- From: Al Gilman <asgilman@iamdigex.net>
- Date: Thu, 14 Dec 2000 16:55:33 -0500
- To: Wendy A Chisholm <wendy@w3.org>, w3c-wai-gl@w3.org
At 03:48 PM 2000-12-14 -0500, Wendy A Chisholm wrote: > >However, here's a new wrench: >Anne suggested that in my example for checkpoint 1.1 of a right arrow that >links to the next slide in a slide set, that the image ought to contain the >text "Next." >[<http://lists.w3.org/Archives/Public/w3c-wai-gl/2000OctDec/0478.html%5D>h ttp://lists.w3.org/Archives/Public/w3c-wai-gl/2000OctDec/0478.html] In >this case, does this qualify as a "graphical function"? > AG:: Not new. Not graphical function. The semantics, here, are primarily operational (move to next page in series). The icon and the word 'next' are, either of them, generally sufficient to evoke this semantic. Just not generally enough. Neither of them is universally accessible. So we want to have both. Yes it is valuable to have the text in the icon. This is blending the modes of communication. This is more important for the quantitative effectiveness of the curb cut effect (benefit to the non-disabled of having both text and image) than it is critical to the access safety net (avoid all-out failure for people with disability). The safety net is complete so long as there is text and there is image. The example should still be a universal design that makes the most effective use of the text and image, not merely satisfies the minimum requirement of "failure to shut people out." Many Aps have UI preferences as to whether to show icons, words, or both. The 3.1 rewrite as it stands is valid for this example. The function of the word in this case is not primarily visual, it is primarily verbal. The best case is still that the icon is drawn by SVG which takes the text from its internal text content. An acceptable compromise is that the icon is a bitmapped image and the text is mirrored in the ALT attribute for the button. [discussion of ALT vs. TITLE suppressed.] Am I reading the rewrite right? Al PS: This reminds me of one of my hobby-horses I have been failing to spit out. Where the content guidelines presently talk about other stuff 'supplementing' textual information, they should rather say that the other modes of expression should be used in a way that 'complements' the verbal expression captured as standard text. While text is first among equals in providing supplemental materials specifically for access, the overall universal design of multimedia includes all cases of leader/follower/peer relationships among the different ways (verbal, visual, auditory, smellyvision, ...) of expressing content. The general model is therefore symmetrical, and we should use the symmetrical language "they complement each other" as opposed to the polarized "one supplements the other" in the top level view. The media expression of the 'next' function here is a good archtypal example: Neither the icon nor the word is dependent on the presence of the other; either is sufficient in and of itself in the main. The communication effectiveness is broadened to cover the preponderance of people with disabilities by providing redundant media expressions using both text language and icon. The redundancy increases the effectiveness of communication with people without articulable disabilities in fact. This 'curb cut effect' is because even 'though many people fail to have functional impairments to the level recognized as a disability, they still have marked differences that make the icon communicate better for some and the word communicate better for others. The time to comprehend and the errors in comprehension are measurably reduced for the general population without respect to disability. In this quantitative measurement, I would expect there to be a statistically significant improvement for the icon with text embedded over the icon used in such a way that it took user action or assistive technology to get at the text. That is to say I would lay money that a competent independent Human Factors experiment would confirm what Anne has said. Al
Received on Thursday, 14 December 2000 16:55:12 UTC