Byte identity and meaning
Separate exact file identity from semantic equivalence when reviewing a model update.
Research note · November 14, 2026Research question
When do different bytes represent a meaningful change, and which question should a file checker answer?
Method
- Create two configuration files with identical parsed values but different whitespace. Keep both originals and calculate their digests without normalization. A byte-level check should report the files as different.
- Create a second pair where a threshold changes while the filename remains constant. Compare the raw file result with a separate, explicitly labeled comparison of parsed values. Do not replace the byte-level result with the semantic comparison.
- Document the identity contract before testing: exact bytes, a canonical representation, or selected fields. Keep each contract versioned so an unchanged label cannot conceal a changed comparison rule.
Records to retain
- Original byte sequences
- Digest algorithm and full digests
- Parser and comparison-policy versions
- Expected result for each pair
Source basis
RFC 6920 describes content naming with hashes, while RFC 8785 defines a particular canonical JSON representation. These answer different identity questions. A canonical digest is meaningful only when both sides apply the same rules.
Interpretation & limits
Neither a matching digest nor equal parsed values establishes that a configuration is appropriate for its task. This protocol contains no measured Vraelis result.
Next experiments
- Add Unicode and numeric representations to the fixture set.
- Review whether users can distinguish raw-file changes from equivalent parsed content.
References
Technical reference
Implementation details supporting this note, drawn directly from the maintained documentation. These are technical references, not additional papers or new experimental results. Expand a topic to inspect its method, fields and limits.