- From: Al Gilman <Alfred.S.Gilman@IEEE.org>
- Date: Tue, 9 Sep 2008 14:04:07 -0400
- To: "wai-xtech@w3.org WAI-XTECH" <wai-xtech@w3.org>
- Cc: Doug Schepers <schepers@w3.org>
** background / on the process Doug Schepers, the Team Contact from SVG WG had an idea based on their design work for SVG Tiny 1.2. Note that this is rather mature, and will likely advance up the Rec track ahead of WAI-ARIA. http://lists.w3.org/Archives/Public/public-pfwg-comments/2008JulSep/ 0018.html I have opened up an issue for consideration by the Working Group because he was not satisfied with the response he got from one of our Participants. This is ISSUE-75 http://www.w3.org/WAI/PF/Group/track/issues/75 -- Member-confidential link. Please discuss this issue on wai-xtech@w3.org because it involves issues of interest to UAAG and WCAG, and because SVG WG works in public. *** summary please review the pro and con discussed below Next step options include * SVG coins new element * SVG coins new role value (for definition in SVG), requests comment from XHTML2 and PFWG. * PFWG adds new document role *** rambles / on the product ** arguments pro * context help use case is endemic Ever since the U.S. Social Security Administration went to Freedom Scientific and got them to recognize a non-standard @contexthelp attribute, I have considered it a priority for us to have a good solution for context help. * improves author comprehension Various of our document roles are not essential for the API binding, rather being a heuristic hint or aid to the author. In particular "description" is in this category. The essential information is captured in the @aria-describedby arc pointing at the element containing the description. The @role='description' adds no new information, it does help the author keep straight what they are doing there. * style guide makes this distinction The keyboard command vocabulary recommended by our friends in the Style Guide caucus distinguishes popup or baloon ask-for help from general onMouseOver- appearing content. http://dev.aol.com/dhtml_style_guide#popuphelp ** arguments con * little red hen There is no need for SVG to be dependent on a @role value introduced in the WAI-ARIA spec to denote the intended semantic distinction (this is about how-to content, not about onMouseOver behavior). SVG can just do it themselves without depending on PFWG action. There are a variety of ways that SVG could just do it themselves. An element type, a role value, or otherwise. SVG can coin a <help> or <contexthelp> element in SVG that has SVG- specified behavior. See the Style Guide for currently proposed command profile for this function http://dev.aol.com/dhtml_style_guide#popuphelp SVG could propose to define a value for @role. This would be a document role, it is not in contention with the PFWG-controlled widget roles. So far the XHTML2 and PFWG working groups are the two in the community adding role names to this vocabulary, but there is no golden rule that says we should own it. I would argue with PFWG and XHTML2 that we should invite SVG to join the swim and coordinate additional terms there as required. [But I think an element name is here a better solution]. In other words, while you have the host-language format open for editing, don't defer to an @role value semantics that you can do in the native markup of the langugage. Just as PFWG decided to mirror the <html5:article> semantics in a role symbol, we could eventually see the wisdom of your ways and add such a role for use of the same semantic in other host languages. A related argument for SVG to do this in SVG rather than in WAI-ARIA is because of the order of maturity. * general semantics later We recognized early that although metadata can be introduced by attributes linking to RDF, and that all kinds of metadata could be handy in deriving adapted user experiences, that for 1.0 we were going to include those things necessary for the user to (a) understand how to operate the scripted behavior and (b) understand how to navigate the pages containing the scripted behavior. Roles for every purpose notion that one could imagine was clearly going to bloat ARIA beyond what we could finish and publish. So we deferred role extension and general semantic annotation beyond WAI-ARIA 1.0. One could argue that the distinction between context help and random popup content (given that the popup behavior and how to get to the information with the keyboard is covered by the @aria-haspopup semantics) is general semantics not meeting this test for inclusion in WAI-ARIA 1.0. *** please discuss Al
Received on Tuesday, 9 September 2008 18:04:54 UTC