Skip to content

Latest commit

 

History

History
164 lines (91 loc) · 1.81 KB

File metadata and controls

164 lines (91 loc) · 1.81 KB

GLYPH_COMMITMENT_CLAIMS_V1

Status: DRAFT

Purpose:

Define the exact statements that a GLYPH Commitment Proof can prove.

Rule:

Every claim must be:

  • deterministic
  • byte-exact
  • independently verifiable
  • machine-checkable

CLAIM TYPE 1

PRESENCE

Statement:

Pattern P exists in corpus C at offset O.

Verifier learns:

  • corpus commitment
  • offset
  • proof

Verifier can conclude:

TRUE or FALSE


CLAIM TYPE 2

ABSENCE

Statement:

Pattern P does not exist in corpus C.

Verifier learns:

  • corpus commitment
  • proof

Verifier can conclude:

TRUE or FALSE


CLAIM TYPE 3

EXACT COUNT

Statement:

Pattern P occurs exactly N times in corpus C.

Verifier learns:

  • corpus commitment
  • count
  • proof

Verifier can conclude:

TRUE or FALSE


CLAIM TYPE 4

INTERVAL

Statement:

Pattern P maps to suffix interval [L,R).

Verifier learns:

  • interval
  • proof

Verifier can conclude:

TRUE or FALSE


CLAIM TYPE 5

CORPUS IDENTITY

Statement:

Corpus fingerprint F corresponds to corpus C.

Verifier learns:

  • fingerprint
  • proof

Verifier can conclude:

TRUE or FALSE


CLAIM TYPE 6

INDEX IDENTITY

Statement:

Suffix array SA belongs to corpus commitment C.

Verifier learns:

  • commitment
  • proof

Verifier can conclude:

TRUE or FALSE


NOT SUPPORTED

Semantic similarity

Embeddings

Meaning equivalence

Approximate matching

Fuzzy matching

Ranking

LLM interpretation


CORE AXIOM

GLYPH proves bytes.

GLYPH does not prove meaning.