Road Meshes Are Delayed, But For A Reason

If you have been following Maps4FS, you probably know that road mesh generation is one of those features I have rewritten more times than I care to admit. What you might not know is that I am about to do it again.

The short version: I am doing it again. And this time I think I finally understand why the previous attempts kept falling short.

The First Attempt Was Not Bad, Just Wrong

The original road mesh implementation actually worked. It produced something that looked like roads, and for a while that felt like enough.

But the more I looked at it the more flaws I found. The geometry was awkward, edge cases were everywhere, and the whole thing was held together by assumptions that did not hold up under real-world map data. It was functional in a narrow sense and broken in a broader one.

The Second Attempt Was Better, Still Not Enough

So I rewrote it. The second version addressed most of the obvious problems. The quality improved noticeably, and for a while I was reasonably happy with it.

Then I started comparing the output to what real game worlds actually look like, and that optimism faded pretty quickly. There was still a fundamental gap between what I was generating and what would genuinely feel natural and accurate on the ground. Better was not the same as right.

This Time It Is A Full Rebuild

That is why I decided to throw out the existing approach entirely and start over from a clean slate. And this rebuild is happening in MapToPlay, not Maps4FS — which means I finally have the chance to design it properly from the ground up rather than patch over something that was already compromised.

The core difference this time is that I am moving away from the tiled approach. It seemed reasonable at first, but tiling is essentially the wrong mental model for roads. Roads are continuous structures that should flow across a map, not stitched-together segments fighting against grid boundaries. I want MapToPlay to get something that actually respects that.

Where Things Stand Right Now

I will be honest: I am in the middle of it and I cannot yet say with confidence that I see the finish line. There are no exciting screenshots to share, no obvious breakthrough moments to point to. Just a lot of groundwork and experimentation.

I do believe the direction is the right one. But I also know that saying "I think this will work" is very different from having something solid to show. Right now I am somewhere between the two.

Road meshes are taking longer than I wanted. But releasing another version that is almost good enough does not interest me. I would rather take the time to get it genuinely right.

More updates as things come together. Thanks for sticking around while I figure this out.

Stan, Developer of MapToPlay