News for hackers. Stuff that matters.
Re:Inventing an Operation to Solve x^x = y (2000) by throwaway71828 on Thursday November 28, 2019 @06:45PM (#21660907)
Here is another fun problem for you folks:

x^x^x^... (infinite power tower) = 2.

If you solved it, try the general case of

x^x^x^... = y, for some arbitrary constant y.

Re:Inventing an Operation to Solve x^x = y (2000) by np_tedious on Thursday November 28, 2019 @10:03PM (#21661674)
Seems to this series converges to 1 for the interval between 0 and 1, and nothing else converges. Is that correct?

EDIT: actually I guess it's more complicated than that. If I'm understanding the order of operations correctly, there's an alternating parity (where n is the number of exponentiations) thing going on with the values between 0 and 1

Re:Inventing an Operation to Solve x^x = y (2000) by JoshuaDavid on Thursday November 28, 2019 @11:55PM (#21662149)
I don't think this is possible for any constant y that is not between 0 and e.
Re:Inventing an Operation to Solve x^x = y (2000) by jwilk on Friday November 29, 2019 @06:30AM (#21663699)
My hypothesis:

• for 1/e ≤ y ≤ e: x = y^(1/y)

• no solutions elsewhere

Can anyone prove the above? :-)

Re:Inventing an Operation to Solve x^x = y (2000) by H8crilA on Thursday November 28, 2019 @07:25PM (#21661061)
How do you define infinite power? A limit of such series, if exists?

{x, x^x, x^x^x, ...}

Re:Inventing an Operation to Solve x^x = y (2000) by lokedhs on Thursday November 28, 2019 @08:15PM (#21661232)
That would be equal to x^y, which means that the solution is the y'th root of y?

However, don't you have to prove that the series converges? Is the approach even valid if it doesn't?

Re:Inventing an Operation to Solve x^x = y (2000) by jwilk on Friday November 29, 2019 @05:54AM (#21663580)
Note that according to your solution, x=√2 for both y=2 and y=4. But of course the sequence cannot converge to both 2 and 4 at the same time!
Re:Inventing an Operation to Solve x^x = y (2000) by a1369209993 on Thursday November 28, 2019 @10:57PM (#21661918)
Fixed point of x^(...), I guess; you don't actually need to define it properly:

  x^x^x^... = y
  x^(x^x^x^...) = x^(y)
  x = y^(1/y) # there you go