Speaking
for myself without my chair hat on, I am not sympathetic to this errata.
RFC 3162 uses the type "Address" in a context that makes it fairly
clear that it is talking about a new data type for IPv6 addresses, rather than
the existing IPv4 address type defined in RFC 2865.
OK, just to satisfy my own mind, were you guys aware at the time
that we were “defining a new RADIUS data type”? I know that I
wasn’t: I just thought that, gosh, “Address” is a good thing
to call a field that holds an address. I don’t recall any
conversation about it amongst us (or in the WG) at the time; was it just so
obvious that it didn’t even bear mention? Put another way, am I the
only moron on the planet that was completely unaware of the “Da Vinci
Code” aspects of the RADIUS RFCs (esp. 2865)?
Previous
RADIUS RFCs did not "Update: RFC 2865" when introducing new data
types, so I don't believe that this was required in RFC 3162 or any other
document.
Actually, had I understood that we were, in fact, altering the
set of RADIUS data types at the time I probably would have suggested updating 2865…
There
are other instances in RFC 3162 we have been discussing, where the data type
used may not be clear (such as Interface-ID), but this specific instance
doesn't seem ambiguous.
…