I think it might be quite simple and very good change to have a logarithmic research costs instead of linear from the size.
Smaller components and hulls are really not much simpler then larger ones, and R&D times for missiles, planes and battleships are nearly equal really.
(I understand that it can be made as home rule, but it's very micro-heavy to implement this home rule every time - we must remember separately to not use this, and this, and this component and GF unit until next year, and keep all research labs busy with some research of straw to load population and economics, and then remember to reward or not reward some scientists... Umpffffff. My hope is that Steve likes to conduct stories and read details as they flow, not calculate and script the very small detail unassisted!)