Mister Blastman, on 13 February 2016 - 09:09 PM, said:
Well I believe anything is possible if you put your mind and effort to it. It sure beats waiting three years for them to fix an existing map with glaring problems on it.
I do too!... but in PGI's case regarding the task being asked of them, there's a couple caveats though. Those being infinite time and resources. Things that PGI (or really, any game) would desperately require to pull off not just procedural generation, anyone can do that, but to do it in a way that would work well, and specifically, with MWO's slow-paced team-based 24 player match-style no-respawn gameplay.
Just thinking about procedurally generating a well-balanced algorithm, that doesn't stagnate, is free from redundancies, etc. makes my head hurt, and it would be a world-breaking record if PGI pulled it off. They would have to define, literally, everything. And run thousands of tests. It's not going to be a singular algorithm, it's going to be several, each packed with a crazy amount of stipulations, right down to the redundancies (like you pointed out, they aren't a bad thing, they do have their place, but it needs to be controlled). You really don't want one side being better than the other... and in fact, you would really want the same amount of control as you would when you make the map yourself. So what makes a good/bad map in MWO? The problem is that the answer depends on who you ask, and when you ask them.
I'm open to the idea of procedural maps, but I think their realistic involvement in MWO's map generation would be limited to generating a decent starter base map for PGI to build off of, and have it tuned by hand to fit the game.
To take a step into the near future, I think it would be VERY cool (because it would be amazing if PGI made it just this far), if PGI built naked base designs/templates for maps with no assets, and the procedural generator could mix-and-max assets to build different visually varied maps. Such as Caustic Valley could turn out to be a snow-land. Obstacle nodes are placed around the map, and the generator would pick assets (buildings, trees, rocks, or nothing) that match the node properties (large obstacle? medium? small?) and populate the map accordingly.
This way PGI can pump out map templates, and leave the generator to applying an asset tileset to the map. I'd say that's probably more realistic, than a full-blown generator. River City might be River City one time, but the next it might be a Frozen City. or a mix of Forest Colony trees and Viridian Bog terrain. And there may, or may not, be a citidel. It could be this huge mountain, or nothing.
*shrug*
I spoke a bit more than I thought I would... sry.
Edited by MoonUnitBeta, 13 February 2016 - 09:43 PM.