$$\gcd(N/q^\alpha,\sigma(N/q^\alpha)) = \begin{cases} \dfrac{N}{q^\alpha \sigma(q^\alpha)}, & \alpha \equiv 0 \pmod 2 \\[8pt] \dfrac{2N}{q^\alpha \sigma(q^\alpha)}, & \alpha \equiv 1 \pmod 2 \end{cases}.$$ $$\gcd(N/q^\alpha,\sigma(N/q^\alpha)) = \begin{cases} \dfrac{N}{q^\alpha \sigma(q^\alpha)}, & \alpha \equiv 0 \pmod 2, \\[8pt] \dfrac{2N}{q^\alpha \sigma(q^\alpha)}, & \alpha \equiv 1 \pmod 2. \end{cases}$$
The style of sentence-ending punctuation in the first display above strikes me as plainly incorrect by any standard, and yet it is frequently seen on MathOverflow and elsewhere on the internet, and that in the second seems reasonable to me.
If you look at something in one font and the same thing in another font and think that one looks aesthetically pleasing and the other hideous, the cause is a large number of little details of design of characters and their relative positions on the page, and it is probable that you don't notice what any of those little details are. When correction of one of those details in a posting was proposed a few years ago, one MathOverflow user said he doubts anyone would even notice the difference. I don't know that anyone has ever missed the point more completely. TeX and LaTeX were obviously designed to make attention to that kind of detail possible and easy.
I did an edit in which, among other things, I changed the first punctuation format above to the second.
Are there consensuses or policies about the appropriateness of such corrections?