Corruption in Civ3 affects the opportunity-cost of infinite city sprawl, while never making expansion a net negative. However, building another super-high-corruption city may not provide as much benefit as building, say, a Library, or a Swordsman to try to conquer a neighbor, or a Galley to go scout for other civilizations.
90% corruption cities isn't really fun, though, just as going bankrupt from expanding too quickly in Civ4 isn't really fun. Both are less unfun than the humungous penalties from negative happiness (also caused by expanding too quickly) in Civ V, IMO.
Which one is worse, corruption or maintenance? Eh, depends on your playstyle. I find Civ4's maintenance less fun in the early-mid game and Civ3's corruption less fun in the late game. Basically each of them is less fun when you notice them the most.
I wonder if a less unfun mechanic, which would still limit city spread and keep barbarians relevant, may have been limiting the number of Settlers a civ can build at a time, similar to how Civ4 limits the number of missionaries that can be built at once?
It's also worth noting that it's possible to mitigate corruption to a significant degree in Civ3, too, even without modding. In my current game, I'm using the Communism government, and while it often underperforms when switched to from Monarchy or Republic, when planned for it can be very powerful at reducing corruption. The usual problem is if only a small core is developed, it will lose more productivity than the outer areas gain, resulting in a worse economy, hence why it requires somewhat careful planning. But done right, it can be powerful in a similar way to State Property in Civ4. In my game, I control about 40% of the world's landmass and corruption is low across the board thanks to Communism + Courthouses + (where useful) Police Stations.
Specialist economies can somewhat sidestep corruption in Civ3, too. Corruption doesn't apply to yield from specialists, so (in a Monarchy/Republic) lots of irrigation + scientists/taxmen/civil engineers can result in decent commerce and building production even in far-flung cities. Drafting allows bringing in a decent number of troops from these cities as well, or police officers can be used to get a middling amount of military production for non-draftable units. The tradeoff is it isn't an intuitive style of play and requires more overhead.
The other tradeoff is that by making maintenance purely based on city count/distance in Civ4, the maintenance per building aspect of Civ3 was lost. This added a more realistic consideration of whether a building was worth the benefit. Is that Barracks in the Tundra really worth 1 gold per turn to maintain? We could build a Stock Exchange in that provincial capital, but if it's only going to provide 3 gold instead of 5 due to its 2 gold per turn maintenance, maybe we'd be better off training some Artillery? Perhaps most commonly, how productive does a city need to be to make it worth building a Factory that then has to be maintained? It's not a huge part of the game but it's somewhat of a counter-point to the argument that it's more realistic for new cities to have to be supported by existing ones (which I also agree with at least when looking at a short-term, brand-new-colony angle. Maybe not from a 100-year-old city angle).