Re: Other comments on EARL 1.0 Schema (editor's draft)

Hi Christophe,

Thank you for your comments, they have been added to the issues list:
  - <http://www.w3.org/WAI/ER/EARL10/issues>

Please find below some initial responses:


On 8.4.2011 16:03, Christophe Strobbe wrote:
> Dear ERT WG,
>
> Below are a few non-editorial comments on EARL 1.0 Schema (editor's draft
> <http://www.w3.org/WAI/ER/EARL10/WD-EARL10-Schema-20090908>).
>
> 1.1 Audience
> "The assumed audience of this specification is developers who are
> implementing EARL in software or processes..." ->  I would expect the
> audience to be evaluation tool developers and developers of tools that
> support manual and semi-automated evaluation. The current wording seems
> somewhat circular (as if you were saying that the audience of WCAG is only
> developers implementing WCAG, when you really want to all web developers to
> implement it).

Agree, this wording should be improved.


> 2.2 Assertor class
> Example 6: Bob using Cool Tool
> The tool is identified by means of foaf:member:<foaf:member
> rdf:resource="http://www.example.org/tool/#cool"/>
> but foaf:member<http://xmlns.com/foaf/spec/#term_member>  does not seem to
> cover software. Could earl:Software be used here instead?

In RDF this means that "Cool Tool" is both an earl:Software (as per 
example 5) and a foaf:Agent (because of the range of foaf:member).

It makes sense to identify the Software as an Agent in this case, so the 
example is semantically correct. Actually the group has previously 
discussed making earl:Software a sub-class of foaf:Agent but decided 
this may not be the case when the software is used as a test subject.

Having said all that, I agree that this example could be clarified or 
changed to avoid confusion between RDF semantics and XML syntax...


> 2.3. TestSubject Class
> How should EARL implementors identify a document that is not available on
> the internet? foaf:document does not seem to have an ID-like property. Some
> document formats, e.g. DAISY, get an automatically generated ID when they
> are created, but others, e.g. the OpenDocument Format, don't have an ID. In
> ODF, it is possible to embed the EARL document in the ODF/ZIP format. Can
> this method be used instead of an ID?

RDF uses URIs to identify resources. A local file could have a local URI 
such as "file:///C:/myfiles/myfile.odf".

Does this answer your question?


> 2.4. TestCriterion Class
> The examples in this section rely on publicly available criteria. What if
> you use criteria that are only available inside an organisation and that
> are not necessarily retrievable through HTTP?

Again, the basis of EARL is URI, which supports protocols other than 
HTTP (including proprietary ones if needed). The group had previously 
decided that resources that cannot be represented by URIs are out of 
scope of EARL. Do you have a specific scenario in mind?

Thanks,
   Shadi


> Best regards,
>
> Christophe
>

-- 
Shadi Abou-Zahra - http://www.w3.org/People/shadi/ |
   WAI International Program Office Activity Lead   |
  W3C Evaluation & Repair Tools Working Group Chair |

Received on Tuesday, 12 April 2011 07:49:22 UTC