I am going to use the below quotes as examples to point out why a many of the proposed "solutions" aren't solution but just another form of unbalancing (as said, these are just examples, other proposals aren't much better):
If you want to get rid of ICS, 3 things you can do:
<snip>
3. Make another strategy competitive (improve bigger cities, rebalance happy buildings) Problem: needs balance testing
<snip> then plop a city down with every tile overlapping, you end up with an extra 3 or 4 unhappy for that city. Doesn't solve the issues,
Any suggestion to help bigger cities in comparison to smaller ones runs at the risk of actually making ICS even more successful, because the bonus for the bigger cities just adds to everything which you already get from the smaller ones.
The second part, though, may harm you in the early game, long before ICS has shown its whole strength.
Imagine you have two good spots for cities in the beginning, yet very near to each other. Your example would literally punish you for correctly making use of both spots for early growth of your empire (not taking into account that this is quite unlikely to scale well with different mapsizes).
-Changing the unhappiness formula from 2 per city and + 1 per pop to something like 4-5 per city and +½ per pop (make it 0.25 for specialist with the right civics).
-Possibly add a "sprawling empire happiness penalty" (depending on mapsize)
-Keep a running tally of culture, and cause civics to deactivate if you go too far into the negative.
The first suggestion is even worse than in the quote above. It might make founding the third city (when we are far away from anything like ICS yet) almost impossible since at that point of time you might not yet be able to cover the additional unhappiness.
The other suggestions are a nightmare to balance. What exactly makes an empire qualify for the "sprawling empire happiness penalty"? When is an empire considered to be "sprawling"? Do 8 cities qualify for this, and if yes, is there a difference between 8 cities in turn 100 and 8 cities in turn 250?
The only reason ICS is so great is because it gives you a ridiculous economic boost.
<snip>
1)Make inflation tie to the number of cities a player owns. the more cities, the bigger inflation.
We have already some kind of "inflation" in the game (rising unit costs by number AND time), which nobody understands. The Civ4 inflation was rather opaque, too.
2)Now make that number completely crippling for large empires (so that the trade routes generated by many small cities CANNOT overcome that inflation).
Once again, this additionally harms the forming of "non-ICS" empires.
3) make inflation resource dependant:
The less oil/coal you have, the worse your inflation. the larger your empire, the more coal/oil you need to maintain a reasonable economy
So to cripple my opponent, I have to kill his resources? Well, this would just make the human combat even more powerful, as I could not only cut the production of units and city improvements, but additionally would be rewarded by him running even deeper into "inflation".
Bottom line:
a) Rewarding city size does not help against ICS, but even accelerates it, as now you will have the benefit from many small cities PLUS the benefit from the bigger cities.
b) Having "fixed penalties" for small cities doesn't help, as this can limit ANY growth of the empire, even when ICS was not intended.
c) Penalties for having many resources are completely counter-productive as being pointed out above.
The main problem is first and foremost the maritime CS.
Since they follow the same logic of bribery as any other form of CS, a solution is not easy to find. They HAVE to offer a remarkable benefit as otherwise the gain isn't worth the effort.
Most probably a fixed amount of food to be spread in equal numbers amongst your cities would be the best solution. Let's say 10 food (numbers are just examples) which will be spread amongst your 10 biggest cities.
That way any city past the initial 10 doesn't benefit from it, but has to be grown in the "traditionial" way.
To fight the income from roads I would propose certain thresholds, after which additional, fixed costs occur (let's say for the road department). That way, after 10 road hexes there could be costs of additional 2 gold . After 25 road hexes there would be additional 5 gold, and so on (once again, the numbers are just for display purposes, and might have to be adjusted per mapsize).
This principle is similar to "inflation", yet is easy enough to be understood by the player and doesn't harm the early build up of your empire.