Thanks for the detailed review. Comments inline below. ## On urgent security conditions
CURRENT: This document specifies that outside of urgent security fixes, and the exceptions listed in Section 4, no changes will be approved for TLS 1.2. Who will make the call about what is “urgent”? Is it the TLS WG? Else? What about extensions that may be required by applications defined in other WGs? A separate sub-thread (from EKR and me) says: yes, just TLS WG. Who else is chartered to do so? :) As for other WG extensions, the only ones that have *EVER* come up are exempted ALPN and exporters. # Abstract ## Reword for better clarity OLD: Use of TLS 1.3 is growing and fixes some known deficiencies in TLS 1.2. NEW: Use of TLS 1.3, which fixes some known deficiencies in TLS 1.2, is growing. Yes, fixed thanks. ## I think “for TLS 1.2-only” would be more accurate as some of these are applicable to TLS1.3 as well. Consider updating accordingly: CURRENT: This document specifies that except urgent security fixes, new TLS Exporter Labels, or new Application-Layer Protocol Negotiation (ALPN) Protocol IDs, no changes will be approved for TLS 1.2. I do not understand this comment. If you mean a new extension is added that *would* be usable in TLS 1.2, we’re saying that it is not defined for TLS 1.2 even if it would “just work.” Is that your point? # Introduction ## Reword for better clarity OLD: Both versions have several extension points, so items like new cryptographic algorithms, new supported groups (formerly "named curves"), etc., can be added without defining a new protocol. NEW: Both TLS versions have several extension points. Items such as new cryptographic algorithms, new supported groups (formerly "named curves"), etc., can be added without defining a new protocol. Changed, thanks. As a side note on “etc.”, I’d like to check if we should be more explicit here given that we have notes in the registry such as: “Although TLS 1.3 uses the same cipher suite space as previous versions of TLS, TLS 1.3 cipher suites are defined differently, only specifying the symmetric ciphers and hash function, and cannot be used for TLS 1.2. Similarly, TLS 1.2 and lower cipher suite values cannot be used with TLS 1.3.” I think that is a confusing level of detail for this document. Except for cipher suites the other registry values are used unchanged. # Section 2 ## Provider examples of “huge impact” mentioned in this text: CURRENT: Cryptographically relevant quantum computers, once available, will have a huge impact on RSA, FFDH, and ECC which are currently used in TLS. I would rather not do so, because I want to avoid explaining Shor’s algorithm, factoring vs the discrete log problem, and the like. Doing this would muddy the waters for the clear statement we want to make. ## On the various NIST citations: Are there any other similar pointers to list for non-US regions? There does not seem to be. At a recent IETF side meeting, several countries said “we’re following the NIST algorithms.” As that is the only standard published so far, it seems okay. # Section 4: ## I think the IANA registries should be authoritative here, not the RFC. CURRENT: No registries [TLS13REG] are being closed by this document. This wording came from discussions with IANA about what to say. ## Call out this is about TLS registries: OLD: No registries [TLS13REG] are being closed by this document. NEW: No registries in TLS registry groups [REF] are being closed by this document. I will change to “No TLS registries …” ## Not any random registry, but TLS: OLD: Any registries created after this document is approved for publication should indicate whether the actions defined here are applicable. NEW: Any TLS registry created after this document is approved for publication should indicate whether the actions defined here are applicable. The whole document is about TLS registries, but sure.
_______________________________________________ TLS mailing list -- tls@ietf.org To unsubscribe send an email to tls-le...@ietf.org
