But given the weapons, the ship types like carriers and the general look and feel we can assume players are comfortable with a parallel to wet-navy ships that they can relate to.
Firstly, the parallels between GCIII ships and real-world ships are extremely superficial. Under most circumstances, real-world capital ships are not going to waste time blasting away 'escorts' when they could be shooting at the capital ships of the opposing fleet. Most real-world carriers are not intentionally going to come within gun range of opposing surface combatants. Real navies have not regularly used the same caliber guns as the primary armament on capital ships as on significantly lighter surface combatants for probably more than three centuries.
Secondly, while it is true that ~50% of a real ship's horsepower can drive it to most of its speed, I am rather doubtful that this is such common knowledge that just about anyone you say it to is going to go "yeah, that sounds right," or for that matter that it's particularly intuitive without spending at least a moment to think about it.
As far as the diminishing returns stuff goes, I agree with DeimosEvotec. The game's UI does not handle diminishing returns well, at least not with the current methods by which diminishing returns can be added to components. If I have a hyperdrive component which provides +1/-10% moves per turn, the reported benefit is going to be +0.9, 0.8, 0.7, 0.6, etc moves per turn for the first, second, third, fourth, etc drive component I consider adding to the ship. A ship with a base speed of 1 and no other movement modifiers save for the ones from the hyperdrive components will see an actual benefit of 0.8, 0.6, 0.4, 0.2, etc and an effective benefit of +1, +1, +0, +0, etc moves per turn from adding 1, 2, 3, 4, etc of those hyperdrive components. A hyperdrive component which costs 10 capacity +10% per hyperdrive component will be reported in the tool tip before being added as costing 11 capacity for the first, 12 for the second, 13 for the third, and so on, which would give the impression that you'd spend 11 capacity for 1 hyperdrive, 23 for 2, 36 for 3, and so on. The actual capacity required would be 11 for one hyperdrive, 24 for two hyperdrives, 39 for three hyperdrives, and so on.
While with the planned percentage of capacity solution, that is not the case and is a lot more intuitive.
There are issues with the percentage of capacity solution if they use the currently-available tools for implementing such. Adding a multiplier to MassCap on each component has the issue of stacking additively with every hull capacity bonus you have; a drive that has a -10% MassCap "bonus" which gets added to a ship will change the ship's maximum capacity by 10% of the base hull capacity rather than the current hull capacity, and while 25/50/75/100/250 for Tiny/Small/Medium/Large/Huge isn't the hardest thing in the world to remember, it can look kind of odd seeing a component which supposedly reduces hull capacity by 10% only reduce the actual hull capacity by, say, 5% since you have +100% hull capacity through technology, Hyperion Shrinkers, Helios Ore, and events, with each subsequent drive component costing relatively more of the current hull capacity, and there's also the rounding issue in the display with fractional hull capacities and hull capacity requirements. Adding a MovesCapMass multiplier to each of the hulls which makes drive components require a more or less constant fraction of the hull capacity has odd interactions with the techs reducing drive capacity requirements (if implemented as no modifier on tiny, +100% on small, +200% on medium, +300% on large, and +900% on huge, a 10% drive capacity requirement reduction from techs matters a lot on tiny hulls but almost not at all on huge hulls) and also requires a MovesCapManufacturingCost multiplier to keep production costs relative to the overall cost of the ship similar, and the addition of a MovesCapMaintenance variable would be useful to keep the drive upkeep relative to total ship upkeep constant and moreover to allow the drive upkeep to be increased for larger ships without also increasing the upkeep for all the other components on the ships, and if you had wanted the drive requirements relative to total ship capacity to remain constant even as hull capacity increases with tech, events, Helios Ore, and Hyperion Shrinkers, you'd need to add at least a MovesCapMass penalty to all of the things which affect hull capacity (and, as with the techs providing a reduction in MovesCapMass, you'd have issues where a bonus that matters a lot on a Tiny hull would have very little impact on a Huge hull).
Um... whether you increase the space required for each extra engine, or decrease the performance of added engines, you are diminishing return on investment.
As I understand things, the 'percentage of capacity' thing is not decreasing the performance or increasing the cost of additional engines. Rather, it's making each component require ~X% of the ship's hull capacity, or perhaps a flat X capacity + Y% of the ship's maximum capacity. I.e. you'd have a drive which requires, say, 10 capacity + 10% of the hull capacity, for a total of 12.5 capacity on a Tiny hull, 15 capacity on a Small hull, 17.5 capacity on a Medium hull, 20 capacity on a Large hull, and 35 capacity on a Huge hull. Additional components would continue to require those same capacities. This is not a diminishing returns system; it is simply a system which keeps the drive costs relative to hull capacity somewhat more similar across hull classes.