BODY: Decompiling a Nintendo 64 game used to be a multi-year pilgrimage. Chris Lewis did one in 84 days.
On August 26, 2026, the engineer announced on his blog that his decompilation of Snowboard Kids — the 1997 N64 racer-meets-snowball-fight — had hit 100%. Every function in the game now exists as human-readable C source code, and compiling that source produces machine code byte-for-byte identical to the original cartridge. That last part is the whole game: a "matching" decompilation isn't an approximation or a clean-room rewrite, it's a proof that the recovered source is functionally the same program Racjin's developers shipped nearly three decades ago.
The speed is what turned heads. The landmark Super Mario 64 decompilation took a distributed community roughly three years; The Legend of Zelda: Ocarina of Time took longer still. The difference now is tooling. A decade of community infrastructure — automated disassemblers that split ROMs into source files, pseudo-C generators that produce a rough first draft of each function, and the permuter tools that brute-force compiler flags until output matches — has compressed work that used to be done by hand, instruction by instruction.
What matters downstream is what the source unlocks. Matching decompilations are the foundation layer for native PC ports, widescreen and high-framerate patches, netplay, and the kind of preservation work that doesn't depend on emulator accuracy. Projects built this way have historically occupied a legal gray zone — the source is newly written, but it describes copyrighted software — which is why most decomp teams distribute code only, never ROM data.
The insider take
Japanese coverage of decomp projects tends to run through tech outlets like Gigazine rather than gaming media, and the framing here is telling: it's an engineering story, not a nostalgia story. Snowboard Kids was a modest Atlus-published title that never carried Mario 64's cultural weight in Japan, and rights to the series have drifted between companies over the years, leaving it in the commercially dormant limbo that a lot of late-90s mid-tier Japanese games occupy. That limbo is exactly why the preservation community picks these titles — nobody is going to remaster them, and the publishers least likely to authorize a port are also the least likely to notice.
Originally reported by はてなブックマーク (Japanese).