2 May 2023

Previous journal:	Next journal:
0071-2023-05-01.md	0073-2023-05-03.md

Raybox

Map grid boundary glitch fixed: I was incorrectly assuming rayDir>0 would lead to partial being in the range [0,1) but it's actually (0,1], i.e. it needs to be able to reach 1.0 but I didn't provide enough bits to allow a 1.0-0.0 result of 1.0.
Quantified movement speed; 3 different options, but only the middle "walk" speed is used for now.
Sim:
- Changed lock overrides and added momentary (one-shot) option.
- Started adding design inspection to On-Screen Display.

Can't get the design to work properly on DE0-Nano. Is it a timing issue or synthesis difference? Maybe need to find out what simulation/verification tools we can use with Quartus.
lzc could be an issue for Quartus; huge series of if() statements might perform badly compared to casez. I've decreased the bit depth from 32 (Q16.16) to 24 (Q12.12) so I would expect it to improve overall. Maybe try going back up to Q16.16, and otherwise try a git-bisect.

Need to see how lower FP precision goes when we bring in rotations of the camera. In that case, stepping the viewplane might not go so well, or might not need to be a clean /512 or /1024 step.
Put FPS counter in sim main window.
Add option for basic text that the design can render.
Create a model for loading the initial state of the "game" into the design, e.g. player position and direction, movement speed, and maybe more dynamic parameters later. Can we make it asynchronous? Should we make the "host" do it following a frame "tick"?