Hello Gorry, Thank you for the review.
On 27 Mar 2026, at 11:39, Gorry Fairhurst via Datatracker <[email protected]> wrote > "All configuration options SHOULD be clearly named to indicate that they are > private keys." - Are there actual requirements where this needs to not be the > case: if not, why not MUST? or if so, please explain those cases? Configuration options are at the edge of the scope of this document. An implementation can call their settings anything, or have any strange form of configuration system, and interoperability would not be affected. That said, we strongly encourage this best practice for all implementations. I would accept the argument that this should be non-normative language due to the scope. > "It is RECOMMENDED that implementations use a bounded exponential backoff > strategy," - I strongly agree, please could you add a sentence to explain why > this is important. We've clarified this in https://github.com/mxsasha/nrtmv4/commit/1f748afaf97956188658192dd2b0b0946361177b by adding: > This prevents unnecessary Snapshot imports on clients, and can prevent > clients overwhelming a server on transient failures. Sasha _______________________________________________ GROW mailing list -- [email protected] To unsubscribe send an email to [email protected]
