@Enrahim I'm glad what I had to say has been helpful to you, though I didn't have a lot to respond to overall in the post. So I'm focusing in on specific bits. Not because I'm ignoring them, but because I felt these were the ones that would get more done, rather than "mhm" or "ehh..." or the like.
People signing up for it expect failure to only indicate task failure
I am not nearly so convinced of this. I don't think most players realize what they're signing up for
at all with D&D, beyond "a fantasy game where I get to play a cool character".
I do think that the very hardcore-sim fans
do have a very specific idea of what they're signing up for. But for the vast majority of people, if you mentioned anything at all--even in the most layman-friendly terms possible--about what they're expecting in terms of "task failure" vs other things, they'd stare at you blankly and ask, "Uh...what does
that mean?"
Or, another way to put this: fresh players genuinely have no idea. Non-fresh, but also non-hardcore, have not thought about it, they just know what D&D does. That leads to a significant group of people who think D&D's way is the
only way, and thus anything that isn't D&D's way is "wrong" somehow because it conflicts with a completely unstated assumption that they've never actually thought about and lack any useful vocabulary for describing. D&D teaches people to do things its very specific way, which is not only not generic, it is pretty dang specific, and there are a lot of other approaches which are either "pretty dang specific" in a different direction, or
actually more general, as in, they don't narrowly require everything to be this one model of what success and failure are, they embrace a variety of concepts for what success-and-failure mean.
This is both the blessing and the curse of D&D. People sign up for it without actually knowing what they're signing up for, and those who do and then stick around often do not realize just how narrow D&D's structures can be. It's a gateway, but it's also really narrowly specific. Like how
Dragon Ball Z was the gateway anime for a lot of Americans--but if you think that ALL anime has to be mostly-comedic vaguely-martial-arts-inspired and drawing heavily on
Journey to the West, you're going to react rather negatively to a HUGE,
HUGE swathe of what "anime" is, because you're coming at it with the mistaken belief that all anime should be like DBZ.
And I think that also shows one of the really weird social...quirks, let's say, of how TTRPGs work within the minds of people. For some reason, folks instantly grok that just because you know one band of a particular style of music, or one work in a particular genre of animation, or one author in a particular branch of genre fiction (e.g. murder mysteries)....doesn't mean you know everything! That you should in fact expect
very different things from different creators, and that you often can't get a full understanding of what it means for something to be "a mystery" or "a glam metal song" or "a romance film" or "an anime" etc., etc. just by consuming work from one singular source.
For some reason, exactly the opposite happens with TTRPGs. People get in via D&D...and then they think everything must use the same concepts and philosophies and structures as D&D, and anything which doesn't is somehow flawed or wrong or "niche" etc., when it very much could be
D&D which is the "niche" thing, it just by total accident happened to be the one that took off first. (In that sense, I am reminded of the rise of Thebes as a power in ancient Greece. They basically bumbled their way, buttocks-first, doing
stupid pointless things, into being one of the strongest city-states in Ancient Greece for over a century, and remained relevant for decades after their heyday. Proof that just because somebody got big, doesn't mean they must have been clearly right and smart and well-prepared. Sometimes, the accidents and outright
follies of history really do shape the course of empires.)
That leaves the question - what if trad try to introduce FF for their group to see if there could be consensus for trying it out? Me myself don't see any principle problem with it in "trad" in general. I still think it is fully incompatible with a game advertised as a living world. Me personally would not want to play with it in anything but one shots or short adventures, but that is due to my prefered pacing of a campaign (more contemplative and deliberate), not due to any more fundamental problems with the concept.
I don't see any problem with it, but I also don't see it as incompatible with a game advertised "as a living world". That is, I don't think a "living world" needs to be one where resolution mechanics are always narrowly laser-focused on the tiniest units of action one can get away with without absurdity (e.g. we don't make someone roll for every step along a narrow ledge, we treat the whole narrow ledge as one chunk even though we
could divide it up). Which, for me, brings up a very important point here: it's not a binary.
