And see, you devolve into English nipicking again with 'increase.'
What you see as "nitpicking", I see as the stated method of the feat. There's a mechanical difference between an effect that grants a bonus, and one that just leads to some kind of overall increase in AC. And "devolve" is likewise inaccurate, since I've held that as a constant standard and it's an important one.
From a direct and practical point of view, "granting" the bonus also allocates direct responsibility for the bonus to the spell's magic, making it logical for the feat to overcome it, compared to any kind of increase, which could result from incidental nonmagical side effects.
PMP asks for a bonus. A bonus naturally means an increase. You're treating it as a static object, a thing, instead of an effect. For some reason, 'bonus' is now an item to you, instead of a shift in a number...which is completely erroneous. Bonus to AC means an increase in armor class, which cat's grace and mage armor both provide in their own way. You're also ignoring the fact that trying to say a modifier to a dex bonus is not the same as a modifier to an AC bonus...your own logic is not transitive, and doesn't follow. It never has. It consistently does nothing to AC like mage armor does nothing to AC...nothing follows from either of those.
A bonus means a positive modifier is being applied, but there's no guarantee that such a modifier will be effective or useful in increasing overall AC, which is why there's a distinction between the two.
And for the Dexterity bonus (we'll assume it's a positive modifier, for simplicity), there's the standard fact that you have such an ability score bonus simply from having a Dexterity score of 12 or higher, and then there's the capability of applying that ability score bonus as a bonus to AC, which is a separate effect from the ability score bonus itself. You can, for example, use the Deepwarden's stone warden ability to divorce your Dex score modifier from your AC, but you still have the ability score modifier in question available to use for all other purposes, such as Reflex saves, Dex-based skills, and initiative.
The difference with mage armour is that the effect (both the field of force, and the +4 bonus it provides) is entirely within the spell's domain of influence, whereas cat's grace grants an improvement to the Dexterity score, with the exact effects of the Dexterity score existing beyond the spell's influence, and consequently not being a part of the spell. Mage armour cannot be subdivided like that, because it's all part of the spell.
Your dispel magic example is conveniently ignoring the limitation of targets and Area of effect. A Dispel Magic will not affect a spell it is not targeting, nor one outside its area of effect. By the game rules, it simply has no power to do so.
By the game rules, having even single target of a multi-target effect within the dispel's range means that you can access the spell for the purpose of attacking and ending its magic.
It's no different, in logical terms, to killing a Colossal dragon with a fireball when only half the spaces occupied by its body are in range; you don't need to be able to reach its entirety to affect the creature (or spell) as a coherent whole, and the parts of its body outside of range still end up just as dead.
What you are saying is that Dispel Magic can now affect things it is NOT targeting, and spells OUTSIDE its area of effect. That is clearly outside the rules, and thus, not true. There are spells that are cakes, where you can take them out...we call them emanations. THere are spells that are spun out and become their own little spells...we call them targeted spells. There is no language whatsoever to say they are still 'linked' and 'one-sourced'. There is plenty of evidence to suggest that targeted spells with durations have absolutely no 'link' to one another whatsoever, and you can eliminate one targeted effect while having no effect on other independent targets.
You're attacking the spell itself, not each individual target that might be receiving its benefit. Some of the targets might be beyond the range of the dispelling effect, but as long as one of the targets is within reach, the spell is within reach, and you're able to end that spell, along with all its effects.
One spell can be universally ended by one successful dispel attempt. What is the evidence you have that these targets are not all still subject to one spell?
I'm not seeing where that would dispel a wall of sand. We've proved repeatedly that the spell isn't providing the bonus to AC, your position, and the position of who you're taking cover from, is providing that. So it's a non-issue.
If what you're saying is that damaging the mage who cast Mass Cat's Grace dispels that Mass Cat's Grace on everyone...well, I find that pretty amusing if you want to apply it that broadly, but you know, it has no effect on my argument on what is and is not eligible. A Wall of Sand is never providing the AC bonus, your position is, and is thus immune to being a dispel target. You ignore the situation (where attacker and defender are) and look at the spell. Wall of Sand never provides a cover bonus unless there's someone to provide cover against...that's the very definition of a situational modifier which PMP is going to ignore, and in this case, isn't even magical.
