Test Concepts for Accessibility
Entire programs now are running in the browser. Word processing, spreadsheets, photo editing, games - the more complex a website is, the more testing becomes necessary. This applies even more to accessibility. Find out here how you can check websites for accessibility.
There are various test options. The usability tests involve people from the target group who are observed using the application in practice. There are automatic testing tools designed to check accessibility. This is quite useful for detecting errors on the web page such as poor contrast, missing alt text, and so on. This is sufficient for the first phase, especially for testing prototypes. You can't uncover real problems with it. I could set up my content management system to automatically allocate some filler text for graphic descriptions. The test tool cannot differentiate between real and irrelevant alternative texts. It also cannot determine whether a markup is semantically correct.
Many people still confuse accessibility with validity. The W3C provides two websites that can be used to check HTML and CSS for correctness. That's important, but valid web pages aren't necessarily accessible, and invalid code doesn't necessarily have a negative impact on accessibility. It's enough if a single character in the head of the file was coded incorrectly and the validator throws out tons of errors for every page with this head, without the screen reader user having to care.
With a little experience, systematic testing can be done. There are extensions for the widely used browsers that can be used to analyze common problems.
Some tools show the webpage as it would be seen by screen reader users. While this is useful, it is mostly temporary. There is a difference between seeing a page as a screen reader and actually having to use it. In practical use, completely different problems can arise than what the web developer can anticipate.
Nevertheless, the test by an expert is the best solution in terms of price-performance ratio. By the way, an expert is someone who is familiar with web design, development and accessibility.
A mixture of different test methods seems to make the most sense. If you were to create a matrix of all possible assistive technologies, all possible versions and all possible browsers, you would get an infinite table. If you then take into account all possible visual impairments and add to that all DIFFERENT TECHNICAL abilities, the testing would never end. And would probably still overlook some problems.
In the usability tests mentioned above, an attempt is made to observe a cross-section of users using the website in practice. Users are tasked with completing specific tasks, observing a record of eye and mouse movements wherever they look. They speak out loud about what they see, what they are up to and what they are thinking about in the context of the website. Such tests can also be used with people implement with disabilities, in this case we also speak of practical tests.
As I said, it is difficult to depict all forms of disability. It should initially be enough if you bring in one or two people from each group and otherwise rely on quality assurance.
You have to say goodbye to the idea of being able to put a completely error-free application online. Of course, every effort must be made to track down possible errors. At the same time, it makes sense to spend resources on subsequent quality assurance. That is, problems are fixed when they occur in live operations and methods are used to systematically capture these problems.