Lots of great comments in this thread.
I have to say I basically totally disagree with the premise of your post, adding City Defence has been a great innovation to the series - no more city captures by a lone warrior on t5!!
Just to clarify: My suggestion was not to completely remove city strength, I just think it should grow less rapidly through passive tech progression and be more tied to fortification level, and I particularly hate the "city strength equals CS of strongest unit in empire" mechanism. When that's said, I actually think that a city strength of zero when no fortifications or units are present make some sense, and the problem you mention can easily be avoided by a) making the palace grant some fortification level (I actually use a mod that does this to prevent those silly very early warrior rushes), and b) having the player use his starting warrior for defense rather than using him as a reckon unit.
I feel like the OPs solutions would end up just being swings and roundabouts. Sure, it would be easier to take a single city, but when you do, the local cities get (effectively) free military units? In my games, that would effectively render them invulnerable. I'd be at more risk of losing a city, but substantially less risk of losing the game. Also, requiring population to build military units would absolutely nerf offensive play in the game. Pop increases so slowly and desired for districts etc that requiring it for military units, until your cities reach that plateau where pop becomes largely irrelevant, would cripple warfare in my opinion.
Obviously linking military units to population is a massive change that could not just be implemented without considering other things for overall balance. But when I suggest a mechanism of a military unit being a citizen assignment rather than just removing a citizen, it was exactly to account for things like district limits. As for the food issue, I'd like to point out that Civ6 currently has the potential for massive food overflow courtesy of the farm adjacency mechanism - one of my favorite features of the game, I could add - but sadly the overall game balance means that for optimal play, carpets of farms are not really optimal strategy because of housing and happiness limitations. My point here is: If population and food tied into military, the game already has a framework to support that, which is currently not really used.
As for the swings and roundabouts thing, if I understand you correctly: Actually what you sketch out is kind of what I would want, not to make things overall easier, but to move emphasis from unit vs. city combat to unit vs. unit combat. The reason why I suggest spawning new units when a city is capture is because of the well known situation from Civ6 (and Civ5) where, once you make an initial military break through, you're basically facing an enemy without any army. Once you reach late game and have artillery, that basically means once you've taken the first city, you can freely take the entire empire (which may not be entirely unrealistic, btw.). I'm not suggesting a fully equipped army should spawn, but if anything between a couple and a handful "militia" units spawn in neighboring cities, that would at least buy the defender some time to rebuild a defense to a certain degree. But the details of how such a mechanism would work is not something I can claim to have an exact image of.
The other way to go would be to make sieges actually play like ... well sieges. An attacking army should cause starvation over time if they occupy enough of the hexes around a city. In an actual siege, the attackers would also have to worry about that, but you can use the attrition damage from the defenders inside the city to fudge it. It would actually open up a lot of interesting gameplay opportunities for both sides. You could give players an option to divert some of the food a city produces into emergency stores, creating a trade off between maximizing growth and preparing for emergencies. Once food runs low, you also have to make choices about where the resources go. Do you keep the army fed and take the loss of population or do you reduce the strength of your defenders to keep your city productive?
It would also give naval warfare a new importance since a coastal city could resupply by sea if the attackers don't have ships to block those hexes. You could also potentially add a mechanic where defending military units could permanently raze their own farms for some food and to deny the attackers the option to pillage those hexes.
I can only say I agree 100 % with this. I love that sieges actually made it into the game, and I would love for it to have much bigger impact (and wouldn't mind it having stricter requirements, like occupying ALL tiles, not just have ZOC over all tiles). Besieging a city should make it unable to work all tiles without the besieged limits, and in general, a town should not be able to work a tile occupied by a hostile unit (as it was in Civ5).
Buffing walls is a huge mistake.
IMO one of the following
-- Reduce city bombard damage to siege by half or more
-- Start catapults with 3 range but nerf movement, sight, combat stats vs units.
It is a bit silly that siege can easily get wiped out by the city center and there is no real way to stop it beyond overwheling with siege weapons. So that makes them good against cities which have their defenses destroyed. Which is when you don't need them.
Giving siege weapons a range of 3 is a pretty dangerous move, as it can extremely difficult to counter them, particularly when you are down to one city. I did like a suggestion made by someone in another thread some time ago: When a city is under siege, it's ranged attack ranged should be reduced from 2 to 1. The logic of this is that the city attack can be interpreted as raids from the city against the hostile units outside the city. When the city is under siege, those raids can only target units imediately next to the city. This would mean that once you bring a city under siege, you can move in your siege engines without having them completely slaughtered.
A ranged unit stationed in the garrison should still be able to shoot to second ring, however.