Joe Theis | Wishlist Infiniwar

YouTube video by LT Gaming 15 Upcoming Roguelikes & Roguelites You Need to Know About!

Infiniwar was featured in a recent video along with some killer roguelikes. Always wild to see the name of my little project mentioned alongside titans like Slay the Spire! Thanks for the mention x.com/THPauthor

www.youtube.com/watch?v=Z0sp...

Mystery solved. Thanks @gigatank3000.bsky.social for the tip!

It definitely wasn't steam traffic (over roughly the same period), which is noisy but mostly flat. So I assume something external?

Suddenly got a spike of over 20-50x my usual wishlists in the past 2 days and I have no idea why. I didn't do anything. Anyone know where this might be from?

YouTube video by Joe Theis Playtest Plans and Busy Schedules | Channel Update

Quick channel update!

youtu.be/ajknGjbJ_tA

This is a very safe bridge and I definitely know what I'm doing. Yessir.

(Showing wireframe) I initially made the system with 10m squares as the basis, but that way to big for realistically scaled roads. 5m squares is better. But I may use 10m for runways and <5m for paths and such.

Meant to do this sooner but at least it was easy to add. Spline-based networks like Roads can now have variable width geometry.

YouTube video by Joe Theis Creating Terrain-conforming Roads, Rails, and Bridges for an Indie RTS | Infiniwar Devlog #9

New devlog! This one expands upon the topic I've been covering in smaller posts recently: Procedural Road Generation

youtu.be/BIdrL-j_Di8

That's a thicc intersection right there.

Vert visualizer coming in handy again as I work to make these intersections thicc.

Procedural crosswalks (no textures!) The stripe interval is unintentionally related to the width of the road, but I kinda don't mind it TBH. What do you think?

Procedural intersection progress. Lots left to do though. Still need to add the side "skirts" to match the roads, use more internal verts for terrain conforming, and figure out what the heck to do about the vertex coloring.

Ish 😅

More glorious programmer art. Reminds me of a spider web. 🕸️

An important breakthrough in the procedural road tech!

Wish I had it back when I started on procedural buildings. It would have saved so many headaches! At least I will have it in the future when I improve them.

I've actually had this for a while, but I finally figured out how to turn it into a standalone script today. This means I can easily attach it to any object with a mesh, like even this built-in unity sphere. But it's mainly useful for procedural meshes.

Procedural mesh debugging tech I've been using. It's very easy to lose track of vert or tri indexes in complex meshes, breaking in incomprehensible ways. This tool has saved me so much time, scrubbing through the mesh to find the exact point where it's breaking.

I haven't done the lane extension/intersection yet, but I did get my sorting issue figured out to correctly determine neighbors.

bsky.app/profile/joet...

Going to need some additional tessellation to make them terrain conforming like the splines though. Alternatively I could force terrain under intersections to always be flat, but that seems less flexible, and I think it should be doable.

Works with different sizes of roads too.

Some progress on procedural intersection geometry. Properly sorting the incoming/outgoing segments, allowing the neighbor connection tris. Center sphere in increasing size to check ordering. Small white spheres on the edge of each road to check left vs right consistency.

Ahhhh! Love this write up, thank you! This is similar to what I was already doing, but the extending the lanes beyond the intersection idea was not on my radar yet. I will definitely give it a shot and let you know how it goes.

As for the UVs, I'm not actually sampling a texture there. It's vertex colors as data channels, and a shader to turn them into road markings/curbs etc. Theory is that I can use it to procedurally alter the markings in code. No idea if this is smart but I'm trying it (allergic to textures I guess)!

Still figuring it out TBH. Current plan is to check for overlaps and progressively back away from the center 1 "tile" at a time until no overlap. But that sounds slow so maybe there's some clever math to do.

Also look at this majestic artwork.

Some progress on figuring out how to approach procedural intersections.

Ok that was faster than expected (obviously the intersection itself is going to need a lot of special work, especially to reconcile different types of connecting segments).

It doesn't look like much yet, but good things start small.