I do liek the AoM Idea, (It slightly resembles the epic FFH).
That's kind of the intent. Cross DUCKS with FfH, add some AoM content, and add a few new twists of my own, and you'd get the early-era mod I'm planning. Likewise, cross RFC with the city-state diplomacy mod, add some Age of Empires content, add a bit more of my own design, and you'd get what I have planned for the middle section. After all, I'm not pretending that my Alpha Centauri mod is unique content; obviously, a lot of it is drawn from SMAC, sometimes by way of the Planetfall mod, but I've added quite a few new elements to it along the way. So expect something along those lines, albeit with fewer new technologies, units, etc. The Mythology mod will include some new techs and a large number of new units and buildings, but the Empire mod will only add a handful of units, very few buildings, and no techs as it'll mostly focus on events and new Lua mechanics.
However, can you do that without the .dll? How far can you go without it?
A lot further than you'd think. Look at the Content mod, for instance; I've managed to get around a lack of DLL access by using a number of randomized effects (like the tech-stealing and breakout), and I've generally designed it in such a way that the AI actually does better with it than in the vanilla game. Combining overly specialized units (anti-air, submarines, carriers, fighters) with their more useful support brethren (artillery, destroyers, battleships, bombers) really helps the AI avoid having too many of a useless unit. No need for the DLL when the mod is optimized for the strategy that the AI already follows.
So consider the mythology mod. Most of the myth units will have abilities that are already present in the vanilla game, or that I've added in the AC mod. Most of the deity bonuses will be things already possible through current policies and buildings, or that I've already added in the AC mod. In fact, I'm going to get around the save/load bookkeeping issues by using the buildings in cities as the benchmark; if a city has a Fire Shrine and the owner is Greek, then I'll know that he's got Hephaestus as a level 1 minor god in that city. "Upgrading" a god, then, is done by replacing that Shrine with a Church (which'd give larger bonuses); no need to save the values somewhere else.
Adding a new Yield type (Favor) allows me to use the existing tables to link to buildings, units, etc., so there's no bottleneck there. Creating a custom interface doesn't require the DLL, and I can get around the AI's inability to handle new concepts by simply removing the Flavor system from the equation. The key is that all spending of Favor will happen at the start of the turn, similar to the notification you get when you unlock a new policy. So where you, the player, would open up a screen to make your decisions, the AI will just randomize it with weightings based on "direction".
To illustrate the "direction" idea, let's say an AI has chosen the Greek Pantheon, and chosen Hades as his central deity (primary focus Death, secondary focus Wealth), with the Pious favor trait (gains extra favor from Priest specialists; this is the default for the Greek pantheon.) Note the primary/secondary split; I spent last night trying to match gods to foci for five pantheons (Greek, Egyptian, Norse, Hindu, and Aztec), and it worked best with each pantheon having four "central" gods you could select from, each with a primary and secondary focus, and then ten minor gods with one focus apiece. In fact, it worked so well that I'm looking to drop the "custom" pantheon idea altogether, and just allow you to rename your gods as you go; I'd really prefer to only use four pantheons for two reasons: 56 icons instead of 70, and I've only got four favor-generation methods (extra from Priest specialists, extra from religious buildings, extra from Missionary units, gain when fighting).
Death is 0 on the Law/Chaos axis and negative on the Material/Ephemeral axis, as is Pious, while Wealth is 0 on the M/E and negative on the L/C. So Hades is in the (-,-) quadrant at the start of the game, probably at -1,-3 with the numbering scheme I've got currently. If he's offered a choice of a minor god at that position, he'd be able to choose between the gods of Arts (Apollo), Healing (Asclepius), Justice, and the Seasons (I think I had Demeter here). Since Justice is Zeus' secondary, it's off-limits to the other major gods, but that'd still leave him with 4 choices. A bit further in the -,- direction is Air, which is Zeus' primary domain (and therefore CAN be added), and depending on which direction you go next you'll probably unlock War (Ares), Travel (Hermes), Knowledge (Athena), and/or Beauty (Aphrodite) as well, so you won't run out of choices; generally speaking, you'll have 3-4 at any given time. You COULD go in the other direction a bit (the Lawful and/or Material axes), but it's pretty much impossible for Hades to go far enough to get things like Fertility, Stone, Crafts, or Water. (Since two of those are the secondaries for Poseidon and Hephaestus, you wouldn't be able to take them anyway. This was deliberate.)
I'll try to whip up a basic diagram to illustrate how these are laid out. It sounds a lot more complex than it really is.
A Human might pick carefully, but the AI would probably just pick randomly, or maybe with a bit of weighting involved. Point is, instead of setting flavors and letting the AI pick internally, I'd do the randomization myself; no need for DLL access when the decisions can be made to always occur at the start of a turn, and when the randomization is so simple. Likewise, if an event comes up that asks you to make a choice between two effects, each of which is tied to a direction on this map, I'd have the AI weight the options by what it's done so far. So, if Hades is currently at -1,-3 and is offered four choices, one per direction, then it'd probably weight 40E/30C/20L/10M (basically, -5 to +5 is the normal range for each axis, so each point shifts the probability by 5%.)
The Empire mod, on the other hand, will absolutely need the DLL. There's just no way around it; adding diplomacy and espionage in a manner that involves units will almost definitely require adjusting the AI, because he has to be taught how to handle a non-combat unit that WANTS to be in enemy territory.
I think you may need some help for that, isn't it?
Depends. Most of the basic mechanics I can do myself, since I'll be scavenging code from my other mods as often as possible. If I decide to add more Pantheons beyond the five I have now, then that could use some outside help, but I don't want to bloat this up to the point where there are separate pantheons for every civ. Beyond a certain point it's just window-dressing; the differences between the Pantheons basically come down to which 8 out of the 18 foci are tied to the four primary gods. The Greek, Norse, and Aztec pantheons generally tie the elemental foci to their primary gods, while the Egyptian and Hindu don't. The rest vary widely; I think Beauty and Knowledge are the only foci that NEVER tie to a primary god.