This week was a weird mix of debugging, refactoring, and accidentally trying to populate an entire continent with animated trees. Progress has been solid, even if half of it came from staring at yellow error logs for way too long.


🌐 Network Mode & Visual Glitch Testing

Started off testing some Three.js changes and ran into this trippy web of broken geometry — mesh vertices exploded into what looked like a virtual spiderweb. Totally unintentional, but it made for a wild visualization. Turns out I had a data mismatch between terrain layer coordinates and node link generation for visibility fog.

Once that was fixed, I was able to refine how visibility lines are rendered between player and pet — and better control what’s shown in the minimap.


🌲 Vegetation Loading & Overkill

I added support for terrain-specific vegetation: palm trees on coasts, baobabs in savannahs, conifers in mountain regions, etc. The system picks a biome type and populates based on tile groups — but somewhere in the math I overshot.

One of the debug logs revealed I’d generated:

  • 13,144 visible tree objects
  • 5,522 on jungle tiles alone
  • Expected: ~2,000 max

So yeah — technically it works… just not optimized yet.

Also had to fix a rendering crash from missing tree instancing logic. Was trying to clone from an empty mesh and forgot to verify whether baseTree.children[0] even existed. Rookie mistake.


🌥️ Cloud Layer and Depth Testing

I’ve also been messing with basic volumetric clouds. They’re just mesh groups floating over tile clusters for now, but they add depth to the otherwise flat skybox. The performance hit is minor, and it gives some of the jungle zones a cool, atmospheric vibe.


🛠️ What’s Next

  • Rewriting terrain generation to limit overpopulation
  • Adding dynamic fog-of-war shrink/grow based on elevation
  • Building a performance dashboard to avoid future 13,000+ asset loads
  • Starting initial UI sketch for faction-specific missions

This was one of those dev weeks where the codebase got a lot stronger behind the scenes, even if it doesn’t look like a huge visual jump.

Still working solo, still doing this in weekend sprints — but the foundation is growing more stable every time I break it and rebuild it smarter.

Back soon with a cleaner map, leaner assets, and hopefully way fewer trees.

—Brian

Facebook
Twitter
LinkedIn