Re: UTF-8 reg tags...

From: Mark Davis (
Date: Wed Sep 11 1996 - 16:21:19 EDT

That is not a particularly satisfactory answer. While I don't know of
any implementation of 10646 that is not in fact Unicode, there could
well be some, which could result in loss of information.


unicode@Unicode.ORG wrote:
> Hi,
> I'd prefer to see the tag as ISO-10646-UTF-8. (I'm the one who registered the
> ISO-10646-UTF-1 char set tag with IANA.) There is a fairly consistent pattern
> of "ISO-nnnn-...." for a number of the registered tags, but also note that
> there are a lot that don't follow the pattern, and a lot of aliases.
> The real-time/Internet Version 7 project ( uses
> ISO-10646-UTF-8 consistently in all text fields in all protocol layers; I had
> assumed the tag was already registered. (from some prior comment in this or a
> related forum; I should have checked of course :-)
> This tag does not (and IMHO should not) specify the version of the repertoire
> used; as was pointed out in a prior note, if you run into a codepoint you don't
> know, having previously seen a higher version number doesn't help you. (You
> don't want to quit/fail on a version higher than you know, because the odds are
> extremely high that you won't see unknown code points anyway! In a large number
> of cases you aren't even going to see codes > 128.)
> Best Regards,
> Robert Ullmann
> Lotus/IBM
> +1 617 693 1315

This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:31 EDT