Re: language tags useful? (was Re: Comments ...)

From: Martin J. Duerst (
Date: Sat Jun 07 1997 - 14:44:15 EDT

On Fri, 6 Jun 1997, Chris Newman wrote:

> On Fri, 6 Jun 1997, Walt Daniels wrote:
> > Lets go back closer to the beginning and see the case for why the IETF
> > thinks that language tagging it needed at all. Then make some
> > technical assessments of whether language tagging solves those
> > problems. Then judge whether the users are competent to correctly tag
> > stuff or even bother. There are opportunities for malicious or
> > politically motivated behavior that could prove quite devisive.
> The IAB charset workshop document (RFC 2130) is not, as far as I know, an
> official or final decision by the IESG or IAB. It simply holds a high
> level of respect in the IETF -- enough so that most people in the
> IETF/IESG will listen to it. If you wish to re-open the issues discussed,
> I do not believe it is too late. On the other hand, if you bring the
> language issue into question, it would also bring the entire document into
> question -- including the rules about using Unicode. I am not sure that
> is a wise move. I was not involved in creating that document, so I'll
> stay out of such a discussion for now.

The IAB charset workshop has definitely done much important work
(in particular promoted UTF-8 and therefore Unicode), and has
earned a lot of due respect.

The IAB charset workshop document also has made the correct conclusion
that language information is in many cases very valuable. Said workshop
or document, however, has NOT concluded that in-stream language codes
are necessary or that they are a good solution.
The problem of in-stream language tagging was left as an open issue
for further discussion, and the document contains some facts and
speculations useful for this discussion, while others are not addressed
directly or have been ignored.
The question is not "why the IETF thinks that language tags are needed",
but "whether and where and in what form the IETF thinks that language
tags are needed".

Regards, Martin.

This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:34 EDT