[Barnacle Bill]

I don't want any new "stuff", just things fixed regarding the AI and diplomacy:

There should be a relationship between the AI agreeing to do something in response to your demand/requst and actually doing it. Really, it should just agree and do it or tell you to forget it. If it positively must be programmed to lie sometimes, then if it agreed there should be only a small change it will not.

AI or human, colonizing a system should automatically claim it. You should not be able to claim or colonize in a system if it is already claimed by a race with which you have any treaty (even NI). Any deal establishing a treaty (NI or better) should have to include resolution of all conflicting claims - for each contested system the deal proposal must either give the contested system to the adressee, or demand that the adressee give the system to the race proposing the deal. Giving a system should automatically abandon any colonies there, regardless of the population size, unless giving the planets was part of the deal.

The AI should have some sort of memory regarding its relationship with each other race, so that its actions make some sense. It should weigh both its strength vs the other race and its feelings about that race as part of decisions about treaties or declaring war. It should only declare war if it is furious. Just being furious would not be enough, though - it should also consider relative strength to the other race and its "war score" (gains/loses at the other race's hands). The "war score" be a running total which moves toward zero on its own by some amount every turn, but goes positive if it loses a ship or planet (varying by the value of what was lost) and negative if it kills/captures a ship/planet of the other race. The positive score would make the race more likely to offer or accept peace if already at war (because it is losing) and less likely to declare war (because it remembers that it lost Last time). Peace breaking out while a race's war score is positive should convert its feelings toward the race it defeated to neutral if they were below that. The idea is that the loser would tend to stay furious and eventually, as its negative war score decayed to zero, come back for a rematch.

The relative strength thing should also cause the AI to slowly become more angry if it is stonger than another race (at a rate proportional to the imbalance and its own agressiveness), but not to become more friendly if it is weaker. The idea is so eventually it would become furious and declare war if it is and agressive race and stronger. This would only apply if systems claimed by the two races are joined by a warp piont.

There should also be a "grudge file" - the AI race would keep track of each system it gave up (surrendered its claim to) in a treaty that ends a war (only), and count this as a permanent bias on its feeling toward whatever race currently claims that system (and in even of war against that race, will reclaim that system). This effect will be extra strong if it is thei AI's home system. Similarly, it would have a bias for every planet populated (in whole or in part) by its species which is held by another race, and in the event of war with that race would claim any system containing such planets.

The more races an AI is at war with, the more likely it should be to make peace, and the less likely to attack someone new.

If an AI race is hemmed it (no unexplored warp points and all the explored ones lead to ststems claimed by another race), at peace, and has no breathable planet it can colonize within its own claimed systems, then it would slowly become angry at all its neighbors until it declares war on one of them. The war score and relative strength considerations would apply to that decision.