- From: Steve Donie <sdonie@zycor.lgc.com>
- Date: Fri, 19 Nov 1999 16:08:22 -0600
- To: "'Mike Brace'" <mlbrace@earthlink.net>, "'W3C interest group'" <w3c-wai-ig@w3.org>
The basic reason is that it is difficult to impossible to write an algorithm that can deduce the meaning of these items from their visual appearance. Also, many times the screen reader doesn't have access to information about the structure of what is presented on the screen. For example, frames. As a sighted person, I cannot tell (by just looking at the screen) whether a web site uses frames or not. I can try to figure that out by, say, tabbing through (if my browser supports that - many don't). Or I could look at the source for the page. But screen readers don't have access to the source for the page - just the visual look of the page, and perhaps some GUI specific things like if there are multiple HWND's (in MS Windows). This is the kind of problem that Microsoft's Active Accessibility was supposed to solve - provide a way for screen readers to get information from an application. As someone who worked on that technology, I will be among the first to say that MSAA isn't a complete solution - there is a long way to go. But when you have two applications that support MSAA, it can go a long way to relieving some of these problems. Of course, when it comes to web pages, it also requires some assistance from the web page authors. MSAA in IE 5 will expose the alt text of an image to a screen reader that uses MSAA to extract that information, but the alt text must still be there, for example. Ditto for forms. When you look at a web form, you can usually tell which field labels go with which fields - but you do this by looking at lots of information - relative proximity, any repetitive features, etc. HTML4 allows an author to specify "this label labels this field" and I think that IE5 exposes that through MSAA, but again the web page author must put that information in. Steve Donie -----Original Message----- From: Mike Brace [mailto:mlbrace@earthlink.net] Sent: Friday, November 19, 1999 3:30 PM To: 'W3C interest group' Subject: Readers Can anyone offer a reasonable explanation as to why readers can't be engineered to accommodate some of the more fashionable trends, e.g., frames, table formats, interactive forms, etc. ?
Received on Friday, 19 November 1999 17:08:02 UTC