Hello, Thank you to the authors and working group for the continued work on draft-ietf-jose-hpke-encrypt.
I would like to offer several implementation-focused observations from the perspective of operational governance, long-term evidence preservation, and cross-stack interoperability between JOSE-, COSE-, and transparency-oriented systems. 1. Explicit guidance on authenticated metadata traceability The draft correctly focuses on cryptographic confidentiality and integrity, but implementers operating transparency or audit systems may benefit from clearer guidance regarding: • which metadata is expected to remain externally observable, •which fields are integrity protected, • and which fields may safely participate in external evidence or transparency logs. In practice, many deployments will combine: • JWE/HPKE, • SCITT-style transparency receipts, • WORM archival systems, • and externally verifiable audit traces. A short operational considerations subsection describing “safe-to-log” vs “sensitive” metadata classes could reduce implementation ambiguity. 2. Interoperability considerations between JOSE HPKE and COSE HPKE As parallel JOSE and COSE HPKE efforts mature, implementation divergence risk may emerge around: • algorithm identifier handling, • recipient structure interpretation, • key lifecycle semantics, • and canonicalization expectations. A non-normative interoperability appendix or implementation note referencing alignment expectations between the JOSE and COSE ecosystems may help prevent avoidable fragmentation for dual-stack implementers. 3. Evidence preservation and future cryptographic migration Because HPKE deployments may persist encrypted artifacts for extended periods, it may be useful to acknowledge: • long-term cryptographic agility expectations, • future PQC migration considerations, • and preservation of sufficient algorithm/context metadata for future validation or reprocessing. This is particularly relevant for archival, governance, regulated infrastructure, and transparency-oriented deployments. These comments are intended as implementation-supportive operational observations rather than objections to advancement. Thank you again to the authors, reviewers, and working group participants for the effort involved in advancing the specification. Best regards, Garth R.T. Smith Project Ari / PoT Research Calgary, Alberta, Canada
_______________________________________________ jose mailing list -- [email protected] To unsubscribe send an email to [email protected]
