ietf
[Top] [All Lists]

Re: Gen-art LC review: draft-ietf-6lo-dect-ule-08

2016-12-20 08:39:04


On 12/20/16 6:26 AM, Jens Toftgaard Petersen wrote:
Hi Robert,

Thanks for your clarification of normative references to work in progress.

I would prefer to keep the reference as in the updated version -09:

"The FPs in such a scenario should behave as Backbone
    Routers (6BBR) as defined in [I-D.ietf-6lo-backbone-router]."

Are you OK with that?
Maybe. I'm not the expert here, so all I can do is ask (what are hopefully useful) questions:

Here's one way to help figure this out: Would you be ok deleting the sentence and the reference? (I'm not asking you to do that, just to consider what it would mean to the resulting set of implementations if you did.)

If you think people will build the right software with that sentence missing, and you'll get the interoperability you're aiming for, then it's fine as an Informative reference. This would be the case if what you're really doing with that sentence is saying "other documents in the base protocol this is building on require you to behave this way, we're just pointing at that here to remind you of that fact".

If that's going to leave implementers guessing at what their implementation is supposed to do, and interoperability will suffer, then it's normative. (What you're really doing with that sentence is telling the implementers something _new_ - that they SHOULD behave as a 6BBR in this situation, and nothing else tells them to do that, and the reference is here to tell them how to do that).

Alternatively, if it's just fine that some large subset _doesn't_ behave as a 6BBR, then its fine the way it is. You might consider saying "One good way for an FP can behave in such scenarios is as a Backbone Router..." instead, to make it clearer that this is just informational guidance rather than something that would affect interoperability in the resulting implementation base.

Either way, I'm happy leaving it up to you and your ADs at this point.

Best regards,
   Jens

-----Original Message-----
From: Robert Sparks [mailto:rjsparks(_at_)nostrum(_dot_)com]
Sent: 17. december 2016 00:31
To: Jens Toftgaard Petersen <jtp(_at_)rtx(_dot_)dk>; General Area Review Team 
<gen-art(_at_)ietf(_dot_)org>; ietf(_at_)ietf(_dot_)org; 6lo(_at_)ietf(_dot_)org; 
draft-ietf-6lo-dect-ule(_dot_)all(_at_)ietf(_dot_)org
Subject: Re: Gen-art LC review: draft-ietf-6lo-dect-ule-08

Just one comment at the end:


On 12/15/16 4:04 PM, Jens Toftgaard Petersen wrote:
Hi Robert,

Many thanks for your review. See my comments and answers inline below.

Best regards,
    Jens

-----Original Message-----
From: Robert Sparks [mailto:rjsparks(_at_)nostrum(_dot_)com]
Sent: 28. november 2016 21:22
To: General Area Review Team <gen-art(_at_)ietf(_dot_)org>; 
ietf(_at_)ietf(_dot_)org;
6lo(_at_)ietf(_dot_)org; draft-ietf-6lo-dect-ule(_dot_)all(_at_)ietf(_dot_)org
Subject: Gen-art LC review: draft-ietf-6lo-dect-ule-08

I am the assigned Gen-ART reviewer for this draft. The General Area Review Team 
(Gen-ART) reviews all IETF documents being processed by the IESG for the IETF 
Chair.  Please treat these comments just like any other last call comments.

For more information, please see the FAQ at

<http://wiki.tools.ietf.org/area/gen/trac/wiki/GenArtfaq>.

Document: draft-ietf-6lo-dect-ule-08
Reviewer: Robert Sparks
Review Date: 28 Nov 2016
IETF LC End Date: 02 Dec 2016
IESG Telechat date: 15 Dec 2016

Summary: Ready with nits

Nits/editorial comments:

First, forgive me, but I need to grumble a little bit:

The way this document approaches standardization makes me very uncomfortable.
The language is passive and relies on inference to the point that it risks 
being vague. If this review were earlier in the document's life-cycle, I would 
strongly suggest a complete restructure focusing on explicitly specifying what 
the implementation is supposed to do.

But, the document has had several reviewers who didn't trip up on this point, 
and the working group believes it is implementable, so I'm going to set that 
aside and provide some concrete suggestions for removing some nits from the 
existing text.

In document order:

1) In section 2.1 "This draft defines 6LoPAN as one of the possible protocols to 
negotiate". That's not what this draft appears to do. Rather, it defines behavior once this 
6LoPAN over DECT ULE has been negotiated. Some other document is defining the negotiation. I 
suggest replacing the sentence with "[TS102.939-1] defines this negotiation and specifies an 
Application Protocol Identifier of 0x06 for 6LowPAN. This document defines the behavior of that 
Application Protocol".

[Jens]: Very good comment and suggested wording. Is implemented in
revision -09

2) The "not recommended" in the last sentence of 2.3 looks like it
should be a
2119 keyword (NOT RECOMMENDED). Similarly, the "shall" in the last sentence of 
the first paragraph of 2.4 looks like it should be a SHALL (consider using MUST instead).

[Jens]: I agree. Is implemented in revision -09

3) At the mention of LOWPAN_IPHC in the second paragraph of 2.4, consider referencing RFC6282. It's 
not clear what the sentence is really trying to convey, though. "all the requirements" is 
very vague - can you point to a specific requirement list somewhere? "It is expected" 
implies that you believe there's a chance that it might fail. Could the sentence be removed (you 
cover this in 3.2) or be replaced with a more direct statement?

[Jens]: Agreed, is covered in section 3.2. Will be removed in revision
-09

4) In the first section of 3.1 you have "The PP MUST be pageable".
Interestingly, the word "pageable" does not yet appear anywhere in the RFC 
series. Please add a reference into the ETSI docs that will lead the reader to a 
definition.

[Jens]: I don't think we have definition of the term we can refer to. Will 
change to an explanatory wording in revision -09.

5) In the last paragraph of 3.2 (before 3.2.1), third sentence, you
introduce using the multi-link subnet approach. Please either add a
reference to
RFC4903
here, or point forward to section 3.3.

[Jens]: A reference to RF4903 added in revision -09

6) In section 3.2.1, third paragraph, you say addresses are derived "similar to the 
guidance of [RFC4291]. I don't believe that is sufficient. Perhaps you should say 
"following the guidance in Appendix A of [RFC4291]"?

[Jens]: Yes, your suggestion is more accurate. Will be done revision -09.

7) The last paragraph of 3.3 says "The FPs operation role in such scenario are rather like 
Backbone Routers (6BBR) than 6LBR, as per [I-D.ietf-6lo-backbone-router]." Is this trying to 
_specify_ the behavior of the FP in this scenario? If not, it's unclear what the sentence is trying 
to accomplish. If so, then the sentence should be "The FPs in such a scenario behave as 
Backbone Routers (6BBR) as defined in [I-D.ietf-6lo-backbone-router]." And that reference 
should be normative, rather than informative.

[Jens]: I agree the wording is bit vague, but I believe we cannot
refer to work in progress [I-D.ietf-6lo-backbone-router] in a
normative way. I will tightening the wording revision -09
You _can_ refer normatively to a work in progress (and I think it may still be 
the right thing to do here). The consequence is that this document would stay 
in the RFC Editor's queue in a state known as MISSREF (for missing reference) 
until the referenced document was approved for publication as an RFC. You can 
see examples of this at work on https://www.rfc-editor.org/current_queue.php.