- From: Ramón Corominas <listas@ramoncorominas.com>
- Date: Wed, 17 Feb 2010 01:28:57 +0100
- To: Phill Jenkins <pjenkins@us.ibm.com>
- CC: WAI Interest Group <w3c-wai-ig@w3.org>
Hello, Phil and all. I totally disagree with any solution that transfers the responsbility for web accessibility to the end-user side. While I agree with the part where User Agents should provide methods and solutions for some of the problems, we should not rely on the user's ability to configure settings, build style sheets or any other more or less complicated task, even with basic or advanced training. Of course there are some solutions that are better than others, but if we don't provide solutions thinking that the user will be able to configure the browser or install a plugin or anyhing else, we for sure will leave many users apart. We simply don't know if our assumptions about users' knowledge will be true. For example, a friend of mine that has also retinitis pigmentosa doesn't know how to configure most of the browser settings that he usually would need to use the Web. I explained him many times, but he explained me that he is not a technician and so he always forgets these kind of things. If he buys a new computer, he will need help again to configure the system. And don't forget that to make changes to a particular setting the user needs first to find his/her way to that setting (for example, when I bought a Mac, that has lots of built-in accessibility features, I needed about an hour to find my way to the "invert colours" setting, because my low vision made it very difficult to locate the right icons, texts, and so on). We can, of course, make some reasonable assumptions about some types of users (for example, a blind user will probably use some kind of screen reader software), in the same way that we assume that a user that browses the Web will probably have some kind of Internet connection and Web browser (wink). We can then study these situations and try to cover the default situation where most users will be, because we learned from many user tests that many users don't know a word about "hidden" settings or browser configuration. It's a pity, but it's the Real World (TM). Anyway, for this particular "new window" case, my recommendation is: 1. Do not open new windows. It is still considered bad practice and has no -demonstrated- advantage. 2. If you MUST do it, inform the user. Any user. - For screen reader users, an icon with alt or a hidden text would be enough. - For mouse-visual users, a title with the text of the link and the warning would be enough. - For keyboard users, CSS :focus class can be used to present the information in a way similar to the title. 3. If you can make the assumption that JavaScript must be enabled to allow new windows, add the behaviour using the onLoad event and non-obtrusive JS, and at the same time add the clues for the different types of users. Best regards, Ramón. Phill Jenkins wrote: > Sounds to me like a classic problem with the user agent and/or end > user settings.
Received on Wednesday, 17 February 2010 00:29:29 UTC