my comments on draft-ietf-idnabis-protocol-14 (second part)

John C Klensin klensin at jck.com
Wed Sep 2 09:49:32 CEST 2009



--On Wednesday, September 02, 2009 11:43 +0900 "\"Martin J.
Dürst\"" <duerst at it.aoyama.ac.jp> wrote:

> Hello John, Vint,
> 
> I'm glad to see that John in particular is understanding and
> appreciating my comments.
> 
> For many of them, I think it's fine to defer them, as long as
> I don't have to keep the task list :-). There are a few (in
> particular I seem to remember section 4.1) where I really had
> now clue what was intended, and I fear that there's a great
> chance that implementations may vary widely, and therefore I'd
> prefer at least those to be addressed now.

I will take responsibility for that, and will go back and look
at 4.1 again.  I'm not opposed to making a change or three; I'm
just afraid that, if we make a lot of changes at once, we are
going to introduce errors.

> Overall, I'm rather unhappy with the general quality of the
> documents, because I think in a big part, they are the result
> of the "we have to hurry up, we can always clean things up
> later" mentality that was very often dominant in this WG.

And at least as dominant in the original IDN WG as well.  The
present situation, about whose state we probably agree, has been
a long time coming.

> I
> know that John (and all our other editors) can write better
> specs, and I think the above mentality has overall resulted in
> less quality without actually using less time. So I'm not
> happy to see the "we don't have much time now" argument
> brought up again.

As the guilty party (or victim, as one prefers), I again tend to
agree with you, although I have a more complex model involving
trying to fix things by inserting patches from many different
people at times of high controversy with not enough WG
involvement in contemporaneous comments on each other's
suggestions.   There has also been the problem of looking too
closely at the documents for too long -- while most participants
in the WG have looked at the documents only every posted draft
or two, if that often, there have been few weeks in the last two
years that I haven't had at least some piece of the collection
open -- under that circumstance,  I start seeing things that
aren't there and missing things that are... and I'm not sure
that Harald, Patrik, and Cary are in much better shape.  I think
those two factors contributed at least as much to the current
situation as time pressure and haste.

While I recognize the particular reasons for urgency that Vint
identifies, I have two different reasons for suggesting deferral
of major editorial work at this time.  The first is that the WG
has gone on too long and we are, IMO, in at least as much risk
of making mistakes through exhaustion as through haste.  The
second is that, at least for me, it is hard to pin down
technical details and editorial quality at the same time.   So
my hope is that we can use this final period of cycling through
the documents to get the technical details as solidified as
possible, get IETF consensus on those details, and then go back,
after a few weeks of not having to look at the documents at all,
and do a complete rewrite on the assumption that it is editorial
work and that any change in technical/substantive meaning is a
very big deal.  If that few weeks isn't enough, being able to
say "this is correct technically, the only issue with the
quality of writing and explanation" permits bringing in outside
technical editing talent to make a pass, an approach I've used
occasionally in the past.   I think that the "later, for Draft
Standard" approach will leave us with a much better result down
the road than trying to make all plausible editorial fixes
before we have formal IETF consensus on the whole package, even
if there were no time constraints.

> Of course, if there is rough consensus to address some things
> later, then I'll accept that.

Thanks and regards,
    john



More information about the Idna-update mailing list