Re: [manet] Mirja Kühlewind's No Objection on draft-ietf-manet-rfc5444-usage-06: (with COMMENT)

[Spencer Dawkins at IETF <[e-mail]>]

Hi, Christopher,

On Sat, Jul 1, 2017 at 11:00 AM, Christopher Dearlove <
[e-mail]> wrote:

> Mirja
>
> Thank you for that.
>
> The short version is that the position strongly recommended was the WG
> consensus. The authors of this draft, and of NHDP and OLSRv2, could
> definitely have lived with a MUST for those protocols. However reactive
> protocols (such as the then still active AODVv2 draft) have a problem with
> unmodifiable messages if a route reply is to collect any information on its
> return path, something that is required if using link metrics. (At least
> not without other serious problems.)
>
> So, in short, there's a use case that would violate a MUST. But it should
> be avoided if possible, hence the agreed wording.


I won't include this in my No Objection ballot, because you and Mirja are
already chatting about it, but if you have a specific reason in mind for
not using MUST ("being used with a reactive protocol"), saying "SHOULD
unless" would leave less room for future mischief.

And your response helped my understanding a lot, so I hope future readers
would also benefit from that explanation.

Do the right thing, of course.

Spencer