W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > April to June 2006

Re: WCAG 2.0 Conformance Claims (from ERT list)

From: Tim Boland <frederick.boland@nist.gov>
Date: Mon, 15 May 2006 09:18:20 -0400
Message-Id: <>
To: w3c-wai-gl@w3.org
Cc: shadi@w3.org

Quality Assurance Framework: Specification Guidelines (QA SpecGL)
  Good Practice 3: "Provide the wording for conformance claims" [1] may be
a resource for the discussion excerpted following.

In addition, both QASpecGL Good Practice 4 - "Provide an Implementation 
Statement pro forma" [2], and QASpecGL Good Practice 5 -
"Require an Implementation Conformance Statement as part of conformance 
claims" [3]
may be helpful in the discussion excerpted following (as well as in future 

Christophe, thanks for the pointer to this discussion!

Thanks and best wishes
Tim Boland NIST

[1]: http://www.w3.org/TR/qaframe-spec/#conformance-claim-gp
[2]: http://www.w3.org/TR/qaframe-spec/#ics-gp
[3]: http://www.w3.org/TR/qaframe-spec/#ics-claim-gp

At 11:17 AM 5/15/2006 +0200, you wrote:

>Last week, there was a discussion about the scope of conformance claims on 
>the ERT list. Below is an excerpt from the discussion.
>For the whole discussion, see the part of the thread that starts at 
>On 9 May 2006, Carlos Velasco asked
>what happens when a delivery unit (...) has a CSS like 
>http://example.com/css/example.css? It is in scope or not?
>Christophe Strobbe responded:
>A stylesheet is not a Web unit but an "authored unit", and when it is 
>"intended to be used as a part of another authored unit", it is an 
>"authored component" 
>Conformance claims apply to Web units, not to authored units. So if a Web 
>unit within the scope of your conformance claim uses a CSS that is outside 
>the scope of your conformance claim, my understanding is that the 
>conformance claim applies to the Web unit *with* the CSS (but not to the 
>CSS in isolation).
>Then Johannes Koch wrote back:
>Let's make it a little more complicated :-) As Carlos Iglesias quoted, a 
>Web Unit is
>   "A collection of information, consisting of one or more resources,
>   intended to be rendered together, and identified by a single Uniform
>   Resource Identifier (such as URLs)".
>Let's consider an HTML document with various linked stylesheets, one for 
>screen, one for print, one for projection, ... They are not intended to be 
>rendered (all) together. So which CSS file belongs to the Web Unit 
>identified by the HTML document's URL? Or are there different Web Units 
>for each CSS file together with the HTML document. But how do I identify 
>them? The HTML document has only one URL.
>Christophe Strobbe:
>You can't make assumptions about how the type of user agent that will be 
>used to render the HTML document: on a screen, through projection, on a 
>refreshable braille display, with speech synthesis, ... So if the HTML 
>document, say 'home.htm', has stylesheets for each of these media types, 
>the following would all count as Web units:
>- home.htm with the CSS for 'screen',
>- home.htm with the CSS for 'projection',
>- home.htm with the CSS for 'braille',
>- home.htm with the CSS for 'aural',
>- ...
>So the question becomes: if you have a conformance claim that has the 
>above 'home.htm' in its scope, does the conformance claim cover each of 
>the above examples? Since you can't make assumptions about the user agent, 
>I think the answer is yes.
>(How you represent this in EARL is another matter.) I don't think it makes 
>sense to define the Web unit as home.htm + screen.css + projection.css + 
>braille.css, ...
>Come to think of it: if you use an object element with one or more 
>fallbacks nested inside it (see the example slightly below 
>I think the Web unit you claim conformance for is the HTML document with 
>the outermost object element (with the TheEarth.py applet). Now let us 
>look at a few scenarios.
>1. Let us assume that Python applets are not in the baseline and that 
>there is a fallback for each Python applet using a technology within the 
>baseline, then the conformance claim applies to the fallback (and the 
>Python applets must not interfere with the content). I don't know if EARL 
>can take baselines into account (I'm too tired to check now). Maybe a 
>"Baseline Description Language" would be useful?
>2. Alternatively, let us assume that Python applets are in the baseline 
>and there is a fallback for each Python applet using another technology 
>within the baseline. What happens if the Python applet cannot be retrieved 
>from the server, so that the fallback is used instead. How does this 
>affect the conformance claim? Can EARL make statements about the HTML 
>document + fallback object?
>3. Let us assume that Python applets are in the baseline and there is a 
>fallback for each Python applet using another technology outside the 
>baseline. (I think this is not very logical, but it is possible.) What 
>happens if the Python applet cannot be retrieved from the server, so that 
>the fallback is used instead. How does this affect the conformance claim?
>Johannes Koch, about the HTML file with different stylesheets per media type:
>(...) how do you identify the different web units? The URL for the HTML 
>document is not enough, because the HTML document is the same in all the 
>web units.
>Christophe Strobbe
>K.U.Leuven - Departement of Electrical Engineering - Research Group on 
>Document Architectures
>Kasteelpark Arenberg 10 - 3001 Leuven-Heverlee - BELGIUM
>tel: +32 16 32 85 51
>Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm
Received on Monday, 15 May 2006 13:19:04 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 16 January 2018 15:34:00 UTC