Before I understood this formula
When I was a high school student, it’s hard for me to imagine a product whose index looped through all prime numbers because primes don’t appear in a regular way: between 1 and 100, there’re 25 primes, but between 900 and 1000, there’re 14.
Even though it’s easier to imagine the infinite sum whose $n$-th term is $n^{-s}$, without learning the $p$-Test and the Comparison Test for the convergence of infinite sums, I couldn’t understand why the infinite sum in the following equality exists.
A heuristic way to understand it
- Note that the geometric series converge (absolutely).
- Borrow the proofs about the Comparison Test and the $p$-Test to convince yourself that the infinite sum on the LHS of the formula is well-defined.
- Know something about convergent infinite products.1
- Convince yourself that the infinite product on the RHS exists.2
- Recall the Fundamental Theorem of Arithmetic.
I think that the last item is the trickiest step. Writing the following lines, I understood this equation.
Steps \eqref{step1} (resp. \eqref{step3}) holds because for each $k^{-s}$ in the leftmost bracket, powers of 2 (resp. 3) can be taken out from $k$. In steps \eqref{step2} and \eqref{step4}, the formula for the sum of geometric series is applied.
Refining the above thoughts
I’ll end this post by wrapping up the above ideas by summation and product signs.