there is a con people tend to gloss over, the editor files are the game files. larger objects demand larger chunks of memory, obsfuscation becomes an issue, extra miles walked for optimization. there is a cost for having editor functionality at runtime.
pro is, this isn’t necessarily a bad thing. the concept of game systems wielding the power of bpy is absolutely nuts, the workflow is speedy and flexible. you get that in exchange for performance.
an interactive component, bge or otherwise, packed with blender isn’t going to be the best option for everybody but is still a pretty cool thing to have, just not going to happen anytime soon. there is simply no interest in developing such a thing right now, that’s what i can infer. we have upbge so i’m a-okay with that, itd be nice if they hadn’t lied to our faces though. dont get my hopes up for nothing maan.
the name of the game is write me an importer. convert these blender datablocks into something unity can read. now do the same for unreal. now urho. now godot. and so on and so on.
honestly i think engine developers should adopt a model more alike openBRF. you pack your data in a very specific way, provide an editor for the small tweaks, everybody is happy. but naaah, lazy .objs for everybody and rebuild every last material. the nerve on these people!