Goran,
Eddie is right, it is what Panther has been working on; a code base now 25 years old. Part of the reason I joined is I said "this is the future of gaming". So, subordinates were not simply units, but intelligent agents. It is a beautiful system, and if no game had ever followed RDOA; that would have been a masterpiece.
I love what Gary has done here, but you really have to invest a huge amount of time. The Panther system scaled better than any game design out there, because of intelligent agents which create plans and orders for their own intelligent agents. Most games have a natural size, but Panther was mainly limited by computing resources.
Is it bad to plan a game which cannot be run? No, I don't think so. First, all RL greatest thinkers saw things before they could be realized. A game visionary is no different. Second, back in the day Moore's Law was in effect. So, from concept to execution, you could expect compute to increase 8X; then over the life of the game 8192X. So, you need not limit yourself with such a small vision.
Moore's Law is not actually dead. We are growing cores and storage; instead of clock frequencies. When I was in the MS/PhD program, they told us that the new compilers would be able to make code parallel. I was over hyped as clearly seen with games. But then AI was also way over hyped, but we have to admit AI has arrived; just not in the 1950s.
As for true agent based games, I think in just a year or two. You will see a CK style game, but that is not script driven, but LLM driven. People like to say, "if you want a challenge, PBEM"; now they are going to say, "if you want chance of winning, PBEM".
2021 - Resigned in writing as a 20+ year Matrix Beta and never looked back ...
