Monday, May 09, 2005

On Accessibility of Online Content

We seem to be on the verge of a minor breakthrough with the WebCT Vista quiz engine in enabling foreign language based quizzes for students with visual impairment. I'm not expert in the underlying HTML issues, but as I understand it for those using a screen reader there are problems at junctures where the language changes, say from English to Spanish or back from Spanish to English. If the screen reader doesn't know it is supposed to change languages, you can imagine the result. There are HTML tags to handle this issue but apparently in the response to the student's quiz answer the Vista software somehow disabled those tags. Now apparently, we have gotten to the route of this problem after a fair amount of investigative work on the part of experts in Web accessibility from my campus and other campuses that are using Vista and interested in accessibility as well as from WebCT itself. Interestingly, the modifications in the code apparently affect this function only with no impact whatsoever on other users. So here is a feel good story about how serious cooperation can produce a tangible result for user benefit. Let's look forward to more of that in the future.

However, it is equally apparent to me that there is a long uphill climb on the accessibility front and much of that has to do with awareness of the designers. The problem is not an issue of good will. It is an issue of being conscious of the accessibility concerns at a very drill down level when writing the code. Here is another example to illustrate, one where the Vista software still is coming up short.

Apparently Vista relies heavily on frames, including an invisible frame which contains much of the fancy programming. The invisible frame includes content that is regularly updated as a consequence of what the user has done on the page. The screen reader goes from frame to frame to check out what is in it. When the reader cycles to the invisible frame it sees new content there and begins its cycle around the page anew. The reader should actually ignore the invisible frame but there is no command in the code at present (I don't know whether this problem is rectifiable or not). This is why some accessibility folks have argued for frameless pages. (The CMS Angel offers a frameless alternative.) But the upkeep of a frameless alternative is expensive. If the version that everyone else uses were accessible, that would seem to be the preferable solution.

Did WebCT set out to design their product to fool screen readers? Of course not. Do they test their product with a screen reader? I doubt they did that at the outset. They do have a particular staff person devoted to these issues now and they might do that in the future. But these accessibility concerns were not part of the original vision of the product that was articulated a few years ago. I should point out that WebCT has had their product reviewed for section 508 compliance by an outside company and they passed that test. The point is that the legal requirement doesn't necessarily map very well into usability for a person with a visual impairment. The latter requirements are sterner and I believe not well understood.

What I'm understanding now, after talking with many people both who work for me and who specialize in accessibility is that these issues have to be addressed in context. Applying general design principles is fine, but that is not sufficient. The principles can't possibly cover all eventualities.

I think we will continue to make incremental progress in this way and I hope we have some goodwill in the group of folks who have been working on these issues. They should be proud of what they have accomplished. But I think we need to be quite humble about what we are offering to people with disabilities and as we are constantly pushed to expand the functionalities of tools such as WebCT Vista we need to understand that it may be at the price that use is less than universal.

No comments: