Abstract

A domain-separated semantic header ID commits to every consensus field except nonce. The canonical block ID and proof-of-work digest each include nonce under their own domains. The miner proves the complete transition once, searches many nonce-bearing headers and reconnects the winner through native PoW and exact boundary checks.

Where should the nonce enter the proof?

State is the authenticated set of current unspent outputs and consensus counters. HistoryStep is the recursive proof of one block's exact transition from its parent State to its child State. It certifies the expensive semantic work, while proof of work changes only a 128-bit nonce—the header value miners vary during search. If the nonce is included in the recursive statement, every candidate changes that statement and the expensive proof cannot be prepared before search.

A nonce-free semantic header

The construction defines three domain-separated hashes:

IdentifierDomainFieldsRole
Semantic header IDSEMHDR__Every consensus header field except nonceHistoryStep statement
Block IDBLOCKHDREvery header field including nonceParent link and canonical identity
PoW digestPOWHDR__Every header field including nonceComparison with the declared mining target

The semantic projection preserves the exact block-header field order and skips only the nonce. Mutation tests change every field independently: changing nonce must preserve the semantic ID; changing any other field must change it.

Prove once, search many

TemplateFix semantic fieldsTransactions, State, target, reward
ProofBuild HistoryStep πBind SEMHDR__ projection once
PoWSearch nonceHash POWHDR__ repeatedly

When a nonce meets the target, the node performs one native PoW check and verifies that the nonce-bearing header names the same proved parent and child State boundary as the semantic template. It then commits the complete block and terminal atomically.

Why the split does not free other fields

The recursive proof omits the nonce, not the mining target or any semantic transition field. Native acceptance requires the terminal's semantic ID and the candidate header to agree on every non-nonce field. The nonce-bearing block ID remains the chain-link identity.

Domain separation prevents the same header schedule from being interpreted interchangeably as a semantic projection, canonical block ID or proof-of-work digest.

Result

Mining boundary

Expensive proof work is performed once per fixed block template. Cheap nonce search may vary the only field deliberately excluded from that statement.

The pattern applies whenever a costly proof certifies a deterministic object and a separate search varies one explicitly non-semantic field. The acceptance rule must reconnect the winning search result to every field the proof fixed.