EARL and describing tests Re: Agenda for F2F at TP

(So people have some idea of what I will say at the appropriate moment ;-)

Hmm. EARL provides some simple (and useful) ways of describing tests. An  
example is the descriptions that can be extracted from among other  
information in  
http://www.w3.org/2001/sw/Europe/200305/axforms/earlinst.rdf - the  
relevant parts are as follows:

<earl:TestCase rdf:about="http://www.w3.org/TR/WCAG10/#tech-color-convey">
    <wcag:priority  
rdf:resource="http://www.w3.org/TR/WCAG10/#wc-priority-1"/>
    <wcag:checkpoint>2.1 </wcag:checkpoint>
    <dc:title xml:lang="es">Asegúrese de que toda la información  
transmitida a través del color está también disponible sin color</dc:title>
    <dc:title xml:lang="en">Ensure that all information conveyed with color  
is  also available without color,  for example from context or markup.  
</dc:title>
   </earl:TestCase>

An explanation:

This description relies on EARL solely for the fact that it has a class  
called TestCase - identifying something as a test that you might talk  
about in EARL.

I have made up some properties and a couple of resources for describing  
different priority levels - these are the things that have the namespace  
prefix wcag in the above example.

I have also provided a Dublin Core title in english and spanish. (It is an  
open question whether I should have used this property, or label and  
comment properties from the RDF Vocabulary Description Language - possible  
the latter solution would have been cleaner).

This information was collected to dynamically build an interface in the  
relevant language in an Xforms-based evaluation tool I wrote, but since it  
is available on the Web I can re-use it in any other application, or add  
to it.

What this doesn't do is decribe how to do the test. I think that it is  
useful to be able to describe procedures for people to follow, and code  
which can automatically run the test, but I don't see that EARL needs to  
do this. The nice thing about working in RDF is that we don't ned to build  
everything into our one vocabulary anyway...

Nor does it describe the relationship between a group of tests, which I  
think should be done using OWL restrictions rather than inventing  
EARL-specific stuff that does the same thing.

But we do have same basic ability to describe tests. We could use Dublin  
Core (for example) to give more detailed information about who is the  
author of a particular test, when it was developed, etc. To make this  
easy, all we do is keep using RDF.

On the other hand, it is useful to have some idea of what we think best  
practices are for some common descriptions...

cheers

Chaals

On Mon, 28 Feb 2005 00:54:19 +0100, Shadi Abou-Zahra <shadi@w3.org> wrote:

>> Does the fact that test descriptions are seperate mean that they are not
>> in scope for the general requirements discussion? (Makes sense to me,
>> actually...)

> As you know, EARL does not describe test cases itself but implicitly it  
> bears some basic assumptions
> about how tests are described. For example, the current spec assumes  
> that each test description has a
> (publicly available) URI.
> So idea for this session is to spend some time refining these  
> assumptions. Maybe developing some really
> rough requirements for test case descriptions. After that we can look  
> back at our EARL requirements and
> tweak them accordingly if necessary.



-- 
Charles McCathieNevile - Vice Presidente - Fundacion Sidar
charles@sidar.org                      http://www.sidar.org
     (chaals is available for consulting at the moment)

Received on Monday, 28 February 2005 21:53:01 UTC