Re: Last Call: <draft-ietf-6man-rfc2460bis-08.txt> (Internet Protocol, Version 6 (IPv6) Specification) to Internet Standard

[Brian E Carpenter <[e-mail]>]

Hi Fernando,

First let me say that I though Ole's summary was fair. One comment below:

On 10/02/2017 10:30, Fernando Gont wrote:
> On 02/09/2017 07:36 AM, [e-mail] wrote:
>> Fernando,
>>
>> Pete asked me to summarize the objections to option 1 - banning header insertion explicitly.
>> I responded with the set of objections I've heard for all options, as I couldn't see a straightforward way of only summarising for option 1.
>>
>> I don't understand your message.
>> Do you disagree with the summary itself? Are there arguments missing?
>> Or is your grief that the I have distilled the arguments wrongly or put them in a bad light?
>>
>> Or are you just rehashing your position on the issue?
> 
> I think that some points are not as clear as they should be:
> 
> 1) The current state of affairs with respect to IPv6 EH insertion is
> that insertion is forbidden. It has always been clear to everyone.

I don't think it has. In fact, that's the whole point: some people
have *not* deduced that rule from the RFC2460/RFC1883 wording.

   Brian

> 
> 2) However, some folks came up with proposals to insert EH, on the basis
> that "RFC2460 does not explicitly ban EH insertion". If there's people
> arguing that, we clearly need to make this clear in the spec.
> 
> 3) There was a consensus call, yes. When the call was made on the
> mailing-list, the vast majority of supporters of "let's keep the
> ambiguity" were folks from the same company as "2)". I have no idea if
> this changes (or not) "consensus"... but this is clearly an important
> datapoint.
> 
> 4) Given "1)" and "2)" above, it should be evident that the spec needs
> to be crystal-clear on this topic.
> 
> 5) Arguing in favor of keeping ambiguity in a spec that has generated
> 600+ messages in the very group that standardizes the protocol pretty
> much reads like "Let's publish a lousy spec!". And I think that would be
> very bad. We're talking about something that is at the core of the
> protocol, essentially "Is IPv6 an end-to-end protocol?". I would expect
> rfc2460bis to answer such a very basic question, and I'm curious how we
> could move a spec to (full) Standard without answering such a basic
> question.
> 
> 
> There's no grief at all. If anything, just a concern that some of these
> items might not be clear enough, and, in particular that without the
> datapoint in "3)", folks might get a misleading interpretation of the
> discussions that happened in th wg. "3" could be read pretty much like
> "all folks from the same company that has a proposal to insert EHs what
> insertion of EHs to be allowed".
> 
> Thanks,
>