- From: Al Gilman <Alfred.S.Gilman@IEEE.org>
- Date: Tue, 26 Jun 2007 23:42:55 +0200
- To: public-comments-WCAG20@w3.org
Prior comments whose dispositions this message replies to: LC-955 LC-956 LC-958 LC-959 LC-971 see also disposition of LC-1208 The case tree in 1.1.1 as currently framed is too tortured. http://www.w3.org/TR/2007/WD-WCAG20-20070517/Overview.html#text-equiv-all We need to start with "associated text" not "equivalent text" and then we can specialize, not exclude, as we walk the cases. Define "associated text" as text with a machine-recognizable association with the non-text content (fragment or object). We do better developing a case tree by looking at the answers to the question "what can I do (with this [non-text] content fragment)?" Unless the non-text content (fragment or object) is part of a larger fragment that has associated text or different-media alternative that meets all requirements at that level (see response to LC-1208), Where do = experience (image of artwork, song, etc. media object) .. required text content identifies the object and AA content describes it Where do = learn (picture or diagram or other item with articulable information to convey) .. required text content expresses the information that one could learn from the non-text content. Where do = choose to browse or skip (section, including form or table) .. associated text (heading) answers "what is _there_?) Where do = go (hyperlink) .. associated text answers "where will I go?" Where do = other (controls and other widgets) .. associated text answers "what can I do?" Where information that text is to communicate is "no information," then associated text may be omitted if this "content free" condition can be recognized by AT from the encoding of the non-text content fragment. Note: examples include spacers [, etc. as listed in the current draft] Where do = complete a required task .. associated text or task-enabling alternate-media content affords the opportunity to reach the same task outcomes. Note that the "required task" concept allows a more general statement of a rule that is in the present draft only represented by the "process" discussion under conformance. This is that if some user task is a required prerequisite for receiving any services of the site, then if the prerequisite is not accessible (including the option of composite alternatives rather than just atomic alternatives), no of the content whose use is dependent on completing the inaccessible task may be claimed to be accessible. [yes please put in the "at the pertinent level" qualification as at present] This includes the 'process' case where defined that all subtasks in the process are required and also more general task graphs but where some subtask is an unavoidable prerequisite for an outcome or another task. [CAPTCHA is covered under this last. -- that can be a note.] [do not give multimedia a free ride.] [do not excuse content "that is not presented to users" -- PFWG and UAWG, in discussions with format specifications (in particular the resolution of the 'override' issue in SMIL) have taken the position that authors cannot make the final determination as to whether content is to be shown to users or not. This falls under the "author proposes, user disposes" management protocol for show/hide profiling of content. In other words, authors must treat all hidden content as conditional content. It doesn't have to be easy, but the user should have a way to drill down to the point that it gets displayed.] Al
Received on Tuesday, 26 June 2007 21:43:08 UTC