The basic idea is a space conquest game. This has definitely been done before. You have planetary systems you can conquer, and you build fleets using them and conquer more. There will be a simple galaxy generator that might populate 50-200 planetary systems (depending on the scale you might choose).
The abstractions are just removing lots of special cases. In most similar games I've played, you can build a wide variety of units (often accompanied by a varied research tree). I intend to replace this with a simple power/disposition system. Basically, variable statistics. So a particular fleet might have an offensive or defensive disposition, or neutral. It might have a fast or slow disposition, or a heavy/light one. Some statistics might be naturally exclusive, for example I might disallow fast/heavy dispositions. The power is basically a combination of the size, disposition and relative technology levels.
It will be turn-based, but each turn will be divided into two phases, decision and simulation. Orders are given during the decision phase. The combat is resolved in the simulation phase. I also intend combat to be a multi-turn process. Planets are captured by laying seige to them for a number of turns. Two armies facing each other will do damage based on their dispositions, only if at least one is in the aggressive disposition will notable combat take place. Two fleets both engaged in an offensive stance should be resolved fairly quickly, with victory mostly based on the power difference between the fleets. If one army is neutral and another is aggressive, the neutral fleet might fall back rather than take excessive damage in a single turn.
The inspiration for some of the interesting ideas come from recently reading The Forever War. I really liked his description of the implications of relativity on interstellar combat. One simple example is that the human space craft travel to the enemy space that a long time has passed for the aliens. This means that the humans are facing technology which literally didn't exist when they set out. I don't believe I can replicate the ideas in the book exactly, but I can offer mechanics which have similar effects.
It will obviously be necessary for fleets to travel faster than light. My ideas are still forming, but one is to drop the relative technology levels of fast moving fleets (in the absence of logistical support perhaps). Another might be to allow the enemy players see the fast-moving fleets from further off. This is obviously artificial - the speed of light forbids this, but this is a game! The point is to design a mechanic
Another idea is to make the galaxy change over time. So the planetary systems will rotate around the centre of the galaxy, but also around each other. This is partly the reason for a separate simulation phase. While individual star systems might have little effect on one another, galactic features like black holes and globular clusters might have noticeable local effects (which I will probably exaggerate for effect). I might simplify the gravitational interactions by grouping areas that are close together. Getting this to work will be the main technical challenge I believe. I will probably have to simulate thousands of non-populated star systems to give the system some stability. I would have to at least draw them anyway to create a nice background effect to give the impression of a galaxy. I'll have to generate stable galaxies, where very few collisions will occur over the lifetime of the game. In particular, care must be taken to avoid the entire galaxy collapsing toward a single point. If it is infeasible to simulate gravity, I might just go for a very simplified system where the different star systems rotate around the centre of the galaxy at a fixed rate. But because different star systems rotate at different speeds, the same dynamic galaxy will be created.
My experience with the critter game taught me that I've a lot to learn about game design. This will be a huge test for me. I don't know if I can make a good strategy game in a ever-changing world. It sounds like an interesting idea, so I am eager to try it out and see if it is any fun. It sounds like a bit of a nightmare though, you have to plan X turns down the line when your planets and ships are all in constant motion. Add to this the various real and theoretical/imaginary interstellar features that can be included, the aforementioned black holes and globular clusters, but maybe supernova, wormholes, nebula. I will include anything that has a bit of character, and can act as strategic key points. I'm not too concerned about realism, I will scale any feature to fit the game, even if it doesn't really make sense. I might even include new ones, such as massive objects which have an inverse effect to gravity. These might be pre-populated features or might be created by the users, I'm not sure yet.
I don't look forward to writing AI for such a system though! I expect it will be hard enough for a human to meaningfully plan ahead in such an dynamic environment. The interface will present a challenge. For example, if you order a fleet to move to a point that will take 10 turns, should that point be relative or absolute? Did the player want the fleet just to approach the galactic core, or did they intend that it end up close to a particular resource of interest. The strategic value of certain areas will change over time as they move. A once important planet might end up moving to the wrong position and become isolated and useless. This could feed into the economic system in the long term, if I take it that far.
I've only barely begun work on this project. The working title is "Conq". This post is intended as a roadmap for future development. At the moment it is a top-down 2D game, but when I look at video's like the Infinity Game Engine by Ysaneya (included below), I keep thinking how cool it would be to do this in 3D. Imagine a fleet surprise attacking - from perpendicular to the galactic plane!