- From: Robinson, Norman B - Washington, DC <Norman.B.Robinson@usps.gov>
- Date: Wed, 15 Jun 2005 10:03:50 -0400
- To: "Sailesh Panchang" <spanchang02@yahoo.com>, <w3c-wai-gl@w3.org>
- Message-ID: <98D2018DE18A2D40BC2A04CFB111CB04020D197D@WADCHQSXM16.usa.dce.usps.gov>
Just ensure that you aren't making accessibility tests secondary to valid code: I would state generically that even pages that fail validation SHOULD be evaluated; you should not require valid HTML documents to test for accessibility. It is not necessary! Section 508 is a legal requirement where valid HTML isn't always necessary because of the way user agents (web browser) work through invalid HTML. There is a role for document type and compliance level (strict) to play in this debate. That said, section 508 does not explicitly require valid HTML DOCUMENTS, but I would argue that to meet 1194.22(A through I, L, M, N, & O) are all performed using VALID HTML CODE. While it may be that not all HTML is valid in a given web page, you can be certain that if you try and any of the above accessibility standards are not met, it is because you tried to do so using invalid code (improper nesting, improper tags, etc.) or your software is defective. Your example of a properly marked up table that the screen reader could not read as expected, you illustrate how SOFTWARE ( screen readers) are tools and subject to defects. It could have been the browser, the screen reader, or any other software interfering. Validation checks eliminate the problems of manually checking your code when you are trying to determine what problem occur in accessing information using assistive technology such as a screen reader. Where does it state you have to make the entire document valid to provide accessibility? You can choose to manually inspect all of your HTML that matches the above references if you prefer. That wouldn't be working smart. I'll be using automated tools - a validator - that allow me to identify defects faster and then focus on the issues of 508 compliance based on meaning and context. Note that you CAN validate all of the above references to meet the section 508 accessibility standard using the W3C validator. As an example, the tags for alt text are pretty well defined. When I validate my web page with a bad alt tag, it finds the defect and makes it easy to know what to fix. You can validate parts of a page while not caring if the rest of the web content may not be valid HTML. Section 508 doesn't explicitly require you to provide valid HTML, but if your code is not correct and a user agent and web browser cannot access the information, then you are not 508 compliant. Hope I added to the discussion and didn't waste your time - if I'm a bit off-topic email me directly and I'll be happy to respond. Regards, Norman B. Robinson -----Original Message----- From: w3c-wai-gl-request@w3.org [mailto:w3c-wai-gl-request@w3.org] On Behalf Of Sailesh Panchang Sent: Tuesday, June 14, 2005 9:44 AM To: w3c-wai-gl@w3.org Subject: Re-post: Influence of valid code on screen readers This raises an important question for tools that evaluate accessibility. Should only pages which do not fail validation check be evaluated for accessibility? But then few pages will qualify for automated accessibility evaluation because Bruce notes: "... is rare to encounter valid html, and such is not an explicit requirement. <for Sec 508>" I remember a recent instance wherein a complex table was apparently marked up "correctly" with headers and id and yet the screen reader / self voicing browser was not able to read it as expected. As it turned out, the id values were not unique in the document and the page failed the W3C Validator. Therefore I feel it is important that accessibility evaluation should be performed only on documents that are correctly marked up and pass validation. Thanks, Sailesh Panchang Senior Accessibility Engineer Deque Systems Reston VA www.deque.com E-mail spanchang02@yahoo.com ________________________________ Discover Yahoo! Find restaurants, movies, travel & more fun for the weekend. Check it out! <http://us.rd.yahoo.com/evt=32658/*http://discover.yahoo.com/weekend.htm l>
Received on Wednesday, 15 June 2005 14:04:04 UTC