When I was working at Babbage in the mid-1990s, I remember that there were three specific computer games that were sitting in the column "soon", looking forever ̵
The original C & C was successful, but the sequels created a real dynasty for games. For this episode of military stories, we've arranged a good technical talk with Westwood co-founder Louis Castle (who also worked on the studio adventure Blade Runner ) to meet the challenges and problems of the Command & Conquer: Tiberian Sun The direct sequel of C & C and one of the most famous games in the entire series.
Road! Way! And there were challenges. One of the goals with "TibSun" (as the game is called when the guy is cool and in-the-know ) is that there is no limit or limitation on the number of players per unit allowed to build and control, and this goal is linked to a large string. More units on the screen meant more units for the computer to monitor and control, and that meant many and many ways to find a way – that is, understanding how to get these units from point A to point B No matter how simple the human eye can look like ("Just Go From Here!"), Discovering is actually a challenging programming challenge – one that, depending on the path, can be NP-full.
"Eureka!" the time came when Castle and their team realized that they should not build a perfect or even really awesome discovery of the road. On the contrary, they had to build a direction indicator that did not look stupid to the player. This meant that instead of having to solve effectively unsolvable NP-complete problems, they could instead apply multiple layers to mitigate the event and modify their basic algorithm. Eventually, after many repetitions, they reached a point where units acted (mostly) normally and did (mostly) what the player expected (most of the time).
The most interesting, at least for me, is that the discovery is Still is still a tricky problem even for modern games and just throwing more processor power to the problem does not really help so much – this is the curse of NP-complete problems. The repairs developed by Westwood developers are still applicable today – and will continue to be applicable until someone proves P = NP.
And if ever happens … well, algorithms for finding a way will be the least of our worries.