Interesting post!One of the things I find interesting about the 'typical style' (if there is such a thing) for laying this kind of stuff out, much like you're outlining here, is that it is STILL very much beholden to the archetypal dungeon room key in some respects. For instance I've rarely, make that pretty much never, seen in a product where there were fully elaborated descriptions of the relationships between things.
We may know, as in Gary's Hommlet exactly the contents of every house, but who's going to stand together with whom? What happens when you kill Fred down the street, doesn't he have a brother? A landlord? Someone must inherit his stuff, want to find out who killed him, etc.
In some sense, I think this 'classic' type of setting design, when it comes to settlements, is entirely inadequate to the type of play that [MENTION=42582]pemerton[/MENTION] seems to be espousing when he asks about the value of worldbuilding. I would say that in terms of his needs these social/cultural/economic details are MUCH MORE IMPORTANT than the trivia about how many coppers are under Farmer Joe's floorboards and what the probability of finding them is (all of which can and probably should in that kind of play be made up as needed anyway).
Its not that nobody ever thinks about this stuff at all, but its weird how you get these detailed maps of buildings but only at best some incredibly vague idea of who owns them, what their allegiances are, etc.
I think that in the Hommlet era alignment was meant to carry a lot of this information.
In my 4e game I've relied on religious affiliation to carry a lot of weight (it's a cosmologically-focused game). The three main settlements that have figured in the game have been Threshold/Adakmi, the duergar stronghold, and a githzerai monastery. For the latter two I basically set it up with two main groupings in each - and framed it so that the PCs may themselves have split allegiances across the two factions. For Threshold, I had a power struggle between baron and patriarch (I think that was my own idea, but maybe I got it from B10?); and then introduced some nuance into the baron situation (with the advisor, the niece who was the spitting image of the PCs' friend from the past, etc).
In my Traveller game, when the PCs assaulted the bioweapons conspiracy outpost, I wrote up the NPCs with connections between them (one ex-army guy had been a comrade of the PC army guy; another NPC was the sister of the shuttle pilot the PCs had hired; etc) - nothing very intricate, but it just added a bit of extra depth and complexity to the resolution of the situation.
In other words, I think the difference between nothing and a little bit goes a long way.
I don't have any of the classic RQ modules, but some of them must have tackled this sort of stuff.