At 9 PM GMT-5 (about 30 minutes from now), I’ll be going over the river propagation algorithm so far and hopefully pushing it to something closer to final form.
I’ll spend the first part demonstrating (visually) the data the underlies the algorithm and my motivations, etc…
I think the stream turned out pretty well and I covered a lot of ground. Anyone with a remote interest in procedural world generation may find that this tickles a lot of brain cells.
I’ve gone back and added chapter markers for those who like to skip around. Preview:
0:00 Introduction
0:50 Quick detour about region mapping
3:20 Rivers Importance
3:30 Embedded Mapper tool
6:14 Land meta-data
7:25 Visualizing water levels
10:00 Water flow propagation
18:20 Water sources
22:00 Cost map visualization
23:50 “Lame” direct to sea rivers
27:00 Slightly smarter river flow
33:30 My problem/goal
35:30 Cost-based approaches
40:00 Processing order
41:25 High level approaches
45:45 Trying various things
54:10 Bug or math?
1:03:00 Temptingly close
1:10:30 Wrapping up