- From: Charles Pritchard <chuck@jumis.com>
- Date: Fri, 28 Jan 2011 11:05:23 -0800
- To: public-canvas-api@w3.org
- CC: Richard Schwerdtfeger <schwer@us.ibm.com>, franko@microsoft.com, david.bolter@gmail.com, public-html-a11y@w3.org, swarren@aisquared.com, tlalor@aisquared.com, dmazzoni@google.com
http://dl.dropbox.com/u/17337752/text-legibility-crop.png Attached is a screenshot from the text-legibility script; demonstrating issues described in the Canvas Accessibility agenda that Richard has been maintaining. The first line shows two normally rendered span elements, with tabindex associations on each, displaying two different fonts and sizes. It treats the text spans as one might treat role='option'. The browser is set to a high magnification level. The third line shows the same text, with textBaseline and fillText being used (canvas). The second line shows another canvas element, with accounting for baseline offsets and window scaling; the paths to be used as focus indicators are outlined in brown. For all practical purposes, they match the focus outlines that are actually drawn on the first line (the spans). At the bottom right of the image is an html select element, with focus on one of the option elements. In this instance, the select element has a focus ring around it, and my selection position encompasses the select element. The script has not been tuned for various browsers; there are work-arounds for scaling, baseline offsets, and focus rings: canvas_a11y proposals are intended to standardize some of these methods. These proposals are relevant to the demo: window.devicePixelRatio - a floating point returning the current ratio of CSS pixels to units. drawFocusRing - to indicate a region corresponding to a focused element setCaretSelectionRect - to indicate a sub-region within a focused element. TextMetrics baseline - to indicate the baseline positioning of text Currently, these techniques rely on various methods of CSS introspection. They would be used more often were they easy to access. Many CSS techniques for a11y involve transparent or non-visible layers, absolution positioned around the viewport. The proposals listed above are intended to obsolete such techniques. Script (work in progress): http://dl.dropbox.com/u/17337752/text-legibility-script.html This document (also WIP): http://dl.dropbox.com/u/17337752/text-legibility-readme.txt I've not finished working on the script so that it only uses our proposals; using various backwards compatibility techniques. -Charles
Received on Friday, 28 January 2011 19:06:05 UTC