W3C home > Mailing lists > Public > xsl-editors@w3.org > January to March 2002

Re: escape-uri-attributes

From: Werner Donné <werner.donne@re.be>
Date: Tue, 29 Jan 2002 15:37:07 +0100
Message-ID: <3C56B393.7030701@re.be>
To: David Carlisle <davidc@nag.co.uk>
CC: Michael.Kay@softwareag.com, jeni@jenitennison.com, xsl-editors@w3.org
The character model is an interesting issue. The conclusion I would draw from
that is that specifications should explicitly use the term IRI, instead of calling,
for example, the type of an attribute "URI" but saying at the same time IRI values
are also allowed. This is, by the way, dangerous. XLink, for example, allows
"partially escaped URIs" in the href attribute. While this is probably to
accommodate IRIs, it also creates the possibility that even reserved characters are
not properly escaped. By using IRIs explicitly this can't occur, because eventually
IRIs will also always have to be in escaped form, IRI escaped form that is.

As a consequence, XML Schema, to take just one specification, should also provide
the basic type "anyIRI" besides the current "anyURI" type.

Mike's escape-uri() proposition is compatible with the character model, because
that function is an example of saying when the conversion from IRI to URI takes
place. Perhaps the name should be "escape-iri()" or "escaped-uri()", in order to
express the fact that what comes out is an URI and what goes in isn't (an
unescaped URI doesn't exists because an URI is escaped by definition).

Werner.

David Carlisle wrote:

> Actually I believe (W3C I18 WG should have the details) that the
> (current) intention is that URI shouldn't have to be escaped in XML (or
> HTML) documents and that the uri-quoting should be performed by whatever
> agent is lifting the uri out of a document and passing it to the uri
> handler.
> 
> the charmod document (still at WD stage) says that all new w3c
> specs MUST follow this scheme:
> 
> http://www.w3.org/TR/charmod/#sec-URIs
> 
> This would mean that when generating XML (or taking a slighlty
> revisionist view of history, HTML) one shouldn't quote the href 
> attributes at all.
> 
> David
> 
> _____________________________________________________________________
> This message has been checked for all known viruses by Star Internet
> delivered through the MessageLabs Virus Scanning Service. For further
> information visit http://www.star.net.uk/stats.asp or alternatively call
> Star Internet for details on the Virus Scanning Service.
> 
> 
> 


-- 
Werner Donné  --  Re BVBA
Engelbeekstraat 8
B-3300 Tienen
tel: (+32) 486 425803	e-mail: werner.donne@re.be
Received on Tuesday, 29 January 2002 09:40:24 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:59:52 GMT