Not long ago I participated in a discussion on a W3C mailing list where a participant on the list contended that a site is not accessible because it did not work right in Lynx. Lynx, for those who don’t know, is a text-based web browser – in other words, it offers no support for graphics, video, or JavaScript. There was a time, years ago, when a site working in Lynx was the litmus test for accessibility. If it worked in Lynx, the argument went, it would work in assistive technologies like screen readers. That was before the proliferation of Accessibility APIs, before screen readers used Document Object Model rendering, and before ARIA. These days, people who still view “working in Lynx” as a viable measure of a site’s accessibility are demonstrating their ignorance of assistive technologies, ignorance of the realities of modern web sites and applications and, frankly, ignorance of accessibility. Further, it is indicative of the myopic view that permeates the topic of accessibility that assumes the only people to be concerned about in discussions of web accessibility are blind people.

If it works in Lynx it works for everyone

The primary argument people make for advocating that a site work with Lynx is the claim that if it works in Lynx, it will work for everyone.

  • Lynx doesn’t display images. Images on a web page will be replaced with their alt attributes. If the alt attributes are good, the lack of images won’t matter. If the alt attribute values are missing or are not a suitable alternative for the image, this will be readily obvious in Lynx, indicating it also be a poor experience for users of screen readers.
  • Lynx doesn’t play time-based media. The on-page content, it is argued, should be sufficient for access by all.
  • Lynx doesn’t support JavaScript. Thus, scripted interfaces – some of which rely on device-dependent events – will be rendered inoperable. Again, the argument goes that if the site works in Lynx it will work for everyone, including keyboard-only users.
  • Lynx doesn’t support CSS. Therefore, CSS positioning, use of color, etc. will be impossible with Lynx. Like before, the argument is that if it is easy to understand the content in Lynx, it is possible for assistive technology users, some of whom may be unable to see the screen or perceive the use of color, size, and positioning on screen.

The Lynx experience is not representative of real users

Taking a look at the complete list of Lynx user-agent strings and looking back at the analytics on all of my sites, I see zero visitors from Lynx, ever. Though some may argue that the user-agent string could be forged (this can be done in the Options menu in Lynx), I doubt that fully all Lynx users do so. This tends to suggest, if nothing else, that the global percentage of users on Lynx is extremely small. Further to my point, this also suggests that the overwhelming majority of persons with disabilities are not browsing with user-agents that are text only, which is in agreement with my personal observations as well. This begs the question: if persons with disabilities aren’t using text-only browsers, what’s the point of recommending that we develop sites that support user-agent that is unused by persons with accessibility needs? The argument that Lynx is the “Lowest Common Denominator” is flawed based on the fact that Lynx is very clearly not common at all among persons with disabilities.

Text-only is not accessible

The arguments I list above, regarding supplying good alt text, not relying on JavaScript, etc. are good arguments, but they miss the point. That point is that text-only is not accessible. There are two reasons for this. First, as I’ve already implied: users with disabilities are not using text-only browsers, they’re using the same browsers as those without disabilities. Second, the arguments for supporting text-only betrays a myopic focus on users who are blind

According to there are roughly 1.5x as many users with hearing disabilities as those with visual disabilities. There are more than 2x as many persons with cognitive disabilities than those with visual disabilities. Finally, there are more than 3x as many persons with ambulatory disabilities than visual disabilities. In other words, it is important for us to consider all users, many of whom benefit from the use of images, color, CSS, time-based media, and even JavaScript.

Accessible is Accessible

For us to build a more accessible web, we should be careful to understand that Universal Design involves ensuring an environment is usable for all users not just those who are blind or who use a specific user agent or device. Lynx can be a good tool for providing quick (and free!) accessibility feedback on your site but it should not be regarded as the ultimate criteria for determining whether your website is accessible. Most tools, be they toolbars, alternate browsers, automated tools, or checklists are useful in testing for accessibility. The ultimate deciding factor however, is whether real people can use the system.

My company, AFixt exists to do one thing: Fix accessibility issues in websites, apps, and software. If you need help, get in touch with me now!