Is it generally preferred that empty products are gotten rid of where possible?

60 Views Asked by At

Is it generally preferred that empty products are gotten rid of where possible?

For example:

Stewart's structure theorem says that for a positive integer $n$, every positive integer $\leq n$ has a representation as a sum of distinct divisors of $n$ iff (if and only if) $n=1$, or $p_1=2$ and $p_i-1\leq\sigma(p_1^{a_1}p_2^{a_2}...p_{i-1}^{a_{i-1}})$ for every $i\in[2,\omega(n)]$, where $\omega(n)$ is the number of distinct prime factors of $n$ and $\sigma(n)$ is the sum of divisors of $n$. Such numbers are called practical numbers. However, we don't have to state $p_1=2$ if we require the special case with $i=1$ to hold, from which $$p_1-1\leq\sigma(\prod_{j=1}^{0}p_j^{a_j})=\sigma(1)=1,$$ thus $p_1=2$, since $1$ is not prime.

But there are other ways to get rid of the empty product than by stating $p_1=2$ explicitly; Multiplying both sides by $\sigma(p_i^{a_i})$ results in

$$p_i^{a_i+1}-1\leq\sigma(\prod_{j=1}^{i}p_j^{a_j})$$

so that for $i=1,\hspace{1mm}p_1^{a_1+1}-1\leq\dfrac{p_1^{a_1+1}-1}{p_1-1}\implies p_1\leq 2\implies p_1=2$.

Another option in this case is to define a function $T_i(n)$ which returns these "truncations" of the prime factorization of $n$ (that appear in the divisor sum) so that $T_0(n)=1$ and $T_{\omega(n)}(n)=n$. In order to avoid having an empty product in the definition of this function we might define the functions $p_i(n)$ to return $1$ and $a_i(n)$ to return $0$ whenever $i\not\in[1,\omega(n)]$, then let $$T_i(n)=\prod_{j=0}^i p_j(n)^{a_j(n)}$$

1

There are 1 best solutions below

4
On BEST ANSWER

As long as you're dealing with a numerical product (or even an operation in a semigroup with identity), defining an empty product to be the identity works out extremely nicely. If you're dealing with more general/weirder sorts of "products" (e.g., Cartesian products, topological products, etc.), you end up having to give up equality and settle for isomorphism in general (e.g., $A\times(B\times C)\cong(A\times B)\times C$ but equality does not hold), so the fact that there's no real identity doesn't seem like such a big deal.

The reason the empty product being the multiplicative identity makes sense is, very simply, that it works nicely with the general rule that $$\prod_{i\in I_1}x_i\cdot \prod_{i\in I_2}x_i=\prod_{i\in I_1\sqcup I_2}x_i.$$