> The above IAB recommendation is therefore a logical consequence from > what are described in the draft because the IPv6 address architecture > specifies the interface ID length is 64 for addresses beginning with > binary 000. > > We could still add the specific recommendation to rfc2462bis. > However, I personally hesitate to do that since I basically prefer > not hard-coding particular constants in general rules as long as the > specification is clear (and, in fact, I believe the specification is > already pretty clear on this point). > > What do others think? Any opinions or suggestions will be highly > appreciated.
I think what you have in 2462bis is the best approach; I agree with you that hard-coding the 000 prefix and /64 in 2462bis would be worse since it would add complexity to implementations and probably make it harder to evolve the IPv6 addressing architecture over time. Erik -------------------------------------------------------------------- IETF IPv6 working group mailing list [EMAIL PROTECTED] Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------