W3C home > Mailing lists > Public > public-wai-ert@w3.org > May 2006

Re: Explicit vs Implicit URIs (aka "blanket statements or not")

From: Carlos A Velasco <Carlos.Velasco@fit.fraunhofer.de>
Date: Wed, 03 May 2006 15:32:47 +0200
Message-ID: <4458B0FF.6020705@fit.fraunhofer.de>
To: Shadi Abou-Zahra <shadi@w3.org>
Cc: public-wai-ert@w3.org

Hi all,

Two Euro-cents before the call:

Shadi Abou-Zahra wrote:
> ...
> There are two related but yet quite different use cases for implicit URI 
> (e.g. "example.org" means all resources available under this domain name):
> 1. to facilitate more compact EARL reports

I disagree with this use case. "Compactability" was not one of the EARL 
requirements. To facilitate aggregation yes, but not this one.

> 2. to enable blanket statements such as conformance claims

Then, let us create a Conformance Claim Reporting Language, but not 
EARL. BTW, I am not volunteering to write that note ;-) Seriously, 
blanket statements can be presented in myriads of ways, basically, one 
per person wanting  a summary (and we have different experiences with 
that and imergo). However, I don't think EARL is the tool to do that.

> CarlosI took up an action item to work out a draft proposal which he 
> sent to the list on Monday 10 April, 2006; and we had subsequent 
> discussions on Wednesday 19 & 26 April, 2006:
>  <http://lists.w3.org/Archives/Public/public-wai-ert/2006Apr/0010>
>  <http://www.w3.org/2006/04/19-er-minutes#item02>
>  <http://www.w3.org/2006/04/26-er-minutes#item02>
> It turns out that such implicit URIs also bring about ambiguity. For 
> example, if a new page is added to "example.org" after an assertion was 
> made, is then the assertion result still valid? Similarly, for blanket 
> statements such as conformance claims, it is unlikely that all pages 
> under "example.org" have been tested but probably only a sampling. It is 
> therefore rather imprecise (but still useful) to give a blanket 
> statement without further description of what has been tested, and which 
> methodology has been used to test.

The ambiguity is more on "which resources are really *under* 
example.org, than with the moment on which the assertion was made. Is 
images.example.org included, or www-intern.example.org included (I owe 
this one to JK)? How did you crawl all resources? Did you include all 
possible POST variants in forms? Did you get also the error messages 
from the server?

> The currently suggested proposal is that EARL should only focus on 
> recording actual test results (ie. no implicit URIs, only explicit 
> ones). In some cases, RDF features such as collections may be suitable 
> to reduce verbosity (still, every tested URI will need to be recorded at 
> least once per report). As to blanket conformance claims, other 
> vocabularies (preferably RDF-CL) should be able to provide the required 
> functionality of expressing these, and pointing back to the EARL report 
> for more detail on what has been tested.


> * "example.org/page.html passes WCAG 1.0 CP 1.1"
> - the "usual" way of using EARL to record test results will still be 
> available
> * "example.org/page.html passes WCAG 1.0 Level A"
> - a contraction of individual checkpoints is possible due to the WCAG 
> hierarchy

It depends. There is still an open discussion on TestCase and 
TestRequirements. Until that discussion is closed, some of the above 
examples might not be valid.

> * "example.org conforms to WCAG 1.0 Level A"
> - not to be expressed by EARL but by a different set of vocabulary (RDF-CL)

I agree.

> We need to confirm that this is indeed the approach we want to take with 
> EARL: it focuses only on recording tests, and relies on other 
> vocabularies for blanket conformance claims. Reports would also become 
> potentially more verbose but also more precise as a result of this 
> decision.


Dr Carlos A Velasco - http://access.fit.fraunhofer.de/
Fraunhofer-Institut für Angewandte Informationstechnik FIT
   [Fraunhofer Institute for Applied Information Technology (FIT)]
   Barrierefreie Informations- und Kommunikationstechnologie für Alle
   Schloss Birlinghoven, D53757 Sankt Augustin (Germany)
   Tel: +49-2241-142609 Fax: +49-2241-1442609
Received on Wednesday, 3 May 2006 13:33:38 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:55:54 UTC