[Top] [All Lists]

Re: IONs & discuss criteria

2008-03-08 07:54:52

On Mar 7, 2008, at 8:56 AM, Spencer Dawkins wrote:

(stuff deleted)

So, for example, it probably IS worth finding out if the rest of  
the ADs who sponsor reviewing bodies

As an AD who sponsors a reviewing body (the Security Directorate), I  
guess it is my
turn to step into the fray.  Yikes!

Before I get started, I would like to note that I would not have  
survived my first year as an AD without
the support of the Security Directorate.  I continue to be impressed  
by the quality of the reviews that
are performed.  It is not an easy task when you are assigned a  
document from an unfamiliar area.   I
am sure that ADs sponsoring other reviewing bodies feel the same way.

also use Russ's division into

- "did you consider the reviewer's comments?", especially when the  
review was issued as
part of IETF Last Call, and

- "how did you address this specific comment, which I agreed  
with?", whether the AD entered the comment as a non-blocking  

I considerate it my responsibility to ensure that reviews from the  
security directorate
are considered thoughtfully.  I requesting that members of the  
community devote their
precious time to these reviews, and I don't want to see them  
ignored.  So yes, I have
filed process discusses of the form "The authors have not responded  
to John Doe's
secdir review."  If the review had unsuccessfully attempted  to  
initiate a conversation
with the authors and they were unresponsive, that may even be the  
only appropriate
path to take.  If the review was in the form of concrete suggestions  
("the security
considerations section needs to address man-in-the-middle attacks") I  
may choose to
issue a DISCUSS using their text instead of involving the reviewer.

If the review was acknowledged, then I review the email thread.  I  
attempt to verify that
any agreed upon actions are included in the current draft or  
implemented by a note to
the RFC Editor.  Modifications promised but not executed merit a  
discuss just to ensure
they aren't forgotten.  (I believe that this is consistent with the  
DICUSS criteria ION, under
the IETF process for document advancement.  Regardless, it doesn't  
present a burden or
unduly complicate the process.  At least, no one has complained about  
this type of discuss.)
If issues were raised in the review but agreement was not reached, I  
try to decide whether
I agree with the reviewer's comment *and* its relative importance.  
So, these "residual"
issues get addressed using the second method.

The more difficult problems come when a review is submitted just days  
before a telechat.
IETF Last Call has typically closed, and the authors may not have  
even seen the review
yet.  This is dangerous territory, since the temptation is to cut and  
paste the entire review
into my discuss to ensure that it isn't overlooked.   Of course, the  
more appropriate and
more helpful course of action is to determine which comments I  
support and separate
them into non-blocking comments and discuss worthy buckets.

[Confession time:  The temptation of cut-and-paste is sometimes too  
strong for a mere mortal,
though. I just revised my discuss on a document from yesterday's  
telechat, where I had cut
and pasted a secdir review, to separate the issues into the discuss  
and comment buckets.
Knowing better isn't the same as doing better.  My thanks to the  
sponsoring AD, who kept
me honest and asked me to review and revise!]

I am not particularly methodical by nature, so I can't claim I  
perform this exact process in
every case.  However, that is a good overview of the process I try to  

Tim Polk
IETF mailing list

<Prev in Thread] Current Thread [Next in Thread>