This little rant has come about after playing some RSS/64k career games but the points I make are also applicable (less so, admittedly) to a stock game. I've found the reliance on IRL engines in any realism overhaul games highly irritating. There are always niches that I really want to fill but can't because of a dearth of RL options (this is most notable with small engines, lander engines, and engines with poorly represnted fuels like methane.)
To that end, I'd like to propose a (hopefully) reasonably easy to implement way that someone with some degree of modding ability might go about making a procedural engine mod as well as some game balance considerations.
Each procedural engine would consist of two meshes, an engine housing, and an engine bell. Housings would be a fixed shape--just scaleable. Housings would also include any turbomachinery, pumps, etc.
Only the bell would be procedural with 3 general options: surface, vacuum, and low-profile. Again, bells would be modified mostly by scaling.
This offers interesting game design considerations. For example:
-Low profile engines would have worse TWR and ISP (bells not ideally shaped, machinery compacted in awkward ways).
-In RSS, choices of pressure fed, turbopump driven, control over the number of ignitions, throttleability, etc.
-Integration with KCT and similar to reward desgning vehicles with a common lineage--each procedural engine would have to be discretely represented.
-Balanced tradeoffs between TWR and ISP would be interesting ways to distinguish engines
-Because stock is much simpler there's less need for a mod like this but if it's pretty enough there'll always be uses
I feel like this is a relatively low-effort way of getting procedural engines without a really involved procedural system. The real trick is game balance.