Re: Character set issues again

Peter Deutsch <> Fri, 15 October 1993 22:13 UTC

Received: from by IETF.CNRI.Reston.VA.US id aa16676; 15 Oct 93 18:13 EDT
Received: from CNRI.RESTON.VA.US by IETF.CNRI.Reston.VA.US id aa16670; 15 Oct 93 18:13 EDT
Received: from by CNRI.Reston.VA.US id aa03757; 15 Oct 93 18:13 EDT
Received: by (4.1/UCD2.05) id AA25103; Fri, 15 Oct 93 14:41:49 PDT
Received: from by (4.1/UCD2.05) id AA18097; Fri, 15 Oct 93 12:55:57 PDT
Received: from expresso.Bunyip.Com by with SMTP (5.65a/IDA-1.4.2b/CC-Guru-2b) id AA01947 (mail destined for on Fri, 15 Oct 93 15:57:19 -0400
Received: by (NX5.67c/NeXT-1.0) id AA14369; Fri, 15 Oct 93 15:52:38 -0400
Message-Id: <>
Sender: ietf-archive-request@IETF.CNRI.Reston.VA.US
From: Peter Deutsch <>
Date: Fri, 15 Oct 1993 15:52:35 -0400
In-Reply-To:'s message as of Oct 15, 8:44
X-Mailer: Mail User's Shell (7.2.4 2/2/92)
Subject: Re: Character set issues again

I'm baaacck!

Okay, first apologies for being incommunicado recently.
We've been working flat out on comapny-related stuff and
travel. This should now be calming down somewhat and I
should be able to get a draft out before the IETF.

Let me pick this up with Dan's comments on unicode. 

[ Dan wrote: ]

.  .  .
> In the time since the draft was written, there has been a lot of progress
> towards adoption of Unicode as an all-inclusive international character set.
> Major companies, including Microsoft and Apple, have committed to switching
> to Unicode.  Unicode is currently supported by Microsoft Windows-NT
> and Plan 9 (the latest research version of ATT Unix), and a version of Xterm
> supporting Unicode was recently posted to Usenet.
> Plan 9's encoding was designed to look like vanilla ASCII when viewed
> by naive programs, including the Unix filesystem's filename code.
> This has the benefit that you can use Unicode filenames even on operating
> systems that don't support Unicode.

I certainly have no problems with adding the words he
requested, but I'm not really qualified to say that
Unicode really is the proper choice here. I _have_ heard
good things about Plan 9's encoding from a number of
people and welcome feedback. Other people care to comment?

> I propose that wnils officially endorse Unicode as the non-ASCII character
> set of choice to be supported someday.
> I also propose that wnils separately consider the issue of
> how to encode Unicode so that plain ASCII and Unicode can be intermixed,
> and suggest that Plan 9's encoding might be a solution.  

My first reaction to this is a question - is this a
general enough problem that we should be addressing it in
WNILS, or should it be decided in a larger context? Again,
I don't particularly care, as long as we actually have a
viable, working solution.

> Finally, wnils should look at how to deal with the difficulties of comparing
> Unicode text strings.  (For an idea of how one company tried to deal with
> Unicode, look at the Microsoft Windows Win32s programmer's reference manuals.)
> Whew.  Hope you guys didn't settle this issue while I wasn't watching.
> I'd love to hear whether you think my suggestion is germane...
> - Dan Kegel (

The only decision that I am aware of is to incorporate the
MIME work as one of the standard output formats and I plan
to spend the next few days with the appropriate documents
and then wordsmithing this in. I have the MIME docs, plus
Greg Vaudreuil's recent posting, Dave Crocker's work and a
strong suggestion from John Klensin to include SGML in our
planning. I'm not sure when I'll actually be posting the
final draft, since I'm going to be on the road, but I will
do my best to actually give people some time with it
before the meeting this time!

					- peterd