- From: Nils Ulltveit-Moe <nils@u-moe.no>
- Date: Tue, 12 Apr 2005 16:32:15 +0200
- To: public-wai-ert@w3.org, shadi@w3.org
Hi Shadi tir, 12,.04.2005 kl. 13.12 +0200, skrev Shadi Abou-Zahra: > Hi, > > At least for the earl:subject element, I think it may be useful to add both dc:hasPart and dc:isPartOf as optional properties. The reason is to be able to describe more complex structures and relationships between subjects, for example: > > * Scenario: > To make a transaction on a Web site, three pages (1, 2, and 3) need to be traversed. > > * EARL Subject: > <earl:subject rdf:resource="http://www.example.org/service/"> > <dc:hasPart rdf:resource="http://www.example.org/page1"/> > <dc:hasPart rdf:resource="http://www.example.org/page2"/> > <dc:hasPart rdf:resource="http://www.example.org/page3"/> > </earl:subject> > > Similar may be applicable to earl:testcase and even earl:assertion but the latter one can bring about unwanted implications on the earl:result and other properties of the assertions. I think it makes sense to explicitly describe which resources that was considered in the test. That may be used to show completeness of the test; e.g. what pages where considered in a scenario where some pages were sampled. With this information the test can be repeated by another tool doing the same sampling. It can also show why two tools diverges, e.g. if one tool is able to interpret a link embedded in Javascript that another tool does not identify, then explicitly showing which resources that were part of the test would reveal what the tools did differently, which would help tool makers as well. Mvh. -- Nils Ulltveit-Moe <nils@u-moe.no>
Received on Tuesday, 12 April 2005 14:28:25 UTC