IDK, I read the article and the original thread and this one, and I feel like a very simple cogent point being made by said article is missed or ignored or bulldozed or something:
Magic in traditional TTRPGs like D&D fails to model or evoke magic in the sources of inspiration they nominally draw from.
That's only because games need rules to play, "magic" in folk lore is literally anything the people of the time didn't understand sufficiently to explain in more cogent and detailed terms. Arthur C. Clarke isn't wrong about sufficiently advanced technology in the view of a Weyland the Smith making "magic" swords that are just exceptionally well made and hardened. Take the idea that a really Weyland did have a way to make really awesome swords, not like Vorpal Swords or sharped using sunshine awesome, but exceptionally well crafted and a century or two later we get Weyland forged magical blades.
On the basis that knowning something will happen and using math to show that thing did happen, and understanding why that thing happened are very different beasts. For example there's a great deal of things in physics that we know happen, can use math to demonstrate will continue to happen, and then when observed happen like the math says. However, when questioned about why that thing happened physicists have to shrug and say, "Don't know, just does, working on it."
Additional replies on a fiction stand point versus game stand point. I'm going to borrow
Brandon Sanderson's essays on magic. This applies as much to superpowers in comics to magic in D&D. Magic is effectively anything that is impossible by today's standards. Just to get that out of the way. I'm aware Sanderon is offering advice for prospective authors, but since a game designer is authoring rules I figure the advice is still pretty solid.
Magic can be a sliding scale from do anything we need as needed, no rules, it does what the author/game designer wants whenever we want as needed to achieve or our intended results. Howard and Tolkien are towards this end of the scale (Howard in particular). We'll call this 0% Rules. At the other end we have rules based magic where we know exactly what it can do, how much it can do it, under what circumstances, what the limits of the magic are, and what any drawbacks it might have include. In effect we know exactly the full scope of the magic and can use it to describe all possible effects with the magic. We'll call this 100% Rules. Newtonian Physics basically works like this, D&D magic is pretty darn close to this as well. In essence in a 100% Rules system the fun is playing with the buttons and seeing what happens, but you're limited to what the buttons do.
0% Rules magic systems aren't that common in fiction, even less so in games. Tolkien for example is not a full 0% Rules, although Howard is darn close. This type of magic tends to leave the reader/viewer/player in awe never knowing exactly what it can do. This is the place where Sam's astonishment at the elf rope comes in because it is magic (from our previous definition), but to an elf it's just well made rope. This also the place where Thulsa Doom turns into snake, or that evil sorcerer summons a demon, or whatever, and Conan stabs it in the face. The important part is the heroes never know what to expect, at least not fully. This is a great system where magic creates trouble for our erstwhile heroes, it is not a great system for our heroes to engage themselves with because 1) in a game the players need to know what they're doing, 2) the GM needs to be able to adjudicate the effects on the game and 3) the GM needs to either have rules they know and the players don't to be able to decide what happens, otherwise the GM just makes it up on the spot to facilitate the game's fiction at that moment.
The 100% rules more or less like real world physics. They describe every possible action/interaction, again I can't think of anything that is 100% Rules in either fiction or games. Asimov's Three Laws are the closest I can think of, he got a ton of leverage from three very simple rules. That's not the kind of magic we want though.
Most systems operate somewhere in the middle, some closer to one end of the spectrum or the other. D&D is closer to the 100% side, I'd say 70% to 80% solid rules. Its not higher because the game just lets the authors randomly add whatever they want, but once we do it works that way all of the time. FATE is close to the 0% Rule, maybe 20% to 30% rules depending on the setting and GM/player world choices, using FATE for a Sanderson setting is pushing you into the 80% to 90% range just because its Sanderson and that's what he likes. All you need to establish is Magic Can Do Things That Are Impossible as an Aspect for the game and bingo, that's your rules.
I'd also say some of the issues are functions of presentation for game play. For example the magic sword is in fiction such Cortana or Excalibur need a game function in the game to differentiate them from say that random guy's sword over there. So we have a rule in the game about what a regular old sword do, and then we have a rule about what a magic sword does. Clearly the magic sword is different (usually "better" in some way) than a regular sword, otherwise why bother having different rules? And lets be honest, folklore differentiates between Mjolnir and a regular hammer. The dwarves that forge it are supernaturally skilled smiths.
As another example from the article we seem to keep talking about: Antimagic zones in D&D. Most European folklore doesn't have a place where all magical things and doodads stop being magic temporarily for sure. But the idea is mostly a D&Dism anyways to make the game work a particular way. Its an intrinsic part of D&D.