I conceded that is probably true for most of the IPv4-Embedded IPv6
Address Formats in section 2.2 of RFC6052, but the /96 format seems
indistinguishable from other IPv6 addresses with embedded IPv4 addresses
described in Section 2.4.5 of this draft.
Could we say that the IID is 64 bits long "for all addresses except those
that start with binary 000, and those that embed IPv4 addresses [RFC 6052]"?
This is demonstrably false (there are lots of IPv6 interfaces with for
instance /124 or /126 masks, among others). I don't expect all of these
IPv6 interfaces to be changed to /64 even if an RFC says so.
So the question is, what do you gain by stating this in an RFC?
(I know, the discussion is old and I don't expect any views to change.
It stills seems wrong to publish an RFC with such an obviously wrong
statement.)
Steinar Haug, AS2116