That is, it's not like the D&D-alike approach, which you have called "task failure", cares nothing at all about intent, manner, or wider context. We
do in fact take some degree of flexibility of scope, even in hard-sim, ultra-"trad" D&D. Even the most simmy of sim fans recognize the utility of concepts like "let it ride", assuming they have informed themselves on what it is for, because they understand that the nature of iterative probability makes it terribly punishing to ask for excessive numbers of rolls. And, as your previous example with the solid wall with crumbly handholds has given, it's also not the case that the D&D-alike approach--"task failure"--doesn't invoke revealing a new key detail about the world
because of failure, either. It does, generally, tend to favor keeping any such reveals "local" (closer to the direct action is much preferred over farther from it), "small" (fine details or easily-overlooked things, rather than broad details that should be easily found even by casual observation), and "narrow" (range of applicability in space, time, amount, and/or persons should be reduced when possible)--but it isn't totally against a broadening of scope or a loosening of direct connection under circumstances that can happen with reasonable frequency in actual games.
Basically, by calling it "task failure" as separate from whatever you would call the alternative, this implies they're two genuinely distinct things...and I don't think they are. Instead, there's a spectrum of specificity. Most folks agree that allowing specificity to fall to zero (infinitely wide scope) would be bad, as that would make it pretty hard for anyone, player or GM alike, to really grapple with it. Conversely, as my "we don't make players roll for every single
step to cross a narrow ledge" example hopefully shows, we
also don't want the inverse, driving specificity to infinity (and thus infinitely
narrow scope), because we know that that leads to problems.
I know I make a lot of video game analogies and these probably aren't desirable for a lot of players, but I'm
really REALLY strongly reminded of games like "Surgeon Simulator" or "QWOP", as compared to games actually meant to
feel like you're doing surgery or running around places. QWOP has narrowed the scope almost infinitely far: you have to independently control not only each leg, but the
thighs and
calves of each leg. This is, in a certain sense, objectively more "realistic" than any other video game that involves running, because the
real physical process of running involves precisely controlling your calves and thighs to physically move your body. However, in a very important sense, it is dramatically
less realistic than most games that involve a running person (e.g. something like
Mirror's Edge), because, even though this is objectively more like the physical motion of running, QWOP's mechanics feel so blatantly unnatural and "wrong" and make the whole task
enormously difficult.
My point with this analogy is that
both your "task failure" approach, and whatever you would contrast it with, cannot possibly be at the extreme ends of the scale we're comparing them on. By necessity, they are closer to one another than they are to either extreme. Hence, there is no bright-line distinction between the two of them. It's a matter of finding the scope-range you like and the degree of scope-flexibility that you're comfortable employing. Having played (and quite enjoyed)
Ironsworn, a game that is...PbtA-adjacent, one might say, I am supremely confident that fail forward can work perfectly well in a "living world" game.
The one thing you
would have to accept, though, is that the GM doesn't actually know functionally all facts in advance of the players knowing them. That's the one key difference. In the approach you describe, the GM
functionally always knows every fact about the world at least some amount of time before the players do, and usually knows any given fact about the world
well in advance of the players. Now, if the players do something that changes a publically-known fact, then of course that's something the GM learns at the same time as the players do, but even in many cases of player action, the players won't know that a particular fact has changed until a good while (hours, perhaps sessions, maybe even many sessions!) after the GM does.
That very very much does not mean that the GM
can't know anything earlier. As I have been taught in this very thread, GM prep is essential and that inherently requires that the GM know things the players don't yet know. But things like the PbtA principle "Draw maps, leave blanks" are there specifically to remind the GM not to prepare
excessively. By leaving room for things nobody knows yet, not even the GM, the system ensures that it is never a situation of "the players are simply uncovering the world the GM already built and fine-tuned for them".