The real problem that this discussion is pointing at is the fact that having the initiative is overpowered (mainly because of the way collateral damage on the attack currently works).
In my private modding, I've limited ancient and medieval-era siege to dealing 20% damage, and then retreating. Not only does this cause siege to realistically retreat more often, but it also puts siege in its proper place as an accessory to battle, rather than as the determining factor. You still end up needing strong, well-promoted units in order to kill the opposing SoD's counter units.
This applies to the player on the strategic offense as well as the person on the strategic defense. In the case of my simple siege mod, having the initiative to engage in active defense does not guarantee victory. If the opposing stack is larger, has better troops, has a good mix of counter units, and/or is approaching along defensive terrain, then the situation becomes very problematic for the defender, and simply taking the initiative to engage in active defense will not be a quick fix to the situation.
It's in a case like this, where having the initiative has been made to be a beneficial but not overpowering thing (it's still beneficial because you can quickly bring in/build reinforcements and counter the enemy's local concentration of force), I would really appreciate the ability to give forts, at the very least, ZoC. That way you could force the attacking stack onto flat land and at least have a chance of defeating it, or at least badly wounding it, after having had one's siege take the stack down to 20% health.
Aside from ZoC, the only other mechanic I'd really like is to be able to conquer territory on a tile-by-tile basis. Like, if you had a "pillage" function that, instead of pillaging an improvement, just flipped the territory to your side (even while, perhaps, not necessarily altering the culture in the tile---sort of like how a master can claim territory in a BFC from a vassal even if the vassal's culture there is higher. In the case of pillaging territory, this might last until the tile "revolts" back to the civ with the higher culture, and just like with cities, tiles would have a certain % chance of revolting back to the real culture owner every turn, mediated, of course, by how many enemy units were stationed on the tile, just like with cities. Okay, this might require a lot of calculations each turn on a large map, but I still think it would be a neat feature). That way you could recreate things like WWI, with forts galore and an advancing line that allows each side to quickly bring up and concentrate force at points on the front as the line moves forward, thus making the taking of territory essential.
Edit: I'd like to add that adding ZoC will always tend to help the AI more than the human in war (and let's face it, in war, it's the AI that generally needs the help). The reason is, humans can (and do) already engage in mostly active defense when confronted by an invading SoD, but the AI oftentimes will hide in its cities and rely on passive defense. ZoC will strengthen passive defense by forcing the attacker to attack strongpoints in order to bypass them. The AI would benefit most from this, as their units that they often have fortified in forts on top of resources would actually be useful for them.