Er, yes? I wish that they would be externally consistent in their mechanical representation of an entity that is supposed to be internally consistent, both because it's less work and because it introduces less inconsistencies elsewhere.
I'll buy that a minion against a level 17 party can be the same as an elite against a level 7 party, but my big problem - which showed itself back when I was playing 4E on a regular basis - is what the creature's stats are when it's just by itself. Or when it's fighting a different monster. Or when it falls down a thirty foot cliff. Constantly re-defining the same creature would be a ton of work, especially as perspective changes.
And I'm pretty sure that the answer from 4E is supposed to be, "Don't worry about it". That's not an acceptable answer to me, though. The whole reason I bought the game and learned the system is so that it can answer those sorts of questions.
I think the core issue here is that you're both looking for entirely different things? I mean, internally consistent and simulationist? That's definitely not 4E, and it's not really any edition prior to 3E either. There are elements of trying to make things gel and be naturalistic in prior editions, but as Gygax himself said, that was never really the point so much as genre emulation. (The genre D&D embodies has varried over the years, but that's a different matter.)
So Third Edition, coming out when it did, tried to model internal consistency much more thouroughly than prior editions, and they did this for several reasons which make sense in context: the general attitude towards game design at the time was disdainful of 'incoherence', and the original intent of the OGL to make D20 the default system across much of the medium also meant that a consistent process sim for creating content was desirable--or so they thought--to keep things consistent between publishers with free reign.
And maybe that could have worked with tweaked circumstances... but this is the thing: 3rd Edition itself died quickly, replaced by 3.5. But even after the .5 update, the game was still a mess as far as balance went. The rules that were supposed to create consistency ended up causing massive monster HP bloat. Fighters were worse at fighting than other melee classes, and monsters were easier to take down with caster abilities that could ignore HP entirely. And of course, the OGL had the inadvertant effect of given other publishers the ability to cheaply create competitive products, some of which sharing only the basic skeletal structure of D&D.
All of this to say: it's not incorrect to value internal inconsistency in an RPG, but I do think it is... not something that D&D can or should implement. It's something that backfired spectacularly the one time it was tried from a mechanical perspective, and was more damaging than constructive financially as well (especially in creating D&D's current biggest competitor, Pathfinder.)
While technically, poor balance, simulation vs emulation, and the unreserved and permenant giveaway of intellectual property are all seperate issues, they converged in a way that makes it seem highly unlikely that D&D will ever go that route again. And I think that's where Saelorn and pemerton are baffling each other. Especially for fans of editions other than 3rd, its design is extremely tied up a series of huge blunders on WOTCs part, blunders that 4E perhaps swung too far in correcting, but not something anyone who liked 4E would want to return to.