W3C home > Mailing lists > Public > www-tag@w3.org > October 2002

IRIs everywhere (including XML namespaces)

From: Jonathan Marsh <jmarsh@microsoft.com>
Date: Wed, 9 Oct 2002 09:09:55 -0700
Message-ID: <330564469BFEC046B84E591EB3D4D59C07E1167A@red-msg-08.redmond.corp.microsoft.com>
To: <www-tag@w3.org>

The XML Activity has been fairly enthusiastic in adopting IRIs [1]
(under various nomenclature).  In particular, we have encouraged users
to put IRIs in XML documents, and described how such IRIs can be
converted to URIs consumable by legacy URI processors and resolvers.

Some examples of IRI use:
- XML 1.0 system identifiers [2]
- XInclude href attribute [3]
- XLink href attribute [4]
- XLink role and arcrole attributes [5]
- XML Schema anyURI datatype [6]

The principle we have followed in these cases is that XML documents (and
web content in general) should wherever possible allow IRIs instead of
URIs, and that processors of those documents should perform the
conversion of IRIs to URIs when interfacing with legacy URI components.

A question arises when attempting to apply this principle to Namespaces
in XML 1.1 [7].  Namespaces in XML 1.0 [8] states that "An XML namespace
is ... identified by a URI reference."  In Namespaces in XML 1.1 we
state instead that "An XML namespace is ... identified by an IRI
reference."  We see this as a desirable change, for all the same reasons
that IRIs are generally preferable to URIs.  The most common namespace
operation, that of comparing namespace names using string equivalence,
would be unaffected by this change.

However, we have received feedback that this "upgrade" is not without
some cost [9].  In particular, applications may be relying on the
constraint that namespace names be restricted to URIs, and performing
additional processing based on this assumption. 

The XML Core WG would like TAG input on whether the desirability of
adopting IRIs into the web infrastructure early outweighs the
anticipated disruption of legacy systems.

The XML Core WG would also like TAG input on the wisdom of early
adoption given the "Internet Draft" status of the IRI draft [10].  So
far adoption has relied on "copy and paste", but there is potential for
these definitions to get out of sync.

[1] http://www.w3.org/International/2002/draft-duerst-iri-01.txt
[2] http://www.w3.org/XML/xml-V10-2e-errata#E26
[3] http://www.w3.org/TR/xinclude/#include-location
[4] http://www.w3.org/TR/xlink/#link-locators
[5] http://www.w3.org/TR/xlink/#link-semantics
[6] http://www.w3.org/TR/xmlschema-2/#anyURI
[7] http://www.w3.org/TR/xml-names11/
[8] http://www.w3.org/TR/REC-xml-names/
[10] http://www.w3.org/International/iri-edit/
Received on Wednesday, 9 October 2002 12:11:11 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:55:54 UTC