And the comparison to cat's grace is what this is all about. I agree that a wall of sand isn't affected by Pierce Magical Protection, and the reason is the same as for cat's grace, since the spell is never providing the bonus to AC.
Dexterity providing a bonus to AC is likewise part of a situation beyond the scope of the spell in question; it derives from non-spell functions of the ability score, and is subject to change by external effects (like the stone warden ability) that have nothing to do with the spell.
Putting aside the issue of cat's grace for a moment, from a mechanical point of view, if you had a multi-target version of mage armour, dispelling the spell on a single target would likewise end the spell effects for all targets, since the spell supporting all those spell effects has ended. You could write a different version of Pierce Magical Protection that didn't work this way, but its current text affects both spells and spell effects.
Your explanation: "My easiest proof is also very simple: Check if there's a bonus to AC in the text which isn't attributed to an external fact (like the operation of Dexterity or cover). Any spell that describes a bonus is a target for Pierce Magical Protection."
Is not simple at all, because 'external fact' can be argued endlessly, and Dexterity is never external to a target, it's as intrinsic as Natural Armor class.
It's not external to the target creature, but it is external to the spell and feat, which is the important fact for my position.
One of the differences between our viewpoints on that is that you care about the effect on the AC of the creature in question, while I only care about the direct interaction of the spell and feat. The fact that Dexterity is part of the creature means that it's not part of the spell.
Your explanation leaves people wildly clawing for what is external and what is not, because you don't have an ironclad, simple definition, and trying to defend it from both directions is going to be the result. It doesn't work at all. You keep attempting to throw in a cover bonus as part of my definition, as I repeatedly point out that it has nothing to do with the spell...you're the one who has to go in there and remove it with a judgement call, and weigh each and every spell.
I'm not throwing in the cover bonus as anything except an example of why the cat's grace logic doesn't work.
See, again we're talking about the internal/external division on a different line of reasoning; the normal function of Dexterity as an ability score and cover from solid objects are both external to the spell-feat interaction.
I just have to look at the end results. You're making mine complex, when yours requires the balancing and explanation for each and every spell that PMP might affect. I just have to look at what would be suppressed or dispelled, and if it affects AC magically.
I just have to look at the spell text. You're making mine complex, when yours requires the balancing and explanation for each and every side-effect of a spell that Pierce Magical Protection might affect, to tell whether or not they're separate.
We do, however, share a common focus on whether or not the spell affects AC magically, we just have different ideas about how to determine that level of effect.
I think it is very clear which of ours is simpler and more logical, and easier to address. Mine's also the one that makes PMP worth taking. It seems to me that your version is seeking to exploit vague wording to neuter the feat, which is word-twisting, and I'm simply taking it on its face and applying it as is, rather then trying to invent new definitions and mechanics which are not part of the feat.
I can assure you that mine is simple, logical, and easy to address.
As to whether or not the feat is worth taking, it seems to me that your version is word-twisting designed to make the feat do what you think it should do, instead of taking it on its face and applying it as is . Definitions and mechanics like "overall increase to AC" are not part of the feat, while a granted bonus to AC most certainly is. It's not good as a feat, but that can be fixed once you accept and realize its problems.
Really, man, there's nothing in the feat dealing with 'spells that mention AC bonus', 'external fact', and what not.
It's spells that grant a bonus to AC. And you trying to interject middle ground disqualiers into that just doesn't fly.
Mentioning an AC bonus is a primary method for telling that the spell is granting a bonus to AC. Really, man, there's nothing in the feat dealing with "increases to AC", "overall boost to AC", and what not.
We both support the limitation on external effects, albeit in different ways; you wouldn't include things like cover, because you recognize it as external to the spell's target, while I also wouldn't ability score results, because those are likewise external to the spell and feat.
It's spells that grant a bonus to AC. You trying to qualify into that effect some spells that don't grant a bonus to AC is what doesn't fly with my position.