From a focused web-research pass — flagged as PulseEngine's highest-leverage open research lead.
The gap: the WebAssembly Component Model has product momentum (Microsoft Wassette) and a security narrative but no soundness theory yet. The position paper The WebAssembly Component Model (Menon, Microsoft; Wagner, Fastly — WAW/POPL 2025, https://popl25.sigplan.org/details/waw-2025-papers/4/The-WebAssembly-Component-Model) argues syntactic type soundness is insufficient — it needs new theorems "inspired by multilanguage semantics" — and the formal spec / reference interpreter / test suite are listed as future work. The people who defined the problem have only posed it.
Why meld: a soundness theorem for component composition / linking (lifting & lowering across the Canonical ABI) is exactly meld's domain (static component fusion).
Opportunity: define + mechanize a soundness result for fused/linked components — PulseEngine could genuinely lead and publish here. Build on WasmCert-Coq / Iris-Wasm (https://github.com/WasmCert/WasmCert-Coq) rather than re-deriving Core Wasm semantics.
(Tracking issue from research; not scoped to a release yet.)
From a focused web-research pass — flagged as PulseEngine's highest-leverage open research lead.
The gap: the WebAssembly Component Model has product momentum (Microsoft Wassette) and a security narrative but no soundness theory yet. The position paper The WebAssembly Component Model (Menon, Microsoft; Wagner, Fastly — WAW/POPL 2025, https://popl25.sigplan.org/details/waw-2025-papers/4/The-WebAssembly-Component-Model) argues syntactic type soundness is insufficient — it needs new theorems "inspired by multilanguage semantics" — and the formal spec / reference interpreter / test suite are listed as future work. The people who defined the problem have only posed it.
Why meld: a soundness theorem for component composition / linking (lifting & lowering across the Canonical ABI) is exactly meld's domain (static component fusion).
Opportunity: define + mechanize a soundness result for fused/linked components — PulseEngine could genuinely lead and publish here. Build on WasmCert-Coq / Iris-Wasm (https://github.com/WasmCert/WasmCert-Coq) rather than re-deriving Core Wasm semantics.
(Tracking issue from research; not scoped to a release yet.)