Annex A — Canonical Serialization

This annex defines mandatory canonical serialization requirements for all Trust-State evaluation inputs, rule definitions, and conformity artifacts.

1. Scope

Canonical serialization ensures deterministic byte representation of structured data prior to hashing, replay validation, or artifact issuance.

All systems claiming alignment with the Trust-State Standard SHALL implement canonical serialization exactly as defined in this annex.

2. Encoding Requirements

• All serialized data SHALL be encoded in UTF-8.
• Byte Order Mark (BOM) SHALL NOT be included.
• No trailing whitespace SHALL be permitted.
• No indentation or formatting whitespace SHALL be included unless explicitly defined.

3. Object Member Ordering

JSON object members SHALL be ordered lexicographically by key using Unicode code point ordering.

• Ordering SHALL be deterministic.
• Nested objects SHALL apply the same ordering recursively.
• Arrays SHALL preserve original element order.

4. Data Normalization Rules

• Strings SHALL be case-sensitive and preserved exactly.
• Numeric values SHALL NOT include leading zeros.
• Boolean values SHALL be lowercase (true, false).
• Null values SHALL be serialized as null.

5. Canonical Form Definition

The canonical form of a structured object is defined as the UTF-8 encoded byte sequence resulting from:

1. Recursive lexicographic ordering of all object keys.
2. Removal of non-semantic whitespace.
3. Preservation of exact string and numeric representations.

The resulting byte sequence SHALL be used for all hash computations defined in subsequent annexes.

6. Determinism Requirement

Given identical structured input, canonical serialization MUST produce a byte-identical output across independent implementations.

Any deviation in byte sequence SHALL constitute non-conformance.

7. Relationship to Hashing

All hash derivations defined in the Trust-State Standard SHALL operate exclusively on canonical serialized byte sequences produced under this annex.