I'll add my bit.
Before 4e came out I was a big fan of the unified character/NPC creation in 3.5, even though it took a lot of time as DM. I enjoyed the character building game on its own, and the unified method really helped me engage with the setting.
When 4e came out, imagine my surprise as I eventually grew to really appreciate the ease of creating monsters, enough to prefer it despite deeply wanting to like the unified method more.
For me, therefore, 5e should help the DM accomplish what they want in the time that they are willing or able to spend while sacrificing as little world consistency as possible. Therefore, like others I'd like to see quick guidelines that get one in the right ballpark fast, a slightly more detailed version akin to 4e, and a full-blown build-from-scratch system.
I would step back from 4e a little bit because of how wonderful 3.5 made building creatures feel to me when I didn't hate how long it took. I don't think basic consistency must be sacrificed, however. Instead, what should be lost is detail. A reasonable test might be if one could in principle swap out a monster for a PC and keep the party trucking without rule contradictions or undefined/inapplicable terms. So monsters use the same health system as PCs, NPC wizards (whether carefully detailed or not) cast the same spells as NPCs, and so on. They just do these things with far fewer knobs, unless the knobs are desired.
I think of it as monsters and NPCs coming into focus: in the game world, the only "confirmed" knowledge the PCs have of an adversary is what they can see or hear for themselves. In other words, the quick and dirty methods don't necessarily define a creature's abilities, they lay a boundary around our ignorance. I think this is a rather more empowering notion for monsters. For example, suppose an NPC wizard is created on the fly and given 3 spells. The assumption isn't that he knows 3 spells, it's that only those 3 spells were relevant in this instance. Later on this hastily-made NPC might makes a more fleshed-out appearance, but only if the DM wants that. And if 5e has different dials for PC detail, the same reasoning holds for all its settings. For example, if all PCs in some game have the lowest level of detail then in that case monsters with the correspondingly low level of detail may in fact completely specify the mechanics of that monster.
As I see it this enables the DM to make stuff up on the fly or detail monsters beforehand, license (and tools!) to change things later, all while maintaining mechanical consistency with the fundamental PC rules and conceptual consistency with the setting.
I also think there is a role for something like the very helpful minor, standard, elite, solo categories, despite their departure from my basic stance. I certainly wouldn't want to withhold those tools from other DMs, but I might repurpose them slightly. In particular, I think the defining mark of the elite or solo creature isn't its raw toughness, but in how it interacts with the action economy. (Early 4e solos were poor in this regard, but eventually MM3 moved in this direction.) I would eliminate elite and solo for humans and other mundane races, and reserve it for creatures that by their nature or very special circumstances can simply do more. Dragons, spellweavers, ettins, aboleths, beholders, hydras, etc. Those creatures might have large attacks and defenses, but what sets them apart is that they mock the idea of 1 action per round. In my opinion this makes counting them as multiple creatures when designing encounters more intuitive than in 4e, and more evenly handles questions about when a creature should be a elite/solo or just very high level.
