EARL 1.0: aggregation and inferences

I found in the working draft three references to test aggregation :

(1) earl:heuristic 

Assertion was made by inference, for example based on several existing test 
results

(2) earl:TestCase

An atomic test, usually one that is a partial test for a requirement. For 
example, checking if an image has an alt attribute which could be part of 
testing WCAG 1.0 Checkpoint 1.1.

(3) Test Subject dct:hasPart and dtc:isPartOf

In order to allow agents to deals with EARL documents in a uniform manner, 
without caring of test specific matters, i think that the inter-test 
relations should be explicited in the schema definition. For example to make 
explicit hyerarchical relations should be helpful for a general purpose EARL 
to HTML converter.

About (1) and (2) i think that expliciting the dependency of a test from other 
ones may be enough general and exhaustive. To create a new class 
for "inferenced" test should be a good idea.

About (3) I have no ideas, probably may be good to put into the EARL schema 
definition(rdf/xml)  to put a "dummy" cardinality constraint(>=0) about the 
suggested but not mandatory properties.

Thank you and sorry for my english,
Cristiano Longo

Received on Tuesday, 4 December 2007 09:43:01 UTC