- From: Dimitris Dimitriadis <dimitris.dimitriadis@improve.se>
- Date: Wed, 11 Jul 2001 19:14:05 +0200
- To: "'Arnold, Curt '" <Curt.Arnold@hyprotech.com>, "''www-dom-ts@w3.org' '" <www-dom-ts@w3.org>
comments inlined -----Original Message----- From: Arnold, Curt To: 'www-dom-ts@w3.org' Sent: 2001-07-09 22:09 Subject: Suites and metadata I had deferred doing anything with test suite definitions, but it seems like things are progressing to the point that we will need to address them shortly. I think that it is fairly obvious that we will have a <suite> element. What goes underneath is not so clear. <suite> <!-- conditions and metadata prolog like test --> <metadata>...</metadata> <test href="characterDataSubstringIndexSizeNegative.xml"/> <test href="characterDataSubstringIndexSizeTooLarge.xml"/> </suite> is attractive, but would complicate DTD validation since <test> is already defined with a different content model, <member> would also be out for the same reason. Any comments on: <suite> <suiteMember href="..."/> </suite> or <suite> <suite.test href="..."/> <suite.suite href="..."/> </suite> [dd] suiteMember sounds better to me, allows both suites and tests. Also, I had been using XML/RDF represent extracted metadata from tests and other sources though the test themselves only had RDF reminiscent elements and adding a dependency on a technology that none of us seem to really know well was not popular. The only construct that was really missing from the existing <metadata> construct in the test definition language was the ability to say that a metadata element was about some arbitrary URI instead of the test or suite that the metadata element was contained it. It would seem that we could express subjects.rdf and some of the other external metadata files within the internal metadata, if we add an optional about attribute to the <metadata> element and allow a <metadata> element to optional contain other <metadata> elements. That would not affect metadata in the tests, but it would allow external metadata compilations (which are now in an RDF format) to be validated by the DOMx.DTD. For example, subjects.rdf which now looks something like: <rdf:RDF ...> <rdf:Description about="uri to some passage in the DOM spec"> <dc:description>some text</dc:description> ... </rdf:Description> </rdf:RDF> could be expressed as: <!DOCTYPE metadata SYSTEM 'dom1.dtd'> <metadata> <metadata about="uri to some passage in the DOM spec"> <description>some text</description> </metadata> <metadata about="..."> <description>...</description> </metadata> </metadata> The resulting file could be trivially converted to RDF by a transform if that eventually added any value. I'd appreciate any feedback and will try to make appropriate changes when I commit the spec-to-xsd.xsl and spec-to-dtd.xsl transforms to the W3C CVS tonight.
Received on Wednesday, 11 July 2001 13:15:04 UTC