I don’t follow. Combining multiple sources of deflective protection reduces and eventually overcomes the disparity. Why can’t that be the case?
So what are shields and cover worth then? There has to be a number, right? If a barkskinned druid doesn't see any AC boost from a shield by itself (implying its bonus is less than +1, because that's the only way a barkskinned druid can wield a shield but gain not extra AC from it), nor any from cover (also implying its less than +1 while barkskin is active)... what are they worth?
They HAVE to be worth something in AC value, right? Because you are telling us that when you combine the two together they at some point boost the effect of the barkskin. The barkskin gives an AC 16... shield and cover together produce an AC of 17. Which means that either barkskin shuts itself off when you pick the shield and go behind cover (and the entire +7 AC bonus come from those two sources)... or the barkskin stays in effect to give the AC of 16 to start with, and the druid then gains the extra +1 bonus from the shield and the cover.
So what are they worth? +0.5 AC each? Or are we doing in proportions? The shield is worth +2/7ths a point of AC and the cover +5/7ths of an AC point?
Oh, but wait...
This other naked druid over here has a 16 Dexterity, normally a +3 bonus to AC. But with the barkskin on, obviously they're too slow to gain any bonus from the Dex, just like someone in chainmail is. The druid's AC is 16 from the spell. That makes perfect sense. And when they pick up a shield for what should be an extra +2, that doesn't help! A shield doesn't speed them up, why would it? So of course their AC is still 16 from the spell.
But hold on... this druid now goes behind the wall with the arrow slit... and their AC JUMPS to AC 20?!? Wait, what? How?!? Where did the druid get these +4 extra points of AC from over from what they get from the barkskin? I mean... their Dexterity obviously didn't suddenly start to actually now work... that would be ridiculous! Did the shield suddenly start learning how to block attacks and two of those points now come from the shield? That seems unlikely! Where did we get the +4 bonus AC points from? What happened within the game world to accomplish this?
Wait... did the 3/4ths cover of the arrow-slitted wall decide on its own to arbitrarily drop its AC bonus from a +5 down to a +4 in order to reach that AC 20? Can it do that? Does the wall get to decide that? How does it know what bonus to give these druids?
That's why what I said above was... "There's no other situation in the game where AC bonuses don't grant their bonuses normally and instead you just make up some number that they do give. " The game has made up this scenario where 3/4ths cover grants a +4 instead of a +5. And guess what? If by some chance that druid's Dexterity modifier was only a +2, that means going behind the wall would only give the druid an AC value of 19. And that means when you take the AC of 16 from barkskin into account... the cover is now only granting 3 points of bonus AC. 3/4ths cover is arbitrarily granting +5, +4 or +3 bonus points of AC to druids, apparently on a whim, depending on the druid's DEX modifier.
And this is why the fluff of the spell and the way it works in the game world make absolutely no sense. And attempts at trying to justify its mechanics are fruitless and a fool's errand. Because after all, you had to eventually say this:
The difficulty of hitting the Druid is represented by their normal AC value. The difficulty of penetrating the bark is represented by the 16 from barkskin. If the difficulty of hitting the Druid is less than the difficulty of penetrating the barkskin, then you use the 16, because any attack forceful enough to break through is also accurate enough to hit. If the difficulty of hitting the Druid is greater than the difficulty of penetrating the barkskin, then you ignore the barkskin, because any attack accurate enough to hit is also forceful enough to break through.
You are trying to narrate the type of hit that happened after the hit occurs... based entirely on what the druid's AC was at the time. "Well, the druid might have had a +2 DEX bonus and a shield, but because they had barkskin on... the hit obviously wasn't dodged or blocked, it just penetrate the druid's skin!" Versus "The druid might have had barkskin on, but that didn't matter because their DEX was +3, they had a shield for +2, and they were behind half-cover for +2, so obviously the hit was just so accurate nothing could stop it, not even the barky skin!"
Really? The shot was so accurate that the skin protection didn't matter? And what was the attack roll total? A 17! Hmm. And what was the AC to hit? 17! Really? THAT was the "super-accurate" attack? The one that just barely hit? I see. Yes. Yes, that is completely understandable. Makes perfect sense. How could I possibly have thought the narrative of barkskin didn't align to the mechanics?