P-152: Virtual World Concept Update 170: Milestone 2: Final Work: Part 1

by PhoenixGames on Sep.04, 2016, under Astral Realms, P-0, Software and Games, Spiritus Astrum

I ran into a few minor issues while implementing my latest solution. It seems that the collision mesh and some of the subdivision code was not updating correctly when the player moved. I solved this fairly easily, but I needed to move my offset code back to the Tier subdivision functions, and out of the rendering function, which means that I may need to implement the same code multiple times for the other tiers.

There also seems to be an additional issue with the noise scale functionality. As the player moves beyond 10k, the scale of the noise seems to change, resulting in sharp, jagged ridges on the terrain.

I have, however, successfully implemented the ability for the player to travel in any direction. I was using a simple movement on the x-axis for testing purposes, but I have now added complete vector support for the player, so that the terrain code is essentially working, and feature complete, even if it is in a very early stage, and requires a lot more work.

Due to the time-consuming nature of the current problem, I have decided that the best thing to do is to temporarily postpone the floating-origin solution, as well as some additional debugging, for the moment. I will likely fix these issues in a “MS 2.5”, before I begin working officially on MS3. I need to spend some time researching the floating point issue in particular, because otherwise I will just end up hacking together a solution that won’t work in the long run.

If I had known that the floating origin implementation for the spherical terrain was going to take this long, I probably would have postponed it until a later stage anyway.

For MS2, I still need to clean up the code, remove unnecessary/test code, add comments, and possibly do some refactoring. I also need to properly document MS2, so that I can return to it easily after a few months. Finally, I need to prepare a testing environment for the main floating origin problem, including some sample data from MS2, so that I can continue to research and work on that problem in isolation, it may be easier to fix this way.

The primary focus of MS 2.5 then will be the floating origin solution, as well as some optimisations that I didn’t get into MS2, the terrain painter fix, shape replicator objects, better script variable integration, and more testing and debugging.

MS3 will feature the introduction of simplex noise and an actual high-detailed terrain surface, currently it is almost entirely flat. I will also add full multiplayer support. I did some testing in MS1 or MS2 for this, and it worked well, but I will need to spend quite a bit of time making it work with the changes to the terrain, especially the floating origin system, since this is known to cause issues with multiplayer games. I will also add instancing (allowing multiple planets, moons, large asteroids, etc to be rendered in the same scene) and some procedural support to the terrain.