This archive is retained to ensure existing URLs remain functional. It will not contain any emails sent to this mailing list after July 1, 2024. For all messages, including those sent before and after this date, please visit the new location of the archive at https://mailman.ripe.net/archives/list/ipv6-wg@ripe.net/
[ipv6-wg] Latest draft for RIPE554-bis confusion with MUST and SHOULD
- Previous message (by thread): [ipv6-wg] RIPE 554 Bis to consensus...
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Lorenz Nickel
Lorenz.Nickel at cs.tum.edu
Thu Nov 25 14:52:08 CET 2021
Hey everyone, I am a bit confused/worried about section 5 (Requirements for IPv6 support in software). There is a list, of what a developer/vendor MUST (at a minimum) support. However, the list contains multiple bullet points that include words that make them seem optional (like "something SHOULD be supported" or "it's recommended"). I don't have any experience with the creation of RFC and just joined this mailing list to be able to see what you are doing here out of interest, so I think this might be (at least partly) intended, as some of these bullet points seem to have been copied from other sources (e.g., https://www.ripe.net/publications/docs/ripe-554 contains this first bullet point with "recommended"). However, I think the standard should make sense on its own and therefore don't think that this is a valid justification. Things should be clearly marked as either mandatory or optional. Therefore, I would suggest rephrasing the introduction sentence from Software developer/vendor must at a minimum do the following things to guarantee this: to [A] software developer/vendor should at a minimum do the following things to guarantee this: However, in case some things in the list should be mandatory and some optional, I think it might make even more sense to split the list in an optional and a mandatory list. If all the things in the list are actually mandatory, I would propose the following changes to the bullet points: From It is strongly recommended not to use any address literals in software code, as described in “Default Address Selection for Internet Protocol version 6” [RFC6724] to Address literals must not be used in software code, as described in “Default Address Selection for Internet Protocol version 6” [RFC6724] From Every place where IPv6 addresses are shown or output the notation as specified in "A Recommendation for IPv6 Address Text Representation [RFC5952]" should be followed to Every place where IPv6 addresses are shown or output[,] the notation as specified in "A Recommendation for IPv6 Address Text Representation [RFC5952]" must be followed From When setting up a connection the software should follow Default Address Selection for Internet Protocol Version 6 [RFC6724] or Happy Eyeballs Version 2: Better Connectivity Using Concurrency [RFC8305] to When setting up a connection[,] the software must follow Default Address Selection for Internet Protocol Version 6 [RFC6724] or Happy Eyeballs Version 2: Better Connectivity Using Concurrency [RFC8305] From These requirements should also be checked in any library or tools used by the software to These requirements must also be checked in any library or tools used by the software I'm not particular interested in what wording exactly is used, these are just some suggestions. Please feel free to let me know what you think. Lorenz
- Previous message (by thread): [ipv6-wg] RIPE 554 Bis to consensus...
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
[ ipv6-wg Archives ]