- From: Andy Heath <w3@axelrod.plus.com>
- Date: Sat, 06 Aug 2011 15:31:30 +0100
- To: w3c-wai-ig@w3.org
Karl, Your comments about the usefulness of automated testing tools seem to be valid and correct to me. I do have a problem with the comment below though .. > The second thing I'd like to comment on is the notion that compliance > with guidelines does not mean a site is accessible. To be frank, I'd > like to see data that backs this up. In my opinion, if you've tested a Isn't this a little like arguing that say one drove a car from A to B and it didn't break down therefore it never will ? Testing is only ever going to reveal properties around what one is testing for and not whether there are properties outside the scope of the tests that fail. The problem as I see it is with limited expectations. "I've met the guidelines therefore I've "done" all I need to do". I am reminded of the anecdote told on one BBC Ouch radio programme that Liz Carr told about not being able to reach the counter in a post office and being informed "Well we meet the Disability Guidelines" - the implication being "you *should* be able to reach". To be accessible to everyone we do need to meet the requirements of every individual at the time needed - and that may be the meaning of conformance to some standards - but unless we *know* the individual concerned and their requirements at the time then that's a little impractical (sometimes we can/do know the individual). So producer-focussed guidelines "average" requirements. In some cases that might mean in practice "accessible to many but optimal to nobody and not accessible at all to some". Is that a case to say its accessible and one need not do more ? It seems to me this is a problem that extends always to the place where design meets use. For example a council might design a cycling facility they think I would want but which is a million miles from what I actually need - of course they didn't ask me because that would have been impractical. Roads are designed for what - to meet the needs of *every* user ? I think not. Show me the freeway designed for me to cycle on. In the world of physical buildings and roads it must be this way because the costs (money, time, material, labour etc.) of reconfiguring entities is very high. But it doesn't have to be that way with computer systems. That is, it doesn't have to be that way if we can shift some of the organisational constraints that are there only because we don't like change and we are still operating as if its the physical world. It doesn't have to be this way - it can be better. To my mind human beings are much richer than machines can ever model and to fall back on the idea that the model we use represents the reality is a cop-out and in many cases an excuse for some other agenda which may have more to do with business goals than any claimed truth. I have no problem with that, its the real world here and there *are* real constraints but lets be honest here and stop pretending we've done everything that could be done. We haven't. andy
Received on Saturday, 6 August 2011 14:31:59 UTC