A list of open problems in how the SNES works:
- What exactly happens in the DMA crash bug on the v1 CPU
- What is the exact timing for the garbage that appears on the screen when switching modes mid frame
- How is the sprite data cached for each scanline and is there a way to abuse the timing to inject data into OAM midframe
- Are there any unexpected interactions with ExtBG that haven't been tested? For example I am not sure if anybody tried Direct color mode with ExtBG yet (probably won't work, but who knows)
- What are the exact decay rates for the MDR
- Emulating Coprocessor bus conflicts
- What about unused pins on the PPU, CPU, and APU -- Even though a game can't use them they are worth exploring IMO
- can we predict the APU crosstalk 16 bit bug?
- What happens to the very first pixel on the scanline in Hi-res Math?
- Various PPU registers still need to know when writing to them is effective.
Most of these problems require access to a logic analyzer.