Thursday, December 20, 2007

Web Accessibility And Usability: Coming Back To The Basics

The Nielsen Norman Group has made available a detailed report on accessibility, which includes the results of several usability tests --- see Going Beyond Alt Text. It's a very good read, and though its conclusions might be depressing to people coming to this area from the outside --- they should be no surprize to users who have been trying to use the Web via spoken output over the last 10 years. From the perspective of the Emacspeak user who lives in a specialized browsing environment that is optimized for performing oft-repeated tasks, there are several interesting take-aways from this report:

  • Though so-called Web Accessibility Standards have attempted to focus on the behavior observed when using screenreaders with mainstream browsers, that thread of work appears to be achieving little with respect to the real metric of task completion. As technologists, we would all do well to remember that users come to the Web, and Web Access solutions not to use the browser but rather to complete one or other task.
  • As described in Specialized Browsers and The Web The Way You Want It, task-oriented access and specialized user-optimized web tools have been around since the inception of the Web.
  • Though the Nielsen study asked users to carry out each of the given tasks by going to a given Web site, it would be interesting to see how such tasks work out in the Emacspeak environment. It's a given that an emacspeak user trying to buy a music CD at an online store would run into a brick wall fairly quickly (see, even online stores are made of brick and mortar;-)). However, as an Emacspeak user I'd never do that I'd either go to Amazon's highly efficient Amazon Accessible Store or faster yet, type an appropriate query at Google and click on the relevant Ad that sports a Google Checkout badge.
  • And more interestingly, it would be interesting to carry out a follow-up user study to compare the rate of task completion as well as observed efficiencies/inefficiencies between users of Emacspeak and generic browser/screenreader combinations for tasks such as:
    1. Play NPR news from the last hour.
    2. Play your local NPR station.
    3. Play BBC News from the last hour.
    4. Skim the top stories from CNN.
    5. Look up today's stock market numbers for the major indices.
    6. And items too numerous to enumerate in this margin.