Brian E Carpenter <brian.e.carpen...@gmail.com> wrote: > On 2011-02-22 01:28, John Leslie wrote: >> Pekka Savola <pek...@netcore.fi> wrote: >> >>> Section S 5.1 already has a MUST requirement for supporting RFC2460. >>> RFC2460 main body and appendix describe the behaviour for assigning >>> flows at the originating node. This was made more specific in RFC3697 >>> that updated 2460. >> >> Saying just-that is reasonable. > > Well, I object to drawing anyone's attention to the Appendix of > 2460, which is explicitly not part of the standard.
I agree with Brian there: the Appendix of 2460 never was normative, and has been thoroughly superseded. > And I think that the main text in 2460 is underspecified, I won't disagree, but the MUST in 5.1 is still appropriate. > whereas RFC 3697 has proved to be plain confusing. I agree, but it's an normative as we've got. > And we don't yet have clear consenus on 3697bis. Does all this belong > in a requirements list? IMHO, pointing out RFC 3697 _is_ appropriate; but the readers would be better served by pointing out that 3697 will probably be superseded soon. I don't actually object to not mentioning Flow Labels, since it's less than clear what is "required" at the current time. Nonetheless, I would hope we could tell folks where to look. Ignoring something that central to IPv6 feels wrong... -- John Leslie <j...@jlc.net> -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------