The Awful State of Web Accessibility
This past week, my company has been promoting web accessibility through a series of seminars that explain its importance for business and some techniques that we can use to make our web sites more accessible. On Friday, I attended a seminar hosted by the Cincinnati Association for the Blind and Visually Impaired (CABVI), in which they demonstrated how web sites appear to users with poor or no vision.
For those with poor vision…
For users with poor vision, they have applications that can zoom in so close that the icon for one application basically takes up the entire screen. These users work more slowly, since they have to pan around the screen a lot, but they can still use most web sites. Mainly, you just want to provide ALT tags for your pictures, since it’s hard to tell what something is at that close range.
For those who are blind…
For users who are blind, they have applications such as JAWS or Window Eyes, which read the screen for you. This is a totally crappy experience on the web. Really, I can’t exaggerate how bad it was. They guy who demonstrated these tools is an instructor at CABVI, and he seemed about as savvy as you can get with the software — and it was still awful. The software provides a lot of keyboard shortcuts, such as a key to move down to the next heading in the text, or a key to open a list of all the hyperlinks on the page. These keys helped a little bit, but you still have no sense of context for what you’re hearing. The screen reader tells you the name of a link, but you don’t know if you’re in the main nav, somewhere in the body (left column, right column, a call out), or in the footer.
Filling out a form is next to impossible. Screen readers operate by reading the HTML behind the page, not the page you would normally see. So when the user tabs to a field, it often just says “edit field” without providing the name of the field. Who knows if it wants your name or your phone number or what. Unless the name of the field is directly before the field itself in the HTML, the screen reader doesn’t give you any context. (And although it seems like it would be easy to write HTML with the name of a field next to the field itself, it’s not. A lot of forms are built with ASP.NET or JSP, and it seems like those languages sometimes separate HTML and forms automatically.)
What can be done about it? - Part 1
In the meeting, my friend Ian asked if the people at CABVI have ever heard of Quicksilver. For those who don’t know, Quicksilver is an amazing program for the Mac. It’s like Spotlight, except much much cooler. You can find any file (let’s say a song) and do almost anything with it (e.g. open it with any application, play it in iTunes, show it in Finder, etc.) all with keystrokes. I think this would be especially cool for the blind because it’s very intelligent. For example, you could type “ffox” to find Firefox. The first time you do that, Firefox might not be the first result, but then next time you type “ffox”, Quicksilver is smart enough to know that “ffox” means “Firefox” to you. The developer, Nicholas Jitkoff, explains the theory behind his application in a Google video (warning: Nicholas is brilliant and worth listening to, but he’s not the most engaging speaker). In fact, his philosophy was a partial inspiration for my Google Social application.
So Quicksilver might be a good solution for the visually impaired on the Mac desktop (has anyone tried this before? can anyone confirm it?), but it doesn’t help too much for browsing web sites.
What can be done about it? - Part 2
My initial impression is that the people who develop screen readers are going to be constantly at odds with the legions of people who make web sites. Web developers are always finding new ways to do things, and web designers are interested in whatever technology helps websites look better (even though those advances generally seem to hurt accessibility), but almost nobody is thinking about accessibility. So these people developing screen readers are working as hard as they can to keep up with technologies that weren’t build with them in mind.
That reminded me of an important idea (that nobody talks about anymore) called the Semantic Web. According to the guy who invented the current world wide web, this is the future of the internet. Basically, it’s a new form of markup that allows machines to really understand what’s on the web.
Imagine, if we has semantically accessible web sites, your computer would know if a link is part of the main nav or part of the body. It would know the name of every form field. The computer would know what’s inside a picture. In the short term, this means your screen reader could actually be useful. In the long run, it means you could give your computer a command like “order me a medium pepperoni pizza from Pizza Hut” and it would do everything for you (like the Star Trek computer!)
Bottom Line
Those of us who are shaping the direction of the web need to support accessibility. My feeling is that making the web more accessible for the disabled will ultimately make it more accessible for everyone.
October 1st, 2007 at 4:16 pm
Your article reminded me of “Bobby.” It was an automated script search thing that would look at your code and determine if it was screen reader-friendly. Bobby has now become “Watchfire” and you can fin dit here: http://webxact.watchfire.com/
oh - and heres the background info on Bobby: http://en.wikipedia.org/wiki/Bobby_%28software%29
Side Note: Do we really want computers knowing what exactly it is that they are accessing? I know I, for one, do not want my computer knowing that I accesss LOLCATs all the time. Not that I do that, I’m just saying that this is one step closer to robot domination! I’m aghast that you would support our future slavery like this.