Modularity
My thoughts as a new A-10C player, Flanker 2.0 "veteran":
I think DCS is on the right path. I feel a single "environment" is key to expanding the DCS product line. But I think the reliance upon the old assets, such as terrain, models, and textures dating back to the Flanker days is a secondary hurdle. These "old" objects, even today, have higher fidelity than many new products I've seen. Time will help with each, Nellis/Nevada is already in production. I see it only as a matter of time before objects get a stream of updates too. (I am inclined to help with this as I used to do 3d modeling from blueprints...for fun. Does ED evaluate player-made models for consideration in patches? What kind of poly counts are we looking at per LOD level?)
ED seems to be much of the way there, in regards to creating a "world" for objects to exist in, but in a sim like this, that is a daunting task for programmers and asset creation alike. A "sim environment" consists of boatloads of math in regards to physics and math I can't even start to begin comprehend. (Isn't Yo-Yo the genius behind the physics and math? I am still learning who does what). I don't include flight models because every vehicle, be it a tank, boat or plane will have its own dynamics since land water and air are quite different mediums to travel, but those dynamics must still interact with the rest of the world. Prop planes, helicopters and jets all fly via different means, but are all effected by lift, inertia, gravity and a myriad of other factors within the same "environment." Props pull the plane through the air while jets push, ...so to speak. Specific modeling for specific vehicles is still necessary. This is where I see specific vehicles getting their own modules within the DCS universe. KA-50 is a module, A-10C is a module, but both exist within a DCS environment. (And will hopefully one day be compatible for joint multiplayer ops.)
If/when ED succeeds in creating an environment in which physics and effecting factors like weather, pressures, atmospheric variances and who-knows-what-else are there for other modules to use, I suspect the creation of flight/dynamics models of specific vehicles can be easier to implement within the same environment.
I think this is ED's long term goal, a grand simulation in which smaller simulations will live. (And hopefully coordinate) So as far as longevity goes, I suspect we are only at the beginning.
And with this diatribe, I end my first post. lol
ED, godspeed and thank you.