First, I would really like the Moloch to either loose his imps, have them turned into a lifelong protection force, or, in general, have the fleeing issue solved in a nicer matter than it is now. At the moment, every pretender with autosummons really ought to get a cost reduction for that special ability, which cannot be what was meant to be.
Second, a general UI overhaul. Probably already scheduled.
Third, and possibly already covered by second, but important enough in its own right, a rework of the blood hunting/blood sacrificing/tax adjusting madness would be high on my list. I like occasionally dabbling with blood nations, but it is a real micromanagement pain.
Some simple examples (that would even do for Dominions 2 in a forthcoming patch):
- Allow blood slave sacrifice directly from a lab present if the priest has no blood slaves but there is a lab present and slaves in a lab. Do NOT reset a priest from Sacrifice to Defence if there are no slaves. If I have started people sacrificing, I want them to continue until I give them another order, and if I have more people sacrificing than slaves and really want to be sure that certain priests get priority, I can still manually give those select priests blood slaves of their own(This would allow you to use the "pool" command without having to then distribute blood slaves to everybody sacrificing). This would allow, e.g. Mictlan players, to set up priest sacrificers and use the "pool" button without the endless tedium of the current system.
- Alternatively, add a "pool slaves EXCEPT FOR THOSE ON SACRIFICERS" button. Not as good as #1, but still slightly better than what you have now.
Unless unrest is completely reworked in Dominions 3, a big red button labeled "SET TAXES SENSIBLY", which would, upon being pressed, set the tax values in all your provinces to a sensible value given the current unrest setting, would prove very useful. You could ignore it if you wanted to and manage all unreset yourself or press it when you felt a need for major readjustment and would like sensible default values. E.g. blood nations would quite likely press this once per turn and then adjust the tax rates for a few strategic provinces rather than, as now, often having to check and adjust the tax rates of a significant fraction of their provinces per turn. A sensible value in the current system would be something like max{110-2*unrest, 0}. A nicer Version would be max{100-v*max{unrest-d, 0}, 0}, where v is value, say, {0...4} and d is a value {0...unrestMax} both chosen by the player, though it and "advanced" (
) Version like that was used it should probably be called a tax policy rather than a Big Red Button.
Another (minor) irritant is the current solution of putting save games in a subdirectory directly in the root of the installation (on windows, at least). This is ugly and unnecessary. A subdirectory "save games" in which the games directories would be placed would improve the aesthetics and also minimize the risks of some mistakes when messing around with the files manually.
--- Architecturally, the franchise would probably win quite a bit in the long run (if one was looking forwards to producing a #4 or using the code base of #3 as the basis for a new game) by finally doing a major rework of the code structure. Cleanly separate it based on components instead of having it all mashed up together and you gain a much more maintainable code. Even splitting it up in just four discreet components would help:
- Server (executes all commands and calculates turns based on input in a well-specified format, returns an output per nation)
- Client (displays turn ouput and accepts input that generates a turn order)
- Netserver (interface between client and server)
- AI (generates a turn order based solely on the information stored in its turn-output and map, just like a player)
Since Dominions is not real-time, there is really very little reason whatsoever architecturally for having the client perform actual execution. While it does make some things much easier (e.g. if you move item A from commander B to C, you just plain move it on the spot rather than storing a "move item A from B to C" action, which has either to be amended to an "move item A from B to D" if the player then subsequently moves it from C to D or to be stored subsequently leading to bloat), it leaves the turn format vulnerable to hacking and corruption.
I appreciate that the Illwinter team is rather small, but this really is an area that is open for major improvement design-wise. (Whether it is economically feasible is, of course, another issue - and one that only Illwinter can answer). If the development team needs (unpaid
) help on this, there are several client/server experts on this forum who might pitch in with ideas and advice if asked nicely. Esben comes to mind though he is currently AWOL.
The AI could very profitably have major factors involved in decision taking externalised as modifiable floats allowing everybody to tweak to their content (or discontent as it might be). I am not asking for a full script language here, just a dozen or a hundred floats I can tweak to adjust AI behaviour, mainly on overall (rather than specific) spell choice and troop composition, though front/garrison values for army movement would also be greatly appreciated. Allow a MOD to point to a specific AI file, and, if necessary, store the AI settings chosen (a few K of data max) as part of save game fatherland file. No breath of winter, shockwave, or immolation chosen by the spellcaster AI unless you really want to? Set the KILL_FRIENDS_WITH_FRIENDLY_FIRE_FOR_THE_HECK_OF_IT value to 0.0