Re: [Gen-art] Genart last call review of draft-ietf-ippm-2330-ipv6-04

[Francesca Palombini <[e-mail]>]

Hi Al,

Thanks for your reply.
I cut selected parts of the email to detail some of my comments, see inline. 

Francesca

> > (About MUST, was there any specific reason not to use the updated
> > boilerplate referencing RFC8174?)
> >
> [acm]
> You mean the Requirements Language RFC has been updated?
> News to me.  Will fix. Seems like the Nits check should catch that.
> Also, it can't be 8174 alone, the definitions of the terms has not changed.
> 

I think it does for Standard track docs.
Yes, I did mean the boilerplate that *also* references RFC8174 (which you can find in RFC8174) :) 

> > I would have appreciated a reference to a load balancing over parallel
> > paths example.
> [acm]
> ECMP is a rather well-known circumstance in IP networking today, as are
> other forms of load balancing, but I don't know of a canonical reference.
>

Ok. (This was just for my -or a generalist reader- sake, since I am not very knowledgeable in the area)

> > To be consistent with the first bullet of the list above ("It includes
> > a valid IP header: see below for version-specific criteria."), I would
> > rephrase the text above with something on the lines of:
> >
> > "For an IPvX (...) packet to be standard-formed, the IPvX-specific
> > criteria for a valid IP header are:"
> [acm]
> Your wording suggestion dropped the clear indication of a requirement.
> We are using the RFC2119 terms consistently for requirements.
> 

I was trying to point out that the first part of the section (first bullet list) does not use RFC2119 terms. I read the second bullet list as a "sub-list" of the first one, which is why I was suggesting removing the REQUIRED term. Or you could rephrase the first list to use REQUIRED too. Anyway, this is nit-picking. Feel free to disregard.

> [acm]
> We can add the pre-5378 disclaimer as a catch-all, but I doubt the original
> authors would make any fuss about the small amount of common text with
> 2330.
> Almes, Paxson, Mahdavi and Mathis are all gentlemen and the best of their
> time.
> 

I'm sure they are :) Just relaying what the id-nits told me.