- From: Kohsuke Kawaguchi <Kohsuke.Kawaguchi@Sun.COM>
- Date: Mon, 26 Feb 2007 10:16:40 -0800
- To: "C. M. Sperberg-McQueen" <cmsmcq@acm.org>
- CC: Leonid Arbouzov <Leonid.Arbouzov@Sun.COM>, www-xml-schema-comments@w3.org
- Message-ID: <45E32408.3090801@Sun.COM>
C. M. Sperberg-McQueen wrote: > The Working Group did not officially record a rationale for this > decision; speaking for myself, I would offer the reasoning that (a) > it's not quite clear what is being suggested, and in particular how > what is being suggested differs from the 'namespace schema information > information items' which are the content of the 'schema information' > property on the validation root, both in XML Schema 1.0 > (http://www.w3.org/TR/xmlschema-1/#d0e17275) and in our current 1.1 > draft (http://www.w3.org/TR/xmlschema11-1/#d0e25780), and No, this is not about PSVI. My understanding is that, generally speaking, XML Schema considers the component model to be the truth and the the XML syntax of the XML Schema to be just one of the many possible ways to represent them. What I wanted to point out is that today there's no way to associate documentation and appinfos to a particular namespace. You can attach them to element declarations, complex types (among many others), as well as the schema-as-a-whole, but the current component model simply lacks a way to "annotate" a namespace, because there's no schema component that represents a namespace. If there's perhaps a "namespace" schema component, then it would have annotations and appinfos as a property, and this provides the most natural place to put such namespace-specific annotation into it. > (b) judging > from the information available, the cost/benefit ratio of the change > falls on the wrong side of the place where the Working Group has drawn > the line for 1.1. The Working Group has been willing to change the > component structure only when absolutely necessary, so the effective > cost of the proposal is high. And the benefit is not as clear as it > would need to be to justify that cost. I certainly understand that changing a component model is a costly change, and if the WG felt that the benefit by this change doesn't justify that, then I respect that decision. -- Kohsuke Kawaguchi Sun Microsystems kohsuke.kawaguchi@sun.com
Received on Monday, 26 February 2007 18:17:53 UTC