- From: Al Gilman <asgilman@iamdigex.net>
- Date: Thu, 17 Jul 2003 17:52:42 -0400
- To: w3c-wai-gl@w3.org
At 06:34 AM 2003-07-17, Little, James wrote: >I don't see any need for META refresh or any such tags - if the page needs >to be refreshed it should be left for the user to do so (Accessibility >aside, it annoys me when pages are refreshed and change while I am in the >middle of reading the page). > >Just my 2p One needs to separate two issues: 1. redirection. Do it with the 'right tool' in HTTP; don't fake it with a refresh. 2. bona-fide refresh to serve as a subscription to an event stream. In either case, there are facilities in HTTP to do the job and so far as I know that's the better way to do it. META HTTP-equiv is a hack to work around problems in the author-server handoff. The problem is that the people who put the META tags in the pages don't control what the server does. http://lists.w3.org/Archives/Public/www-tag/2002May/0210.html Getting that fixed takes more than a pronouncement from us. The User Agent group took a good approach in stating the access requirements in terms of the experience offered to the user, and leaving the HTTP header vs. html:meta protocol-repair hack to the web architecture folks. Orthodoxy is good for accessibility, but we should not go off on our own to frame "adhere to standards" techniques. This is an area where we need to work with the QA group, the Technical Architecture Group (TAG) and the working groups behind the specifications in campaigning for adherence to the published specification and their best use. In terms of bona-fide refresh, this is a mixed blessing for accessibility, not a uniformly bad thing. Affording the user supervisory command of the event-handling policies is a cross-disability position. Disallowing or discouraging refresh is not. There are users with motor disabilities and users with cognitive disabilities where the values are all the other way. Here animating the web application by allowing events to be pushed from the server and alter the state of the client- side process are access enablers, not breakers. This is why the User Agent WG elected for configurability, not prohibition, in this area. Please review the W3C Recommendation in this area http://www.w3.org/TR/UAAG10/guidelines.html#tech-configure-content-retrieval Events that push from the system to the user interface are as close as your alarm clock. The INCITS-V2 committee working on a remote console standard is working on the premise that the system has to support pushing events from the system to the user and not be limited to a pure-pull protocol. Would you want to have to manually poll your doorbell? There are things going on like tornado alarms that can be known to the computer network and are worth interrupting a user for. http://www.incits.org/tc_home/v2.htm Please give the architecture sketch of the Multi-modal Interaction group a looking over. This lays out their idea of the general dialog fabric that will replace "pure pull of documents." Stock tickers we should be able to get recast as Timed Text media objects so it is clear to the assistive technology that the content change on the push event is limited to the latest bit of stock quote and not the whole page. http://www.w3.org/TR/mmi-framework/ http://www.w3.org/AudioVideo/TT/ Popular uptake of the Web has been powered no more by "access to all the World's information" than it has been powered by access to up-to-the-minute information. Event streams bearing information that simply wasn't available before *now* is a class of web content that is of value to people with disabilities as much as it is of value to the general public. Al >James
Received on Thursday, 17 July 2003 17:52:53 UTC