Moving onto another section in the WAS Body of Knowledge, quickly approaching the end. I’m postponing going over the “Test for End-user Impact” section in order to work through the “accessibility testing tools” section. The summary says it all for me:
“No accessibility software tool can find all the accessibility issues on a web site, but software tools can expedite the process of finding accessibility issues, and increase the overall accuracy when supplemented by a skilled manual evaluation of the same content.”
Or, as the Web Accessibility Initiative (WAI) sums it up:
“We cannot check all accessibility aspects automatically. Human judgement is required. Sometimes evaluation tools can produce false or misleading results. Web accessibility evaluation tools can not determine accessibility, they can only assist in doing so.”
Things I accomplished
- Read W3C WAI’s Selecting Evaluation Tools section.
- Browsed WAI’s Web Accessibility Evaluation Tool List.
- Followed some links to different tools, in order to get a better sense of what’s available.
What I learned today
I hadn’t considered this before, but not all tools are meant to target one audience (developers). Each tool is created with a specific audience in mind, whether it be:
- non-technical content authors,
- quality assurance testers, and
There are SO many options. How intimidating for anyone trying to decide what software, plug-in, or consultant to use!
Automated testing involves different considerations based on audience, need, conformance standard and level, site complexity, and accessibility experience. Various types of automated testing include:
- site-wide scanning and reporting (SortSite, Tenon.io, AMP)
- server-based page analysis from one page to entire site (Cynthia Says, SiteImprove)
- browser-based developer/QA plug-ins that evaluate one page at a time (WAVE, AInspector)
- unit testing during development (aXe API)
- integration testing before deployment (aXe API)
It strikes me that using a combination of tools with differing purposes could help speed up the process and ensure accuracy even more. By no means, would they replace manual checks and end-user testing, but it’s incentive to not pick just one tool to do a job meant for several tools.