Because then the character just shrugs, gives up and goes home? No. they try something else. Perhaps not regarding that particular obstacle, but the larger situation at hand.
What "something else"? How do they know? What gives them the impulse to try? Why didn't they try those other things first, rather than the thing that failed? Why is it that their failure didn't change ANYTHING WHATSOEVER about the situation--unless it did, in fact, change the situation somehow?
The answers to these questions are the very process of making it NOT "you literally just fail and nothing else happens." If the situation legitimately changed, it's not zero consequence failure. If there is time pressure, it's more than
literally exclusively "you just fail and nothing happens."
Also, don't build adventures so that there is some specific thing that must happen "in order to proceed."
Evidence suggests this advice, while good, is significantly harder to actually employ than you're accounting for.
I mean, in the very simple case, what if you prepare four methods of solving the problem, but the players completely don't even notice the first method, accidentally lock themselves out of the second method, and then roll poorly on both of the other two? This is hardly an unlikely sequence of events, especially for any game that runs long enough. This is why "fail forward" exists. It emphasizes that failure still happens and can still be costly, without permitting dead end situations whether through DM error by making a single point of failure, or by bad luck resulting in all roads being closed off or failing to pay off.
It's not the only form the failure can take. It is a form failure can sometimes take.
Oofta replied by saying he likes failure to happen sometimes ("I don't
want to always succeed"); this statement is only meaningful if it is understood to be rebutting a claim that tasks
should always succeed. But since I was responding to your claim that bare, unadorned, doesn't-change-the-situation failure is useful, and NOT saying that all possible forms of failure are a problem, the only possible way one could get my argument to mean "players should not fail" is to presume that bare, unadorned, doesn't-change-the-situation failure is the ONLY form of failure.
Hence my question: why is that the only form failure is permitted to take? I did not, in any way, at any point, imply or even vaguely suggest that players should always succeed at everything forever. I am solely taking aim at doesn't-change-the-situation failure.
I never said it was. But sometimes it's the only logical consequence. I don't want to play the game with unrealistic safety bumpers. I don't make adventures success or failure rely on a check (or even a multiple checks), just that there will be other costs. Miss the secret door? You have to fight to get past the guards or find some other way to get in.
Nothing whatsoever in my post indicated any of this, so I am deeply confused why you would respond as though it did if you were not, as noted above, assuming this. What gave you even the slightest hint of "unrealistic safety bumpers"?
Bare, unadorned, doesn't-change-the-situation failure IS boring. Period. Even in your own example you have given dramatically more than this: failure has the cost of fighting, or delaying, which may be a serious cost indeed.