- From: Robinson, Norman B - Washington, DC <Norman.B.Robinson@usps.gov>
- Date: Tue, 8 Nov 2005 08:46:07 -0500
- To: "John M Slatin" <john_slatin@austin.utexas.edu>, <Becky_Gibson@notesdev.ibm.com>, <w3c-wai-gl@w3.org>
On Monday, November 07, 2005 5:37 PM Becky_Gibson@notesdev.ibm.com posted RE: Summary of arguments FOR validity -- and another against -- and a third of alternatives; "a level 1 validation requirement may prevent innovation of new technologies that use new attributes not included in current specifications (for example implementation of the DHTML roadmap [http://www.w3.org/WAI/PF/roadmap/DHTMLRoadmap110505.html] using HTML)." I can just as easily state that a level 1 validation requirement may prevent new technologies from being accessible. As was mentioned in one of your posts, the proper approach is to get "innovative and new" technologies supported by the existing specifications. In your DHTML example referenced, the specification only allowed tabindex for two elements. You're trying to work around the specification that exists. This is an example of why validation is required - not an example of why it hinders 'innovation'. If I build a screen reader, logically I'm going to follow the specification. I'm willing to bet I haven't tested my screen reader for the conditions of tabindex outside of what was specified. What happens? Maybe my screen reader can't handle the condition and ignores it, maybe it has unpredictable results. If I was testing I would be able to determine that 1. You weren't following the specification and 2. The screen reader isn't expected to deal with "innovative and new" technologies. I'd debate that new and innovative might be expected to not follow the standards, but I would also expect them not to validate. If I were a program manager in a large corporation, I would see that as risky. I would then be able to determine I was being offered something that was not standard, and then be asking the questions of who else has implemented this technology approach and if they were successful. I don't particularly care to be a test case for unproven technologies, especially where the support by existing (and legacy) assistive technologies hasn't been proven. It doesn't matter that your example is all about providing a new way for accessibility to happen, it still fails badly. This further enforces my belief that if you do not *require* validation then all sorts of vendors will be 'new and innovative' without being able to determine (or not care in some cases) if the technologies exists to support accessibility. Regards, Norman B. Robinson
Received on Tuesday, 8 November 2005 13:46:19 UTC