Brief Encounter: Dr. Graham McAllister talks about why companies should do more to make their sites accessible to the visually impaired.

Graham McAllister of Queen's University, Belfast, lectures at undergraduate and postgraduate levels in computer programming and how it is applied to the areas of digital audio software and computer games creation. His research focuses on the area of HCI, in particular, accessibility and usability. His current research project, which received funding from IT services group Eduserv, explores the area of representing visual information using multi-modal feedback. In 2006, Graham formed Ninespace, a company specializing in Accessibility and Usability consultancy.

NMK: What is the problem when it comes to visually impaired people and the Internet?

Graham: What I have found is that many websites are built in an extremely visual way, perhaps especially those that are consumer-facing. It's sad, but many e-commerce sites are especially bad. Take the [well-known airline ] site. It's full of forms and tables that the screen readers used by visually-impaired people don't tackle well. If there's a timetable, the reader will just start listing a series of figures, which haven't necessarily been structured in a logical way. Then there are forms for the e-commerce side which will time out before people with disabilities have time to fill them in.

Visually-impaired people actually use the Internet in a different way than sighted people. They tend to be task-focused. They've decided to buy a digital radio, for example, and want to get on and buy it. Because of the way that they're accessing these pages, it tends to take them longer to do things, and so they need a very direct path. Unfortunately, though, consumer sites are often very marketing led and put barriers in the way of achieving these objectives.

NMK: What would be good examples of sites that work well for the visually impaired?

Graham: Well, the BBC News site is a good example of the way things should be done. There's also a good directory of resources for the visually impaired at White Stick. However, a lot of designers would look at that and say it was boring. That doesn't have to be the case. The Zen Garden website, for example, is a good example of the way in which strong design can live alongside accessibility. All the design elements are applied through cascading style sheets, so the actual substance of the site remains sematically intelligible and could be displayed using a user's own style sheet to increase its legibility for them.

NMK: For what other reasons should companies and designers be thinking about the visually-impaired?

Graham: Well, as I mentioned earlier, visually impaired people aren't just on your site to browse. By and large, they're there because they want to buy something. Internet shopping should be a great convenience for people with all sorts of disabilities. There are two million registered visually impaired people in the UK, so that's three percent of your audience purely as a proportion of the population. But because they want to buy things on the Internet since it's often more convenient for them, it could be an even higher proportion of the total e-commerce spend that an inaccessible site is missing out on.

Second, companies should be well-aware that it's illegal to discriminate against the visually impaired. The DDA (Disability Discrimination Act) means it's against the law to offer a service to able-bodied people and not offer that same service to other groups.

Lastly, on a purely commercial plane, there's a certain amount of evidence to suggest that a site that's built for accessibility will work better, i.e. produce more sales, than an inaccessible site for all sectors of the population. That's certainly true of Google rankings as well - an accessible site is semantically intelligible to Google's spiders and that will improve its rankings.

NMK: What advice do you have on what companies should be doing to improve accessibility?

Graham: The main thing is that your designers are familiar with and working towards the W3C guidelines on accessibility. There are three different levels of accessibility you can comply to, with 'triple-A' standard being the best. Note that this isn't the same thing as XHTML compliance, which looks at whether your code is well-formed. A site might well be completely XHTML compliant but utterly inaccessible.

It's a very good idea to get a copy of the text-only web browser Lynx. Looking at your site using this tool will give you a very good idea of what your site 'looks like' to the visually impaired. It shows you what the site means rather than how it looks.

There are a number of services that will do an automated check on compliance levels, such as WebXact. These are worth a go, but they won't do as thorough job as a real person testing the site. Take the issue of pictures, for example. Accessibility guidelines tell you to use the 'alt' attribute to give each image a caption. However, those captions might just say 'picture', which isn't an enormous amount of use. It's a far better idea to get visually impaired people to have a go at using the site. This sort of process can give designers an enormous amount of information, not just about accessibility, but the overall usability and design - in the sense of how it works - of the site.

One last point. Courses and accessibility audits may seem expensive but it's a lot cheaper to get a site correct the first time round than it is to try to bolt-on accessibility as a second thought.

NMK : So how is your current research helping?

Graham: We've developed a prototype browser that gives tactile feedback about the layout of web pages. It will give a nudge, for example, when the cursor passes over a picture or a link. Using this technique, visually impaired people can get a very detailed idea of the layout of a page, and this really helps with understanding it.

Five years ago, it might have been alright to simply provide a text-only version of the page as an alternative for the visually impaired. But today, their expectations are quite rightly rising. They want to know the pleasure of using the Internet that has been denied them. We think that the sort of ideas we're playing with in this research can help deliver that.