Message-Id: <3.0.1.32.19970507160442.006d9ce0@riq.qc.ca>
Date: Wed, 07 May 1997 16:04:42 -0400
To: "Martin J. Duerst" <mduerst@ifi.unizh.ch>
From: Alain LaBont/e'/ <alb@riq.qc.ca>
Subject: Re: "Difficult Characters" draft (in URLs)
Cc: URI mailing list <uri@bunyip.com>
In-Reply-To: <Pine.SUN.3.96.970507203249.245n-100000@enoshima>
References: <3.0.1.32.19970507124205.0073ca78@riq.qc.ca>
Mime-Version: 1.0
Content-Type: text/plain; charset="iso-8859-1"
Sender: owner-uri@bunyip.com
Precedence: bulk
Content-Transfer-Encoding: quoted-printable

A 20:37 97-05-07 +0200, Martin J. Duerst a =E9crit :
>On Wed, 7 May 1997, Alain LaBont/e'/ wrote:
>
>> A 11:23 97-05-07 +0200, Martin J. Duerst a =E9crit :
>> >I think we pretty much agree that we should discourage URLs with
>> >accented uppercase letters.

[Alain] :
>> I personally do not, nor any of my colleagues... We would agree with t=
his
>> if upper case were not allowed.

[Martin] :
>To be more precise, could you then agree to the suggestion
>that uppercase accented letters should be discouraged in
>URLs intended for a general "France-French" audience, given
>as an example of how case might affect URLs, the reason for
>this being the current widespread use of keyboards that make
>it difficult to enter such characters?
>
>Regards, 	Martin.

No, we could not agree on this as long as upper case will not be
discouraged for everybody. Even if upper case letters are not available o=
n
all machines, in France, all environments which use browsers have the
possibility to copy (as you suggested for URLs) the exact characters
required from either a system-provided table or an existing URL. Short-te=
rm
considerations have made people take bad decisions in the past, that woul=
d
be a bad decision too.

And upper case accented letters are definitely required for French.

I have to be frank on this, this is a firm position. No compromise possib=
le
from our point of view without limiting expression. People have suffered =
so
much with typewriters and later on with computer languages that when we
build new models we ought not be bound again by artificial restrictions
based on short-term considerations. The non-placement of <oe> and <OE> in
the final draft of ISO/IEC 8859-1 in 1987 was due to such a bad decision
taken by a very small team, based only on the fact that a French company
had no printers with these characters.

Alain LaBont=E9
Qu=E9bec