- From: Wendy A Chisholm <wendy@w3.org>
- Date: Mon, 16 Oct 2000 15:27:37 -0400
- To: w3c-wai-gl@w3.org
Hello, I have reread this entire thread both on IG and GL. I have given this a lot of thought. I expect disagreement with my conclusions. The question is: Does checkpoint 3.1completely outlaw the use of text in images. <blockquote> 3.1 When an appropriate markup language exists, use markup rather than images to convey information. [Priority 2] For example, use MathML to mark up mathematical equations, and style sheets to format text and control layout. Also, avoid using images to represent text -- use text and style sheets instead. </blockquote> I have boiled the arguments into the following two positions: 1. Checkpoint 3.1 of WCAG 1.0 is open to interpretation because people believe that adequate support for "an appropriate markup language" does not exist. CSS and SVG as languages exist, but the support for them does not. Therefore, it <em>should be</em> possible to use images to create text effects and as long as those images have alt-text a site can claim WCAG 1.0 AA Conformance. 2. Checkpoint 3.1 of WCAG 1.0 is <em>not</em> open to interpretation because CSS and SVG exist and it is possible to create text effects in markup language. Therefore, it <em>should not be</em> possible to use images to create text effects and claim WCAG 1.0 AA Conformance. The primary reason for using markup is for the user to control the presentation, particularly to increase the font size. This is a priority 2 checkpoint. People can clearly claim WCAG 1.0 single A conformance and use text in images. Conformance to Double A becomes cloudy when we discuss logos or proprietary (or rare) fonts. It is possible to create logos in SVG, however SVG is not widely supported. As 3.1 is worded, "an appropriate markup language" exists, but the support does not. Therefore, as it is worded non of the W3C sites can claim higher than single A conformance because we include logos on the majority of our site. Note that many of those pages currently claim AAA conformance. Therefore, we can infer that we have been interpreting this checkpoint to mean that an appropriate markup language does not exist. In Opera it is possible to increase the font size of all text, including alt-text for images, up to 1000%. It is very easy to turn images on and off in Opera. One only need press "g" or click on a button on the menu bar. Therefore, if someone with low vision were using the page with opera and had difficulty reading the text in an image, they could easily toggle between seeing the image and reading the "magnified" alt-text. With the current state of browsers, I do not believe it is possible to avoid using text in images. Tools exist that will allow users to magnify the alt-text of those images. There are also tools that will read text out loud. I believe that many of the pages in the W3C site can continue to claim Triple A conformance to WCAG 1.0 because the appropriate markup language is not supported. As Lisa pointed out, our own WCAG conformance logo does not conform to checkpoint 3.1. Why? Because it is very important to establish the "brand" of both the W3C and WCAG, same with Bobby, and all other logos. By saying that people can not use logos and claim anything higher than a single A would make the other levels of conformance pointless. Therefore, I have to agree that sites may use text in images and claim Double A conformance. However, we should limit the use of text in images to only what is necessary for branding. I propose limiting text in images to logos and navigation buttons. It should not be possible to put an entire paragraph of text in an image and claim Double A conformance because you lose the structure of the document. The only case that remains that I find difficult to satisfy is the case that Cynthia pointed out where a company is using images in headings as part of branding. My concern is that people could claim "branding" for so many things to avoid converting something to markup. I believe this example violates another checkpoint, 3.5 Use header elements to convey document structure and use them according to specification. [Priority 2] Therefore, someone that wedded to appearance is probably only going to claim Single A and we should all be satisfied with that. I propose we add the following to the errata: 3.1 When an appropriate markup language exists, use markup rather than images to convey information. [Priority 2] For example, use MathML to mark up mathematical equations, and style sheets to format text and control layout. Also, avoid using images to represent text -- use text and style sheets instead -- except for logos and stylized navigation buttons used to create a distinctive look to a site. Thoughts? --wendy -- wendy a chisholm world wide web consortium web accessibility initiative madison, wi usa tel: +1 608 663 6346 /--
Received on Monday, 16 October 2000 15:19:47 UTC