(For instance, they can't just nerf translation: they've promised and delivered a gameplay mechanic for G-force penalties for certain manoeuvres. Because of that, your ship has to be able to pull vertical accelerations of 10G or more; and so given full control over your ship, you can use and abuse this however you please. A choice has to be made.)
I disagree.
Let's look at this from practical perspective. I expect most atmospheric-capable ships to also be VTOL capable which obviously already means they must provide at least 1g of acceleration in upward direction, plus the margins necessary for climbing and arresting descent - so let's say 2g capable hover thrusters at minimum. 3g-4g capable hover thrusters are not impossible either.
The other directions have no practical requirements for such high thrust capacity. And, turns out if you really do want a "planes in space" type of flight model, this will actually be the only way to get anywhere close to such experience (with the hover thrusters substituting for lift in turns). And you definitely do not need a 10g translation system in every direction. 2-3g's would be very much sufficient in enabling "aircraft-like" turning ability for the ships without making them seem completely ludicrous.
However since there are ships that are not designed for atmospheric operations - like the Aurora - it would likely have about the same translation acceleration in up, down, left, right and backwards - but a higher acceleration forwards. Sure, the fly-by-wire system could keep the nose aligned with velocity vector if you wanted, but I don't think massively capable translation thrusters would provide a very good gameplay. If the only thing you need to do is point and shoot, the main part of gameplay will be decided by the ship and weapon stats, with little room for player skill being a meaningful component of gameplay.
Skill progression is an important component of satisfaction I get from games, and it should be possible to gain an advantage by executing maneuvers manually if you have the skill necessary to do so. Good example would be doing turns by turning your ship 90 degrees off velocity vector and using your main engine to generate the highest possible acceleration, to cause a rapid course change. Whatever the RCS translation fly-by-wire can do, it should NOT be able to provide a higher acceleration than with the main engine, that just makes no sense at all - either from gameplay perspective or within the internal rules of the game universe.
In other words: Different ships should behave differently. If the spacecraft is atmospheric VTOL capable, sure, it could use hover thrusters to mimick upward lift. Other directional translations should be much weaker - in line with what you would expect from the RCS thrusters. They must already have some given thrust level for rotational control, after all.
Spacecraft that are exclusively designed for microgravity operations should behave like spacecraft. They should of course have their own advantages - like better rotational control, or better forward acceleration.
This would force players to use different tactics with different ships. For some ships, the auto-alignment fly-by-wire could be the best compromise between performance and intuitive controls, for others you might have to practice a lot to get them to perform optimally.
sounds a lot like the stab-aug feature from space combat. it ate fuel like it was nothing.
i personally like lift fans for vtol flight. then have 0.5g maneuvering thrusters. this way you can land anywhere with an atmosphere, as well as low gravity bodies. if you need to go to other places, have your ship fitted with disposable thrust pods. so in addition to a way to land, you also have an excuse for not having ludicrously large thrusters in space.
