You've got to love how a thread on the latest patch notes turns into a spirited debate about the pros/cons of civ4 style maintenance and local happiness vs civ5 style global happiness
So in the tradition of off topic debate - here I go, once more into the fray
Both systems have their flaws and neither system is what I'd call "realistic". Realism is not something I've ever associated with a civ game. In civ the game mechanics are always so abstract that for any given mechanic I can come of with both justification for it and a reason why its completely unrealistic at the same time. But maybe that's just me...
So I'll stick to game mechanic arguments.
In civ4 you have city maintenance as the main wide empire/number of city limiter, with local happiness as the main growth/raw population limiter. The city maintenance will punish you (quite severely at times) if you expand too early - ie prior to courthouses/markets/+1 trade routes. After that initial speed bump you can quite happily invade and annex the entire world without much constraint (not something I've found possible in civ5 btw).
In civ5 you have global happiness instead, the primary purpose of the mechanic is as an empire growth limiter (total population not so much number of cities) - the number of city limiter is secondary. Additional cities do initially cost you extra unhappiness but these costs can be largely (but not entirely) overcome with social policies + forbidden palace wonder. This however incurs an opportunity cost via both increased social policy costs and choosing the more ICS friendly policies over others.
In an ICS game (as I understand it) you need to artificially limit your population size in your non-core cities so you don't waste your happiness. This means that the mechanic is actually doing its primary job - limit raw population (and indirectly science output).
In civ5 in a normal play game I will often (due to conquests for rapid expansion) find myself with happiness problems forcing me to slow down. This happens to a degree that never happened to me in my civ4 games. More often then not I mitigate this by going down the Piety SP line to take theocracy costing me 3 SP that could have gone into say rationalism for science.
I prefer the global happiness model as I believe it is a base design that is very adaptable and allows more interesting game options to be added then a local based system. It is a much more extensible design, you can add all sorts of cool effects and trade-offs based on rewarding excess happiness, punishing excess unhappiness with interesting interaction with the SP system (like bonus culture, bonus science for excess happiness). As a raw population limiter I find it works better then local happiness system (in civ4 between luxuries, multiple religion buildings and culture slider the local happiness cap is entirely mitigated) and is about the same as a wide empire limiter (early limiter only).
I think where civ5 breaks down is through overpowered great scientists and research agreements, which breaks the science equals population potential times science infrastructure design. Hopefully increased science costs will fix this.
The issue is the interaction with mechanics outside of the global happiness mechanic that make ICS a viable/optimal option not the happiness mechanic itself.
The argument that civ5 global happiness model is broken/rubbish as it allows/encourages ICS is based on a couple of false assumptions.
#1 ICS is evil and must be destroyed at all costs
I personally don't have a problem if the system allows for an ICS style of play, as long as its not forced on you as the only viable game play option.. Its not a style of play that I enjoy (same goes for RA blocking) but if that's what others enjoy that's cool (the more styles of game play that are possible the better imo).
From what I've seen its not the dominate style of play for multiplayer (NC start is preferred I believe) so its clearly not the one optimal way to play ie it has weaknesses against thinking opponents. In single player no-one is forcing you to play in any particular way.
#2 that the primary design goal of a population growth/number of city limiter system is to prevent ICS
There are other design considerations at work where. To set "must not have ICS" up as the ONE primary over arching design goal is very one dimensional thinking.
#3 ICS is viable because of global happiness system
No its more to do with a combination of trade route income, city tile bonuses, per city bonuses, GS, RA abuse. All of which are balanceable: tweak the numbers to give high population city more advantages over the equivalent population in small cities.
There is nothing fundamentally broken about a global happiness system. As a design it is both reasonable and justifiable, and adds new and interesting game elements, tradeoffs and modding potential.
If you don't like the system that's cool - some people like anchovies, some don't.