Exactly like that, but completely different.
[MENTION=42582]pemerton[/MENTION], I think you’re nitpicking wording to force a conclusion at this point. DnD doesn’t allow “killing the orc” to be an action, but als doesn’t let “gain an army” be an action. DnD breaks things into single actions, or close enough to best abstracted like it’s a single action. Each action is resolved using the same system.
“Win the fight” is an action in some games, in other games it isn’t. Some fights will be won with one action, most won’t. That doesn’t change that actions are resolved in the same way either way. Some puzzles will be resolved in one action, others won’t.
Its the same shstem of resolution regardless.
I'm not nitpicking. I am trying to engage in serious analysis about how 5e's resolution works. I believe you are also, which means I'm a bit puzzled by your disagreement with me but am going to plough on!
Some context: when it comes to analysing the play of a RPG I'm basically a loyal pupil of Vincent Baker (I've read many but not all of his lumplely/anyway posts) and Ron Edwards (I've read many of his essays and posts but not the whole corpus which is vast!). The games I'm drawing on in making the analysis include a range of "trad" games (Classic Traveller, Runequest and its offshoots, classic D&D), mid-80s through 90s games (AD&D 2nd ed in a Vampire-heavy culture), 4e, and a variety of "indie" games (Burning Wheel, Prince Valiant which is from the late 80s but plays like it was written yesterday, Cortex+ Heroic, HeroWars/Quest, Maelstrom Storytelling, Dungeon World, various Vincent Baker games like In a Wicked Age and DitV).
If we think about the relationship between action declaration (a move made by the player at the table) and the fiction, there is no reason why "I kill the orc" can't be a legitimate action decelaration. (It is in Burning Wheel, for instance.) The fact that D&D treats it as a move in a wargame-like resolution framework - ie a possible trigger for hit point attrition - is not driven by the fiction.
But nor is it simply a case of breaking down the action into smaller steps. Doing 4 hp of damage to an orc isn't a step in the fiction; it's purely a mechanical change of state (although it is possible to narrate some colour around it, but that colour could be whatever the player or GM wants which is consistent with the orce being worn down). This contrasts with, say, Runequest where the granular resolution of combat actually does produce ascertainable changes in the fiction.
A GM who says "To enter the castle you first have to succeed at a STR/Athletics check to swim the moat, then a STR/Athletics check to scale the wall" is setting out some granular resolution stages, but this is different from D&D combat as (i) there is no mechanical subystem here, and (ii) each success correlates to something in the fiction.
A 4 skill challenge is another case intermediate between the previous two: each step correlates to something concrete in the fiction that actually changes the fictional positioning (unlike the mere colour of whatever narration accompanies hit point loss), but it also generates a mechanical change in a subsystem. Cortex+ Heroic is similar to this in its resolution; but unlike 4e it uses the same system for everything (from fighting to talking to gaining an advantage by taking the higher ground).
When people complain about skill challenges being "dice rolling exercises" they're disregarding the step where the fictional positioning changes and treating it all just as colour (or even completely ignoring the colour). When people complain about 4e combat being nothing but a tactical skirmish game, I think they are mostly ignoring that - while hp loss doesn't generate any changes in the fiction other than colour (with "bloodied" and "you're at zero" as exceptions), positioning, terrain and status effects do generate ficitonal positioning, and keywords provide the anchors for exploiting fictional positioning via subsequent power use. 5e seems to use fewer status effects in combat than 4e, and so for fictional positioning in combat seems to rely more heaviy on positioning.
But anyway, it is not the same system of resolution as the check-result-correlates-to-fictional-outcome of an ability/skill check.
(EDIT: Vincent Baker's
"clouds-and-boxes" represents the contrast in visual terms.)