In linear algebra, you typically use juxtaposition to represent multiplication, rather than explicitly with a symbol. For some reason, this actually seems to capture noncommutativity better than a symmetric symbol.

Another example would be string concatenation, which people often debate the merits of using a + to represent.

I guess it originated from my synthesis of the article.

The article gives an example: copy() bad, copyFrom() good. The single word "copy" doesn't imply which side is which, it is kinda "symmetric". However "copyFrom" does show the side, it is kinda "asymmetric". I just considered symbols instead of words as method names. (After all, if we are to make things shorter.. TLDR is a double joke here.)

And yes, you're correct, in mathematics, this is often not the rule. It's not even a rule in Haskell. But perhaps it could be a rule, at least for new operators?