- From: Mary Brady <mbrady@nist.gov>
- Date: Mon, 9 Jul 2001 11:00:58 -0400
- To: "Curt Arnold" <carnold@houston.rr.com>, <www-dom-ts@w3.org>
- Cc: <www-dom-ts-submission@w3.org>
The information in the test-matrix.html file only captures info regarding the test purposes (or semantic requirements), not necessarily all of the metadata for a particular test. There is an additional description for a test that more fully describes what happens in the test itself, and should be used to supplement the test purpose. It is this information that resides in an older rendition that needs to be brought into the metadata file. I would also suggest that it, along with author, possibly creation date, revision date info be carried along with each test, and translated into prologue documentation -- probably javadoc for the java transformation. I noticed that you are still using rdf. Why? --Mary ----- Original Message ----- From: "Curt Arnold" <carnold@houston.rr.com> To: <www-dom-ts@w3.org> Cc: <www-dom-ts-submission@w3.org> Sent: Friday, July 06, 2001 1:11 AM Subject: nistmeta.xsl > I've made up a little XSLT transform that creates RDF metadata for > the NIST tests by analyzing the Test Matrix at > http://xw2k.sdct.itl.nist.gov/brady/w3c-dom/test-matrix.html I first needed > to make that document a legal XML document by adding an /> to <br> and <col> > elements. > > Seems to generate garbage for the HTML tests, but does an okay job on the > XML tests. Basically, it scans the description of the test for any > parameter name, exception code or 'returns' and emits a corresponding > <dc:subject> element for the parameter, exception code or return values with > a resource URI that should be the same as the corresponding entry in > subjects.rdf. For example: > > <rdf:Description > about="level-both/xml/nodeAppendChildNewChildDiffDocument.xml"> > <dc:title>nodeAppendChildNewChildDiffDocument</dc:title> > <dc:description>The appendChild method raises a WRONG_DOCUMENT_ERR > DOMexception if newChild was created from a different document than the one > that created this node.</dc:description> > <dc:subject > resource="http://www.w3.org/TR/1998/REC-DOM-Level-1-19981001/level-one-core# > ID-184E7107"/> > <dc:subject > resource="http://www.w3.org/TR/1998/REC-DOM-Level-1-19981001/level-one-core# > xpointer(id('ID-184E7107')/raises/exception[@name='DOMException']/descr/p[su > bstring-before(.':')='WRONG_DOCUMENT_ERR'])"/> > <dc:subject > resource="http://www.w3.org/TR/1998/REC-DOM-Level-1-19981001/level-one-core# > xpointer(id('ID-184E7107')/parameters/param[@name='newChild'])"/> > <dc:creator>NIST</dc:creator> > </rdf:Description> > > The nistmeta.xsl is in the xmlconf CVS (viewable at > http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/xmlconf/www-dom-ts/nistmeta.x > sl). The sample of the generated output is at > http://home.houston.rr.com/curta/domtest/tests.rdf > > The metadata processing code could append extracted in-test metadata with an > external metadata source to produce an overall external metadata file that > would be used for the test matrix and other generation. > > The values of the about parameters would need to be generated to reflect the > eventual w3.org locator. > > What it the significance of level-both? Wouldn't any test that tests "both" > levels really be a DOM level 1 test? > > >
Received on Monday, 9 July 2001 11:02:39 UTC