W3C home > Mailing lists > Public > www-tag@w3.org > April 2003

Re: [Minutes] 14 Apr 2003 TAG teleconf (URIEquivalence-15, IRIEverywhere-27, xmlIDSemantics-32, abstractComponentRefs-37, namespaceDocument-8)

From: Chris Lilley <chris@w3.org>
Date: Wed, 16 Apr 2003 05:07:22 +0200
Message-ID: <3030342875.20030416050722@w3.org>
To: Tim Bray <tbray@textuality.com>
CC: "Ian B. Jacobs" <ij@w3.org>, www-tag@w3.org

On Wednesday, April 16, 2003, 12:29:30 AM, Tim wrote:

TB> Chris Lilley wrote:

>> In the XML instance, this may be accomplished by having them appear as
>> themselves (unless you deliberately chose a less portable and less
>> comprehensive encoding than the two universally understood ones and
>> declared it as such in the xml encoding declaration in which case, ask
>> yourself again what you did that for) or via NCRs.

TB> I write XML in ISO-8859 all the time.

And like I said, thats your choice to use a more restrictive encoding
that not all parsers can read and that cannot represent most

TB> Furthermore, the original source code of the XML 1.0 specification
TB> was in ISO-8859-1. For me, ISO-8859-1 is culturally appropriate
TB> and storage-efficient.

I wonder, then, why it is not mandatory for all XML parsers and why
the equally culturally appropriate Shift-JIS and Big5 and assorted
ISCII encodings and .... well you get my point.

Assuming Latin-1 is accepted by everyone is a Euro-centrism that
initially helped the Web spread (in Europe, where going beyond ascii
had been a pretty amazing thing at the time) but it is certainly not
universally understood.

TB> 'Ongoing' is written in pure ASCII, (which, left undeclared,
TB> pretends to be UTF-8).

That is a better choice.

TB> 'Ongoing' makes regular use of non-ASCII characters, this is made
TB> easy by XML's entity & NCR mechanism.

Yes. As long as it does not make much use of it, the inefficiency is
readily masked and the authoring convenience is greater for the users
of less than well internationalized OSs.

TB> XML is explicitly designed to allow people to use the encodings
TB> that are appropriate for them, and there's good support for this
TB> in deployed software.

Provided they are willing to take the risk, yes, they might well get
away with it especially if the audience is geographically narrow.

 Chris                            mailto:chris@w3.org
Received on Wednesday, 16 April 2003 00:08:45 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:32:37 UTC