Point-by-point responses (was: Re: Protocol-00)

[John C Klensin]

--On Thursday, 22 May, 2008 18:32 -0700 Mark Davis
<mark.davis at icu-project.org> wrote:

> I had a number of comments posted some time ago on each of the
> documents; to the best of my knowledge you didn't respond on
> them. It sounds like you want me (and others like me) to go
> back and review each new version to see which have been
> addressed and which haven't, and revise my comments
> accordingly. It would be more effective to respond on the list
> with commentary as to why you did or didn't incorporate
> responses to the comments made.

Mark,

You may recall that, when "rationale" was posted, I included a
note about your comments.   Each one of them was addressed in
that document in some way with the exception of those that asked
for reorganization of the document set and those that were under
very active WG discussion in the weeks of the charter
discussions and thereafter but without any resolution.  

I believe that note also discussed the fact that a point by
point explanatory response to an omnibus set of comments is
immensely time-consuming, especially when those comments
intermix editorial comments, issues that the WG has not
addressed yet and that are certain to be controversial (in this
case, including reorganization requests), substantive but
largely uncontroversial issues, and repetitions of issues the
group had already resolved and even more so when many of the
change suggestions do not include an explanation of why you
believe they should be made (essentially the mirror image of
what you are asking me to provide).  Such point by point
responses are rarely provided in the IETF, especially at early
stages of work on a document in a WG, partially because it could
be an easy avenue to using extensive comments as a denial of
service attack on the group.  

While your comments --at least in my entirely subjective
judgment-- fall into a rather different category, consider, for
example, some of the comments that have been recently posted to
the list and that some of us believe are off-topic or very
unclear in the context and vocabulary of the group's
discussions.   Item by item discussions to many of them would
bring us to a dead stop.  

If it seems wise to Vint, we can move to an environment in which
we are using an issue tracker, with specific requirements for
defining and creating issues.  That imposes some additional
burdens on everyone, but is the only way I know of to get and
track a complete set of responses to issues and suggestions on
which there is some WG agreement that they represent issues that
are actually worth consideration.   Without that, while I was
very careful with your notes and "rationale", if one person
raises a topic that I believe is controversial (perhaps
controversial as to whether it is an issue at all) and it
generates no on-list discussion at all (or even no discussion
other than a "me too" from your close colleagues), it is likely
to be ignored or deferred at this early stage in the document
work.   

Especially in the case of Rationale, you will note that there
are now explicit placeholders in the document, noting the points
of controversy and asking (implicitly or explicitly) for group
discussion, on the suggestions that you and others made that I
considered too controversial to incorporate without further WG
discussion.   I believe that is an appropriate way to respond to
your comments and one that is actually more useful to most of
the group than the point-by-point list you ask for above.

The "move this to some other document", "lose this text
entirely", and "reorganize things in a major way" comments are
in a slightly different situation.   The WG has an explicit
charter item to consider the organization of the documents.  We
haven't had that discussion.  Within the WG, I know that
opinions as to what should be done range from "nothing" (i.e.,
the current organization is fine"), to "drop all explanatory and
contextual material so that only algorithms and tables are left,
presumably as no more than two documents", with several possible
positions in between (including, fwiw, my personal one).  I know
that some of us strongly oppose the second of these, in large
measure because of what we  believe to the context in which IDNs
operate and the various pieces of the implementation environment
and have explained that position on the list.   Those who
advocate the more drastic reorganizations and dropping of
material have not been equally explicit (e.g., I understand
about the desire to drop explanatory and rationale material in
general but am unable to determine whether some material fell
into that category).   

What I've tried to say before is that I would like to see
explicit and detailed position statements/ proposals that
address the reorganizational topic posted, compared to each
other by the WG, and debated.    Until that is done, I'm
reluctant, as an editorial matter, to start moving blocks of
text around or, more important, dropping text, simply because it
might confuse all of us about precisely what is being suggested.
I made too exceptions to that preference with the current set of
documents.   The first was to remove the material from
"rationale" that constituted specific critiques or explanation
of IDNA2003.   I did so because there has been a specific
request on-list to do so and because, while there had not been
any extensive discussion, I sensed general consensus that
material was no longer necessary.   If I got that wrong and
people want to see the material back, people should say so soon:
the text can be easily restored now, but it will get much harder
to restore as we rev the documents in the WG and especially
after we start reorganizing.   The other was the move of the
"contextual rules table appendix" from "rationale" to
"protocol".  That was largely a matter of convenience that I
have already explained on-list and in change logs.

As far as your comments on Protocol are concerned, they were
much fewer, after the above explanation about deferring
reorganizational and controversial comments is understood, than
those about Rationale.   My recollection is that the vast
majority of them were largely editorial and small details and
that I've incorporated them into the document.   However, it was
a long time ago (I did most of the work on the two sets of
extensive comments -- yours and Cary's-- during the interminable
charter debate and did not come back to them when the first WG
versions were actually generated) and I'm a continent away from
my notes.  I'm a continent away from my notes at the moment and
finished up the work on Protocol and queued it for posting while
on the road.  I will be home late Friday evening and will be
happy to recheck my notes on your suggestions to be sure that
nothing has fallen through the cracks.  If anything has, I'll
get a new version up before Monday.  But, again, you should
expect changes and annotations in the document, not a replay of
your comments with explicit point by point explanations of what
changes were made.

Obviously there are other ways to proceed.   If Vint is
convinced that one of them would be more efficient, I will, of
course, defer to his judgment.  But, until there is either
consensus on a different approach or instructions from him, I'm
going to continue to try to edit documents, insert explicit
placeholders and requests for further discussion, and no do
point-by-point responses unless I need clarification of what is
being requested or can focus things for WG discussion.

best,
    john