Function application without parentheses

From Why start at x, y, z


It's sometimes OK to write a function name followed by its argument, without any parentheses. Because there's there is no function application symbol, this can look like a multiplication.

This is an example of juxtaposition means combine in the obvious way.

This is most usually done with the trigonometric functions and logarithms, e.g.

\[ \sin \theta, \ln x \]

So is it only OK to omit the parentheses for well-known functions? Or is it for any function whose name is longer than one letter? Christian Lawson-Perfect asked[1][2], and got mixed responses.

Some people interpret \(\sin(x)^2\) to mean \(\sin(x^2)\) rather than \((\sin x)^2\), even though \(f(x)^2\) normally means \((f(x))^2\).

Deyan Ginev searched for occurrences of f x in arXiv papers[3] and found about 7,000 papers containing that pattern, containing a wide range of exotic uses. In a handful of which, "f x" was indeed a function application.

"To simplify expressions we often omit parentheses writing f x = f (x) etc." [math/0201098]
Clear statement of the "f x" shorthand, from arXiv:math/0201098

References