I’ve launched a crowdfunding campaign for OpenRecomp, an open-source project exploring deterministic static recompilation for software preservation.
OpenRecomp already has a PROVEN bounded RV32I synthetic baseline. MIPS32 is the second architecture: existing bounded MIPS32 fixtures pass their documented validation paths, but that does not establish arbitrary MIPS32 compatibility. Wider MIPS32 support therefore remains CANDIDATE.
The funded work focuses on ingestion/extraction, MIPS32 semantics and translation, bounded indirect control flow, native/WebAssembly repeatability, rights-safe fixtures and reproducible public evidence.
I’m particularly interested in technical feedback from the rev.ng community around MIPS control flow, calling conventions, ABI/compiler-pattern quirks, indirect targets and validation methodology.
Funding does not make MIPS32 PROVEN. Evidence does.
Campaign: https://www.crowdfunder.co.uk/p/openrecomp-mips32
Source/evidence: