numbers, positive and negative signs, powers, etc.

        • Puffin [any, they/them]
          ·
          4 years ago

          Ok I have to come clean, this is a bit. There do exist integer solutions, but you have to use algebraic geometry and in particular the theory of elliptic curves to solve this. The smallest solution has about 80 digits for each of x, y, and z.

          • asaharyev [he/him]
            hexagon
            ·
            4 years ago

            lol, next time i'll open the chain letter from grandma.

            this is a good bit, haven't seen it before that i can remember.