- From: Jim Ley <jim@jibbering.com>
- Date: Fri, 5 Nov 2004 16:09:32 -0000
- To: www-svg@w3.org
"Nigel McFarlane" <nrm@kingtide.com.au> wrote in message news:418BA2B7.7030803@kingtide.com.au... > Rather than pick nits over <canvas> or other tags, it would > be better if the "yea" sayers would defend SVG by explaining > how it *can* support accessibility, UA content extensions, and > XForms, and how much CSS it has to lean on to do so. How much > can it do? How SVG can support accessibility of a stock graph (something that Ian stated was better done with a canvas element) can be easily demonstrated. You can use sXBL to convert source data in an appropriate format (I imagine you'll find something appropriate at http://www.service-architecture.com/xml/articles/finance_xml.html ) to its representation. However, even without this ability to re-use known semantics, the SVG representation can be more accessible, since each element of the graph can be labelled with a title and description, which is then available to AT's as a user interacts with the graph. (e.g. Vladimir Bulatov's bump printer described in http://www.svgopen.org/2004/papers/SVGOpen2004MakingGraphicsAccessible ) Even without a title/description, information can be better obtained from an SVG document, than a bitmap (which is all canvas has, it just paints bits onto a canvas, there's no state). > Can SVG perform accessibly without leaning on CSS? They're both rendering languages and completely orthogonal, in SVG the semantics are in the rendering - and in CSS there are no semantics in the rendering, there's nothing to lean on each other. CSS is actively harmful to accessibility in SVG, you can't style a rendering language and maintain accessibility (If you're not aware of the reasons why, you cansee the problems the CSS cascade produces described in http://jibbering.com/2002/8/text-mixup.svg ) Making SVG graphics accessible is not about CSS - CSS isn't about making anything accessible, it just removes the problem of seperating the rendering from the semantics. Jim.
Received on Friday, 5 November 2004 16:09:56 UTC