Posts

5 posts


Notes from building and measuring large native codebases. Every claim ships with its artefact — the spec, the diff, the commit, the test output. The failures are the interesting part.

All Posts

Done is where the evidence runs out

FourShades finished at 155 of 165 test ROMs. The ten that are left cannot be fixed without evidence that does not exist, and saying so is the interesting part.

FourShades

The screen works. The scoreboard barely moved.

The emulator can draw now, and you can play it. Fixing one systematic error corrected 44,000 wrong pixels and flipped exactly one test.

FourShades

Two ways a scoreboard can lie

The emulator went from 0 to 85 passing test ROMs. The interesting part is the two ways it nearly claimed more than it had earned.

FourShades

499 out of 500, on purpose

The CPU passes every SingleStepTests instruction but one. The one it fails, it fails deliberately.

FourShades

Starting the scoreboard at zero

Why I am building a Game Boy emulator in C++, and publishing every number it scores.

FourShades