Augh, I just had a whole huge post responding to a bunch of people, and then my computer ate it. For the love of...this is going to be much more brief than it was intended to be.
Yeah, it might be a good subject for a Dragon article or even some errata. Because, frankly, as I read the DMG, I got the impression that if the PCs encountered a wall at 3rd level and the same wall at 15th level, the climb DC should be increased. (The famous "chart on page 42" seems to feed into this, with the same maneuver doing more damage the higher levels the PC are -- I guess when an epic PC slides down a bannister to kick a monster in the teeth, he slides REALLY fast!)
I think it's easy to dismiss the abstraction here as not making sense, but I can think of a number of ways it does. Maybe you're just better at improvising maneuvers, maybe you're physically stronger, maybe both. Why does a longsword do more damage at higher levels (not just 4E; think about how stuff like Power Attack worked in 3E). We accept the abstraction as a function of your general prowess increasing.
Some parts of the PHB and DMG seem to use "objective" DCs (i.e, this kind of door is DC 10, this kind is DC 20) while others use "subjective" DCs ("This is a Hard DC for the party").
I think, in many cases, they are clearly called out as examples. You start with the tools (page 42 and 61) and create specific examples. That being said: the quote from page 23 talks about the importance of realism. Those static numbers are examples of how to create realism and consistency. They are not the be-all, end-all, though, and we give the ability for DMs to improvise challenges when those fixed rules don't cover it.
Why are all the threats in the city of brass epic - why are all the inhabitants epic? The premise of why you need the open lock DCs to all be higher based on tier seems to be based on the idea that everyone the PCs interact with at a given tier is also going to be at that same tier, which seems awkward.
Not everyone and everything that the PCs interact with--just the people that are meant to challenge them! It's perfectly OK to have lower-level peons in the City of Brass, but they're not going to be as much of a challenge. The idea of putting epic threats in the City of Brass is to present a challenge to the heroes. If you want to present a lesser challenge, or none at all, then by all means do so. However, it's just as easy to assume that the City of Brass has epic locks on the doors to keep out the epic thieves that roam the city, not just the PCs.
I know that if I saw a list of, say, six doors with DCs from 10 to 40, I'd be a lot more likely to make up a dozen more than if I just saw a generic DC by level chart. A tiny nudge to the imagination can go a long, long, way, and it also helps make the world more real.
So a list of 5 doors isn't good enough, but 6 is? Of course, I'm just ribbing you here, but it does bring up a point: how much is enough? How many examples do you need? I could give you a list of a description of doors/locks/walls/whatever at every DC from 1 to 40...or I can give you the tools to make your own and 5 ideas, which is what was done.
Also: I believe that there is some truth to the saying, "to define is to confine." If I give you too many definitions, you have the problem of the DM/players/designers being straightjacketed in the definitions you've already given in the DMG. You should be consistent, but some people will say, "OK, I need a DC 24 door for the PCs to break down...guess it HAS to be an adamantine-banded door made of treant wood." And, of course, the first time you do something different in a published module, well, suddenly your designers are hacks, your editors should be fired, and you're not publishing quality material.
Now for the prize, tell me... what type of ledge, to balance on, do you put in a paragon tier locale? Because it's DC isn't connected in any way to tiers... this is where I find the problem with the clarity of skills in 4e.
This is exactly what the DMG section on Realism is talking about. If it's just a ledge, you can use the sample DCs which assume it's Just a Ledge. If it's a ledge of ice covered in jagged shards of icicles, it might just be a paragon tier ledge...so you use the scaling DCs. Just because you become paragon tier doesn't mean that every ledge is a paragon ledge; it only is a paragon ledge if you, the DM, need it to be. And then you need to describe it accordingly.
It's a consistency issue that could have easily been resolved if the designers had decided on one way to represent DC's for in-game challenges...either through objective description or tier based assignment... but why both?
Because sometimes you want specificity, and sometimes you don't? Because unless you know a lot about how locks work, specific descriptions don't help you visualize it any easier? You can argue that the Open Lock DC descriptions are not to your taste, and I'll buy it. To say it's some larger flaw of game design seems more like a stretch, since playing and DMing a roleplaying game demands some level of common sense and the ability to put together different mechanical concepts. We can safely assume you understand both what specific examples mean, and what "paragon tier" means.
I have no problem imagining a DC 40 Acrobatics check ledge. I only have a problem with the idea it becomes a DC 10 ledge if a 1st level PC walks on it.
I think this is where the section on Realism from page 23 has to come into play. Besides, if you've already defined the ledge at DC 40, then that's the DC. You use the improvisation rules to define, mechanically, the current situation. If you tie a description to a DC, it's a two-way street; once you've tied a DC to a description, you should stick with it.