- From: Simon <simonstl@simonstl.com>
- Date: 19 Jun 2001 11:44:38 -0400
- To: Vun "Kannon," David <dvunkannon@kpmg.com>
- Cc: "'www-xml-linking-comments@w3.org'" <www-xml-linking-comments@w3.org>
On 19 Jun 2001 11:24:52 -0400, Vun Kannon, David wrote: > The XLink spec currently defines the attributes of the Xlink > namespace using DTDs. There are a few aspects of the new XML Schema > specification which may be helpful to XLink applications. Defining XLink in > terms of XML Schema will allow a certain level of validation of the XLink > information by the general process of schema validation. This could include > namespace correctness, datatyping and enumeration values and defaults, and > uniqueness constraints. > XLink currently uses the xlink:type attribute to distinguish various > kinds of elements. This function could be handled in XML Schema by abstract > element definitions and the use of element substitution groups. > > My request is that the XLink specification include a normative XML > Schema as part of the spec. While I wouldn't necessarily object to inclusion of a normative XML Schema, I would strongly oppose any effort to make XLink processing dependent on XML Schema processing. While xlink:type could perhaps be abolished by abstract elements and the use of element substitution groups, doing so would render XLink far less useful to a large group of developers who are working with DTDs, alternative schema languages (notably RELAX NG), or with no schema support whatsoever. Simon St.Laurent
Received on Tuesday, 19 June 2001 11:43:43 UTC