Re:Procedural Dungeon Generation Algorithm
by carlisle_ on Friday September 04, 2015 @03:48PM
(#10172522)
I really like the idea to just use a physics engine as a means of randomizing the layout. Smart, succinct, and easy to implement.
[ Parent
| on HN ]
Re:Procedural Dungeon Generation Algorithm
by neolefty on Saturday September 05, 2015 @12:02AM
(#10173914)
I delighted in this article because of its opinions about what is easy and what is hard. Physics engine: easy; of course you have one; everyone has a physics engine. Graph data structure & algos: suggests a few places to find one.
As an older programmer who is in the midst of retooling, I would assume the opposite: graphs easy, physics hard. But that's what retooling seems to be all about -- finding the new "easy" and learning it.
Re:Procedural Dungeon Generation Algorithm
by yoklov on Saturday September 05, 2015 @10:49AM
(#10174927)
A physics system good enough to support this is probably ~200loc. Probably less if you cut out everything extraneous. All it is is AABB collision detection+resolution, and then some update code.
I think the graph + graph algos would be trickier most of the time, but that could just be since I've written the above code many, many times.