W3C home > Mailing lists > Public > wai-xtech@w3.org > February 2007

Re: Updated ARIA docs for Firefox

From: Aaron Leventhal <aaronlev@moonset.net>
Date: Sun, 11 Feb 2007 15:38:00 -0500
Message-ID: <45CF7EA8.108@moonset.net>
To: boland@nist.gov
CC: wai-xtech@w3.org

Hi Tim,

Please bring up anything that doesn't make sense or agree with the spec. 
That said, this doc is meant as a developer guide for authors using the 
Firefox ARIA implementation, not as a spec. For example, in the specific 
case of our mention of tabindex that you've mentioned, it's only 
referring to the code example right before that table. It's merely 
explaining how ARIA works with a simple checkbox example -- not as a 
general definition.

On the other hand, the other tables in the document are more general -- 
and list what's supported in Firefox.

Basically what I'm looking to have someone check is this: where is there 
a difference in the ARIA features listed in our docs when compared with 
the spec? For example, I found that our docs had previously listed 
"multiselect" as a property name, while the spec called it 
"multiselectable". In reality, these docs, the specs and the RDF all 
need to be compared for differences.

- Aaron




boland@nist.gov wrote:
> Hi, I was wondering how specific ("formal"?) the consistency check should be, 
> in terms of the similarity of the prose used in the "Description" column of the 
> table following Section 3.2 ("Supported States and Properties")of [1], with the 
> prose used in the "meaning" column of the "basic support for assistive 
> technologies" section of the document referenced following. 
>
> For example, in the referenced document, for "tabindex=0", the "meaning" is 
> stated as "put the span-based checkbox in the tab navigation cycle, and follow 
> click-to-focus with the mouse".  In contrast, in the "Description" column of 
> [1], it is stated that "elements of types div, span, a.. etc., with a zero 
> value of tabindex are navigated next.. these elements are navigated in the 
> order they appear in the character stream".
>
> The language (prose) used is different in the two references, even 
> though "tabindex=0" is being discussed in both cases.  I think that language 
> (prose) may be somewhat subjective (open to different interpretations), and 
> that "meaning" may not be exactly the same as "description".  If the 
> description in Section 3.2 of the ARIA States and Properties Document 
> is "normative" (as I'm getting by implication from the Section 3 heading being 
> listed as "normative"), is it possible to start from this language in the 
> description and derive the "meaning" for implementation from that description 
> directly using as much of the same language as possible and document the 
> derivation?   Otherwise I'm concerned that the "meaning" may be a 
> reinterpretation of the "requirements" contained in the "description", and may 
> possibly affect inadvertantly those requirements.
>
> My apologies if I'm missing something..
>
> Thanks again for all this great work!
>
> Best wishes
> Tim Boland NIST
>
>     
>  Quoting Aaron Leventhal <aaronlev@moonset.net>:
>
>   
>>     
> http://developer.mozilla.org/en/docs/ARIA:_Accessible_Rich_Internet_Applications
>   
>> Feedback welcome. Would love if someone did some consistency checks with 
>> the ARIA role and state docs.
>>
>> - Aaron
>>
>>
>>     
>
>
>
>   
Received on Sunday, 11 February 2007 20:38:26 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 13:15:42 GMT