Reading the thread, I think there are two primary modes in which people think about D&D: First, as the rules considered specifically as a game, as well as the implications of those rules. Second, as a (sub)genre, and how well the feel of that genre can be summoned, matched, or imitated by some set of rules.
Is D&D first a game (i.e. mechanics), or a subgenre? As far as I can tell the answer is "yes", but for any given individual the more specific answer is something like {yes, yes}, {yes, no}, or {no, yes}. If you like, add a ", but" after any of those. In your view, is one usually subordinate to the other? If this were linear algebra, they might be basis vectors and could be weighted individually. If this were logic, one might be a premise, and the other a consequence.
If D&D is first a sack of some fixed (here unspecified) mechanics, then the flavor implications of those mechanics are what define D&D settings. And any settings broadly compatible with those mechanics can be D&D. It also lets people expand and flesh out settings, by allowing what the mechanics allow to become part of the setting's reality. For optimizers, this can be quite a lot. If this is the view, then the specific mechanics in the sack is important, hence Mearl's list.
If D&D is first its own subgenre, then the mostly fixed set of flavor requirements is more likely to restrict or define exactly what mechanics are in play, and how. For example:
Celebrim said:
Again, that's based on setting and adventure design. The rules don't force it on you. I frequently gamed in 1e without access to a cleric, and if you change the assumptions of the setting and the assumptions of adventure pacing you can dispense with clerics entirely in any addition should you desire.
Clearly there can be feedback, where setting informs mechanics, and mechanics informs setting, but when one of these mechanisms is much stronger than the other what counts as "D&D" can rapidly diverge depending on beholder. These feedback mechanisms might work in surprising ways. For example, I think it is likely that most D&D players identify most strongly with their first campaign as the image of D&D. If that first campaign was run by someone with mechanics first tendencies, letting the mechanics define and redefine the reality of the setting, a setting first person might actually define D&D how they do because of the mechanical implications of that first campaign. Similarly, a mechanics first person could do the same for some particular houserule in their first campaign that only exists because their setting first DM wanted to emulate some particular aspect of a non-D&D story.
As for Mearl's list, it isn't clear if it is the union of all the answers he got, the intersection (very unlikely!), a hastily agreed upon consensus in R&D, or what have you. I think union is most likely, in which case individual lists were probably as diverse as ours. Rather than nitpick every item, although I enjoyed reading others' thoughts about them, I asked myself this question: "If all of these things are present, am I possibly or likely playing D&D? If none of them are, am I possibly or likely playing D&D?" My answers are yes and no, respectively, which suggests that even if the list isn't perfect, for at least one subset of those items the answer changes. I think it is likely there are multiple such subsets for me, so that there is a "phase change" from non-D&D to (possibly) D&D that can occur at multiple places given all possible combinations of those mechanics. In that case "D&Dness" is an emergent property of the interaction of its parts, which I think favors the big tent philosophy Mearls is so clearly trying to pursue. That this could be different subsets for the same person, and clearly for different people as well, is exactly the sort of thing they should be finding out.