Language tags, the phillips draft, and procedures

John C Klensin john-ietf at
Fri Jan 7 17:14:26 CET 2005


I've just reviewed the last 48 hours of these threads and a very
high volume of associated postings, many or most of them after
the Last Call formally closed and the tracking system
automatically moved the status of the document into the "waiting
for AD" state.   While Ted Hardie and his colleagues try to sort
out how to proceed (an activity on which they should have the
sympathy, support, and best wishes of all of us), and in
deference to those who follow the IETF list but who would prefer
to not be involved in the details of these discussions, I'd like
to suggest that everyone voluntarily declare a cooling-off

The threads themselves indicate to me that we passed, at least
several days ago, that critical point in mailing list
discussions after which very little that is being said is new;
instead, we (and I have certainly been guilty) are repeating
variations on the same arguments, sometimes accompanied by
rising emotional temperatures, without convincing anyone to
change their minds or positions.  That, too, is an indication
that it is time to stop writing messages and try to regain a bit
of perspective.  

If people are feeling an overwhelming need to find something to
think about in this area, I'd suggest two things:  The first is
a reerading of Kristin Hubner's posting yesterday morning.
Whether one agrees with her analysis or not, or believes that
the dichotomies are as precise as her note implies, the note was
the first attempt I saw in many days, by someone who has not
been immersed in the discussion, to take a different cut at
understanding the differences in perspective and assumptions
that clearly exist.  I think that, at this stage, any such
thoughtful and serious attempt, by someone who has not generated
dozens of messages on these threads, deserves careful reading
and thought by the rest of us.  

The second is that it seems to me that there is an apparent
contradiction in some of the discussions.  It may not be a real
contradiction, but it is adding to the confusion.  A
simplification of one of the positions is that this work is
needed because the design, structure, or semantics of 3066 are
insufficient to deal with distinctions (about decomposing tags
or about substantive language issues) that are required in the
real world.  The contrasting position, also simplified, is that
this really is not a change to 3066, is forward and backward
compatible, and merely represents writing down some restrictions
on what can be registered that 3066 leaves to a matter of taste.
While it is certainly possible for both of those things to be
true, the question that has not been asked on-list is whether
the combination actually implies that 3066 is hopelessly broken,
that compatibility with it is a bad idea, and that we should
really bite the bullet and look for a syntax that does not
require trick parsing and that can really support, unambiguously
and perhaps extensibly, the various language, country, script,
region, time period, phonetic styles, and whatever else, might
now or in the future be at issue here.  Please don't try to
answer that question today, especially on the IETF list.  But
let's all think about whether the real source of these
discussions is, at least in part, the possibility that a key
problem is the way in which both 3066 and the proposed
replacement are overloading both a style of using ISO 639 and a
small amount of syntax and, if so, whether that conclusion
should be our real starting point.

regards to all,

More information about the Ietf-languages mailing list