- From: Chimezie Ogbuji <ogbujic@bio.ri.ccf.org>
- Date: Mon, 19 Mar 2007 09:39:50 -0400 (EDT)
- To: public-grddl-wg <public-grddl-wg@w3.org>
With regards to both of my actions [1] [2] registered in our last teleco, I've incorporated comments from Jeremy and Dan, made some additional edits, finished up the body of the test document, cleaned up supporting text for certain tests, removed as many absolute references as I could (there were a few), had a thorough walk-thru of the tests, and put together some publication instructions as there are certain server requirements (content negotiation, redirects, and such) that are necessary for this document's functionality. The URL for the draft again is: http://www.w3.org/2001/sw/grddl-wg/doc50/grddl-tests.html Noteworthy, was the removal of the XSLT literal result test as it was broken per the current spec - see notes below A class='NetworkedTest' attribute was added on <li> elements for tests which can not be run remotely, this is extracted into additional statements in the GRDDL result of the test document. I also had to make an assumption that http://www.w3.org/TR/grddl-tests/ would be the base URI of the test document once published (I attempted to convert to as many relative URIs as I could in case this assumption was wrong) The zip was created by an automated script [3] which used the GRDDL result of the test document as an index for identifying the files to archive, so at least that verifies that the links from input to outputs are all valid. The test zip only includes 'localized' tests (i.e., those tests that do not require network connectivity to run). I also checked [4] in GRDDL.py to the test directory. Previous links to test cases from the main specification will need to be updated. For instance, the link to the atomttl1 test from #issue-output-formats: http://www.w3.org/TR/grddl/#issue-output-formats and the link to the Xinclude test from #faithful-infoset Additional notes (detailing changes by files): ---------------------------------------------- Assumed base URI for tests: http://www.w3.org/TR/grddl-tests/ Misc. notes: Fixed absolute -> relative URI's in: - conneg.html.de - conneg.html.en - grddlonrdf.rdf - titleauthor.html - xinclude1.html loop.xml - added self document reference - fixed absolute URI testhtml3 - langconneg3 links were fixed (were pointing to wrong result documents) aboutTests.xsl: - Added RDFS document for subsumes/alternative terms and NetworkedTest class - Fixed absolute URIs in GRDDL result of test document ## Special redirects: ## sq1ns -> sq1ns.xml sq2ns -> sq2ns.rdf loop-ns-b -> loop-ns-b.xml ## Special response headers ## grddlonrdf-xmlmediatype.rdf -> Content-Type: application/xml ## Content / Language negotiation ## conneg.html -> conneg.html.en (Accept-Language: en) conneg.html -> conneg.html.en (Accept-Language: de) Broken tests ------------ litres.xml seems IMHO. The output indicates that the source document does *not* have a GRDDL result. Even if served up as application/xml, a GRDDL-aware agent would still be required (by the base rule, see: http://www.w3.org/TR/grddl/#txforms) to compute the RDF graph that corresponds to parsing the document *as-is* into a GRDDL result. The fact that it has an @xsl:version attribute on it's document element is of no consequence to GRDDL and the second dc:description *does* indeed parse as RDF/XML: @prefix _10: <http://www.w3.org/2001/tag/issues.html#>. @prefix _8: <http://www.w3.org/1999/XSL/>. @prefix _9: <http://www.w3.org/TR/xslt#>. @prefix dc: <http://purl.org/dc/elements/1.1/>. @prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>. @prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#>. <litres.xml> dc:description """The start tag of the root element of this document seems to say that it's an RDF/XML document. But it also says that it's an XSLT literal result element. And the element after this one doesn't parse as RDF/XML. But after you run it thru an XSLT processor, the result *is* RDF/XML. So when it's served up as application/xml, what GRDDL results does it have, if any? """, [ a _8:Transformvalue-of; <select> "2+2"]; rdfs:seeAlso _10:nsMediaType-3, _10:xmlFunctions-34, _9:result-element-stylesheet. There would be no reason to run it through an XSLT processor as no transforms are nominated. ---------------------------------------------- [1] http://www.w3.org/2007/03/14-grddl-wg-minutes.html#action03 [2] http://www.w3.org/2007/03/14-grddl-wg-minutes.html#action07 [3] http://www.w3.org/2001/sw/grddl-wg/td/archive-tests.py [4] http://www.w3.org/2001/sw/grddl-wg/td/GRDDL.py Chimezie Ogbuji Lead Systems Analyst Thoracic and Cardiovascular Surgery Cleveland Clinic Foundation 9500 Euclid Avenue/ W26 Cleveland, Ohio 44195 Office: (216)444-8593 ogbujic@ccf.org
Received on Monday, 19 March 2007 13:39:57 UTC