NFO basically was just lamplighting that there is a difference between what the player knows, and what the character knows.
Simply because I happen to know about nitrates, and their usefulness as an explosive precursor, does not mean that Wizzy McWizzface does. I fully accepted that, because it is quite apparently true.
However, it was not the argument I was railing against. The argument I was railing against was more straw-manish (as NFO does not do it anyway, but many DMs DO): Treating magic as "Fundementally unknowable", and so, denying any attempts to understand why you need a fist full of dried peas to cast a certain spell--- etc. (pathfinder example.)
If you know why you need the fist full of dried peas, you can possibly do a substitution, or possibly even revise the spell so that it is not needed at all. This alters the casting mechanic of the magic, through applied magical knowledge, which many DMs will forbid because it changes the mechanics of the game-- and also, "I dont like thinking about magic, I like swords and clobbering things-- so I invoke rule-zero on you because 'not fun!'" also rears its head.
I countered that Newton did not need to know about the Higgs field to formulate his law of universal gravitation (which was plenty good enough for several hundred years, until it was replaced by Einstien and general relativity, and then added to with quantum theory and quantum gravity-- through increasingly sophisticated and expensive experimentation)--- He just needed to do some simple experiments with weights and measures, and do some math.
Giving an explanation for why an experiment's outcome is a certain way is the DM's job, if the player wants to do the experiment. However, the results of the experiment must become just as real as fire being hot is; Something the player can then build on to learn or do other things. He and I agreed on that, as long as the process is not tortuous or joy-killing.