>>>>> On Fri, 5 Jan 2007 08:50:38 -0500,
>>>>> James Carlson <[EMAIL PROTECTED]> said:
>> This is more than strictly required for interoperability.
> I think that's the key point. It's very important to make sure that
> the things marked as "MUST" or "MUST NOT" are actually requirements
> for interoperability or anticipated development, and not just needless
> strictures.
> Otherwise, it becomes much more tedious to support enviroments that
> require slightly different values or implementations that trade off
> simplicity against precision.
FWIW, I agree. I've not assess whether each of the RFC2461(bis)
constants can be a source of an interoperability issue, but if it's
not, "MUST" or "MUST NOT" is clearly a wrong choice.
Although SHOULD (NOT) might help clarify the intent (and I don't
oppose to adding it in this case), I'm not sure if it has a real
effect; there's always a vendor that misunderstands or even ignores a
particular part of a specification even if it's specified as a MUST.
So we won't be able to eliminate such abnormal implementations by
giving a spec a SHOULD or MUST. We may be able to convince such a
vendor of fixing their implementation with the specific keywords, but
I personally don't see significant difference for that purpose in this
case, with or without an RFC2119 keyword.
In any case, I have no strong opinion on whether to add an RFC2119
keyword to 2461bis (as long as it's not an unnecessary MUST).
JINMEI, Tatuya
Communication Platform Lab.
Corporate R&D Center, Toshiba Corp.
[EMAIL PROTECTED]
--------------------------------------------------------------------
IETF IPv6 working group mailing list
ipv6@ietf.org
Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------
