Ron Edwards spiega la differenza tra task resolution e conflict resolution

Imagine that you are playing a game, a very ordinary RPG, nothing special or fancy about it. You have a character and this character has a big sword. There is a wooden pole set in the ground in front of the character. You say that your character swings his sword to cut that pole. We will assume for purposes of this explanation that this is, in-universe, physically possible for the character to do. 
Now we split into two possible approaches to this situation. 
Always, forever, no matter what, you must roll or apply a given game mechanic of any kind to see whether the character cuts the pole successfully. Or perhaps this "always" exists in vague range of difficulty as judged by the GM, but still, inside that range, the mechanic must be applied. 
You consider whether cutting this pole is opposed by any other character or character-equivalent. Does any current action or pre-established action (even in the GM's notes) act as an opponent to this action? Does the timing of this action play a relevant part in some other circumstances in play? If yes, then you apply the relevant mechanic. If no, then you don't, and the character simply cuts the pole. 
The first is task resolution. Tasks are resolved (mechanically) in the presence or absence of conflicts. A given conflict is resolved only if the cumulative tasks involved finally and eventually add up to its resolution, and this is an opportunistic outcome. 
The second is conflict resolution. Tasks are only mechanically interesting (i.e. mechanically resolved) insofar as they are relevant to an existing conflict of interest, and that conflict of interest will be resolved through the application of the mechanics, no matter what tasks are or are not involved inside it. 
The difference is enormous. It is not trivial, and there is no spectrum between these approaches to play and to rules. This is a binary and real distinction that applies to any role-playing rules ever written and played. All of the above statements that "it makes no difference" or that "once you learn it you can forget it" are false. 
It has nothing to do with the scale of the resolution, i.e., how much time or effort or how many tasks are involved. It has nothing to do with whether the potential results are pre-stated ("stakes") or emergent. 
Please think about Dogs in the Vineyard. Once the dice begin to roll, whatever problem or confrontation was in play will come to some kind of relevant outcome, and that problem or confrontation will at the very least undergo a profound change. This is what the dice are for. The various actions inside that conflict are not trivial, but they are not considered units of mechanics usage. Or to put it a little differently, this is why you do not roll dice in Dogs in the Vineyard when your character decides to fire a bullet into a tree for fun when he rides by it. 
Many people have arrived at conflict resolution in practice when using task resolution rules, simply because they informally avoid using the system unless a conflict of interest is involved. This is a serious change from the written rules to the in-play (real) rules, for that group. The fact that they do not realize they are doing it, or if they do, they merely call it "playing right," does not mean the change is not real.