Imagine you're renovating a house and the building code says every door must be a standard 80×36 inch slab. Works great for bedrooms. But you also need a garage door, a bank vault, and a cat flap — and the code insists they're all the same slab. That's the state of post-quantum cryptography for blockchains right now. NIST standardized ML-DSA, SLH-DSA, and Falcon/FN-DSA as general-purpose signature schemes. The blockchain world looked at them and said: close enough, we'll swap them in. Lie and Marsh argue that 'close enough' is a catastrophic engineering posture when the signature layer must simultaneously serve transaction authorization, consensus quorum certificates, light client verification, and committee rotation — each with different size, latency, aggregation, and accountability requirements. The paper's core contribution is a formalization of what blockchain signatures actually need to do, decomposed into nine properties: canonical bytes, priced invalid-input rejection, stable transaction identifiers, hybrid downgrade resistance, public aggregation, merge semantics, accountable signer evidence, forward-secure committee rotation, and light-client consequences. These aren't wish-list items — each maps to a concrete failure mode. Stable transaction IDs prevent malleability attacks. Priced rejection stops DoS via garbage signatures. Forward-secure rotation ensures that compromising a past committee key doesn't retroactively break finality. The paper argues these are requirements, not nice-to-haves, and that no current post-quantum scheme satisfies all of them simultaneously. The evaluation is genuinely broad. Fourteen signature schemes are benchmarked: ML-DSA, SLH-DSA, Falcon/FN-DSA, HAWK, MAYO, SNOVA, UOV/QR-UOV, FAEST, SQIsign, LaBRADOR-Falcon, Squirrel, Chipmunk, and DKKW/LeanSig. Each is tested against three deployment profiles — Bitcoin (low throughput, high decentralization), Ethereum (medium throughput, validator sets of ~1M), and a Sei-Giga-style high-throughput BFT chain (aggressive latency targets). The headline result is negative in the most useful sense: none of the 14 is a drop-in replacement. ML-DSA signatures are 2–4× too large for high-throughput consensus. SLH-DSA verification is too slow. Falcon has canonical-bytes issues. The lattice-based aggregate schemes (Squirrel, Chipmunk) show promise for quorum certificates but lack maturity. SQIsign is compact but verification is orders of magnitude too slow for consensus timelines. The analogy to hash-function engineering is the paper's sharpest intellectual move. The cryptography community didn't solve hash-table DoS or arithmetized proof systems by picking a better hash function from the existing menu. It built SipHash and algebraic hashes — domain-specific primitives designed for the actual threat model. Lie and Marsh argue post-quantum signatures need the same treatment: blockchain-specific profiles that specify aggregation behavior, rejection cost bounds, and committee-rotation semantics as first-class design constraints, not afterthoughts bolted onto a generic signing API. The integrity posture is solid for a systems-analysis paper. This isn't claiming a new scheme — it's claiming a new requirements framework and showing that existing schemes fail to meet it. The benchmarks use three distinct deployment profiles with different stress parameters, which prevents the usual trick of picking the one profile where your favorite scheme looks good. The limitation is that the evaluation is analytical and simulation-based, not deployed. No on-chain experiments, no real validator sets under load, no adversarial testing of the rejection-cost bounds. The practical implication is uncomfortable for every blockchain project with a 'post-quantum roadmap' that amounts to 'swap in ML-DSA when NIST says go.' The paper says that plan is incomplete at best and dangerous at worst. The missing piece isn't a better signature algorithm — it's a consensus-layer signature specification that the cryptography community hasn't written yet. Until that spec exists, every migration plan is building on sand.
