- From: Jim Allan <jimallan@tsb1.tsbvi.edu>
- Date: Wed, 13 Oct 2004 08:53:02 -0500
- To: W3C WAI-UAWG <w3c-wai-ua@w3.org>
Jon Gunderson Jim Allan - scribe Matt May David Poehlman Aaron Leventhal Cathy Laws Will JG: merging the Mozilla Keyboard Model (mozilla, cita) al: manages mozilla accessibility effort need to generate a list of things that need to be done dp: Sun to come out with new version of the keyboard requirements document. working with Bejing al: sent Bejing javascript that uses arrow keys to move by element. Add to mozilla components directory. "keyboard listener" works with every new page that opens. Would be useful for specific disabilities that need different keyboard access. could works based on user preference or user profile, or presence of screen reader. jg: hard for non-screen reader users to know and use different keyboard models. Screen readers use keyboard differently cl: screen reader will have additional keys over standard mozilla keys. al: dhtml calendar. on page javascript that changes the way arrow keys work. may conflict with screen readers. user must be able to access keyboard. should be at most 2 keyboard layers - caret browsing - move cursor, when hits something that is focusable moves the focus. Review mode. API for focus is based on caret position. ATK - by Sun. jg: compatible with Composer? yes. share keyboard code. al: how about structural and visual navigation. cl: "line" is relative to what is on the screen. screen readers define "line" differently. not visual. jg: what is difference between structural and visual navigation? want to avoid different function of arrow keys. al: screen readers, etc. use up and down arrow to move by element. visually move by onscreen line. dp: structure nav is something we need to keep. browsing is different from editing. describes how JAWS reads "lines" in browser. editing is more granular. cl: hpr allows user to control the line length. discussion of display models and how assistive technology interacts with these. al: things that everyone needs in standard model, things for screen readers goes into overlay/add-on to standard model. similar to shift-f10 for context menu, use f12 for accessibility options. will: need predefined user models, configure keyboard for them. work vs. home...higher verbosity at home. al: need per web page scripts to change keyboard interface for web applications. this scripting model allows browser to function like jaws scripting language. jg: need to keep in mind the average user. they are not going to write scripts. al: right. not the intent. have a repository of scripts. jg: with these scripts could create a hpr model of mozilla. cl: right. dp: personal configuration would be useful. may be confusing at the onset, screen reader users not used to using other browsers. dp: what does screen reader developer need to do to work with mozilla. al: mozilla on windows works with window eyes because they use msaa. JAWS does its own parsing in IE, then syncs with displayed page. Linux/gnome is different. uses cursor/caret navigation. not sure how to screen readers with work with mozilla on linux...may make mozilla self-voicing. dp: JAWS only works for IE (html parsing), will not work with other browsers. if there is a way for users to access information directly without having to way to screen reader developers to make an interface. jg: up/down arrow moves by line (standard model), screen reader model (keyboard extensions) would move by element unless overridden by user. al: keyboard api is javascript, and can include c++. there are DOM issues with screen readers and what is not available in the DOM (such as, specific item in ordered lists do not contain the actual number of the item in the list). differences between windows and linux accessibility. No plans to work on Tiger (Mac) accessibility. mm: Mac is using Safari as their browsing solution. and the Self-voicing of the operating systems and applications. Not Apples priority to work with other browsers. Telecon with Bejing next Friday. issue with document loosing keyboard focus. DOM does not tell you that the keyboard has left the document. there are other APIs from which to get that information. jg: UAAG could develop test suites for testing keyboard models. 1. Mozilla Keyboard Model (guest Peter Korn from SUN) http://www.mozilla.org/access/unix/keyboardproposal.html http://cita.rehab.uiuc.edu/mozilla/mozilla-keyboard.html Some issues: 1. Screen reader users verus sighted users keyboard requirements 2. User configuration of keyboard shortcuts 3. Test for user agent test suite to support the development of keyboard support 4. Developing a common document, defining editors, and what should be in the document Jim Allan, Webmaster & Sta
Received on Wednesday, 13 October 2004 14:18:10 UTC