You seem to have misinterpreted my comment since we agree that the "no-space-before 'foo' way" is truer to the syntax, and to repeat myself, I had said "char* foo" is the bad cue and even mentioned @onre's multiple identifiers supporting point (which you reiterate). It can be hard to see whitespace on HN comments, though unless you do the 4space indents to get a monospaced font. You can in-page search, though.

The only difference with the spacing against syntax in "x+y * z" is that there are fewer type-relevant operators (e.g. '+' does not work in a type expression). One does NOT hear people advocating for more space around '*' than '+' (though I have sometimes heard "equal space"). Your non-disagreement has not persuaded me that this is a bad way to explain the problem, although I agree I could cut down on the double negatives. :-) { Even grade school kids have some sense of precedence of '+' & '*' which is so broadly adopted that even hand calculators do it. }

"How misleading" spacing against syntax can be is admittedly more subjective. I think whitespace can really matter to human readers. There are many other examples. gcc added -Wmisleading-indentation a while back. Off-side rule languages like Python embrace things that sometimes prevent that problem. The Nim compiler was once sensitive to exactly this kind of operator spacing where it would parse "x+y * z" { space around '*' } as "(x+y)*z" in an attempt to have the compiler "Do what I mean".