nope... success and failure can both change the state of the game without giving the desired result...
You can climb the tree and find no clue
You can tear the bar off and still not be able to open the door
You can intimidate the kobold and still not get him to tell you.
not sure what you are saying 'nope' to here. In my description of how 5e works I think I said basically exactly the same thing. You could intimidate the kobold, but the 'objective state' of the unrevealed backstory could make achieving your goal impossible. In DW this cannot happen, and that's the contrast. In 4e its more ambiguous, but I would say it is more like DW.
Just because something can go either way (uncertain) doesn't mean you get what you want...
Well, now if we're talking about a check being made, which would match in 5e with 'uncertain', then the problem here is the 'valence problem' with 5e, it doesn't really say what 'success' entails, as a general game principle. However, my reading of various specific sections of the 5e DMG makes me believe that the principle is that the PC gets what they were after when they pass a check. If no success was possible, then the conditions requiring a check are not present. So we will have to disagree, though I do agree with you that 5e doesn't really prevent GMs from 'squibbing'. DW, for example, does though, you simply cannot deny the intent of a move that comes up with 10+. The REAL DIFFERENCE though is that nothing in DW is 'impossible' (or that is rare at least) because there isn't unrevealed backstory.
an action can be in question (and need a roll) and still not be what you need to accomplish your desired goal.
This would then be one of two cases. Either A) there is unrevealed backstory which makes it impossible (thus no check by RAW) or B) there's a problem with the description of the fiction. In case A either the action itself is impossible (the rope is rotten) or the action doesn't lead to any meaningful consequence and is thus still not worthy of a check. And if this is just a 'lets see if he hurts himself' kind of thing, well, this is why I don't really play 5e, such things are time wasters IMHO and don't contribute usefully to the game as I see it.
Skip I don't play dungeon world.
I would love to talk 4e (it is my favorite edition) but now is not the time or place
I use these examples to shed light on exactly what choices 5e is making in terms of how it envisages play to proceed.
Okay, but what you think only matters when YOU DM, if not it is up to the DM in that game.
Who can talk about what someone else thinks? The 5e rules DO however seem to state that, since the outcome was not in doubt, no check is required at all. Maybe my interpretation is wrong? I don't think so, but in this case I think what I'm talking about is what is on the printed page, lol.
no check should ever be meaningless (I think we agree). CHanging the in game story though is a meaning. Changing a resistant and stubbern kobold into a cowering fearful one is such a meaning. climbing the tree is such a meaning, ripping the bar off changed things.
In the case of the kobold though I doubt that was uncertain. In the case of the bar on the door, wouldn't a knock spell deal with that as well as the magic? I mean, OK, the bar being missing is a change in game state, but I am dubious about it being worth a check. Why do I care about climbing a tree? I mean, OK, if the tree leads to some location that is interesting, maybe. If not then who cares? The character comes back down, and we go on our way.
Now the DM CAN rule an auto success or and auto fail, or they can ask for a check. Just becuse they ask for the check doesn't mean you get what you want... it just means the story the setting the character or something has 2 possible and probable future states and you roll the dice to decide.
Well, I don't think we can discuss that except in the context of an AGENDA. So in a Gygaxian sort of approach then 'climb the tree' could be adjudicated on the simple principle of "roll to see if you fall on your head and take 3d6 damage from the fall or not." That CAN be compatible with an interpretation of 5e, you can use it that way. OTOH if the agenda is more 2e-like where the goal is "tell a story concocted by the GM and let the players provide characterization." then there's not really much use for the check, as it doesn't really bear on the story. Likewise with other forms of more cooperative story telling, the tree is irrelevant and a distraction, though I guess it could become more than that.