New proposal (was Re: UTF-8 and URLs)

Edward Cherlin (
Fri, 25 Apr 1997 14:44:35 -0700

Message-Id: <v0300783baf86d0f45756@[]>
In-Reply-To: <>
Date: Fri, 25 Apr 1997 14:44:35 -0700
From: Edward Cherlin <>
Subject: New proposal (was Re: UTF-8 and URLs)

François Yergeau <> a écrit:
>À 09:56 24-04-97 PDT, Larry Masinter a écrit :
>>I think given its likely controversial nature, we should clearly
>>make these recommendations in a separate RFC, and perhaps with
>>a new working group.
>Meaning what?  Two separate standards?  Or worse, a standard and an
>experimental/informational/BCP?  I thought we had already buried that one.
>Who wants a two-tier Web, with only the lower tier internationalized, raise
>your hand!

How about this: Two coordinated standards, where we put in some language
here acknowledging the work on international URLs, and then we take up that
problem in its own document, since it is important enough to do that way.
Then that document can acknowledge all of the other work on
internationalizing the Net, and take over the job of coordinating it in one

When I say "language acknowledging the work" I mean something that permits
changes in the direction of implementing Unicode, with appropriate measures
for backward compatibility, and forbids taking any other new direction.
Since the other new document will start with the proposal to use UTF-8/%HH
in URLs, and will be acknowledged here, we can prevent divergence and
worsening chaos without offending backers of the current and proposed
standard, and we can carry out the necessary experiments in peace.

If that statement is too convoluted, try this: We agree on splitting this
document in two procedurally, and letting one segment go on to the next
stage while we write the other, but we treat them logically as part of the
same proposal, making sure that the segments are compatible from the start.
Then we will have a working improvement to the current standard, and a
clear direction for development of the next version, which is also a clear
guide for implementors and users.

As you know, I have argued strenuously for allowing UTF-8 in URLs. Larry
now says he is willing to do that, but prefers a different procedure than
putting the experiment into this draft. I don't care much about procedure
personally, and would be willing to go either way. François and Martin now
need to say whether there is an alternative procedure they can live with,
such as sufficiently tight coordination between the current document and
the proposed one. They and Larry could propose language for coordinating
the documents. Anyone else could suggest something useful.

What do you say?


>François Yergeau <>
>Alis Technologies Inc., Montréal
>Tél : +1 (514) 747-2547
>Fax : +1 (514) 747-2561

Edward Cherlin     Everything should be made
Vice President     Ask. Someone knows.       as simple as possible,
NewbieNet, Inc.                                 __but no simpler__.                Attributed to Albert Einstein