DIVERSASIA - Best practices

While lists of tools and services are useful, their applicability and suitability depend on context. To that end we have a few observations. Whether addressing website accessibility is internally (re-design) or externally (regulatory pressure) motivated, a fundamental question can be: which sites need testing? Larger universities with sprawling web estates of internally and externally hosted sites may not know all of the sites within their scope. Regulators and visitors aren’t concerned with ownership technicalities. It all looks like one website. You may need to start by identifying all the sites in a web estate. Automated testing is the essence of computer programming: exhaustively cycling through algorithms. In this case, comparing web page elements against guidelines (usually, WCAG2.0). But, parts of WCAG2.0 aren’t amenable to automation. They require human intervention. For example, automated tests find missing ALT attributes, but don’t verify the descriptions being used are appropriate. Automated testing can do some, but not all of the work. When the need for manual intervention is combined with variations in test logic implementation, it’s unlikely that only one tool or service will work in under all conditions. In other words, one tool can do some, but not all of the work. We’ve only included tools and services that analyse page structure, as this process represents the greatest opportunity for testing-at-scale and is the biggest challenge across complex pages and large websites. We’ve included tools and cloud-based services that test individual pages as well as entire websites. Browser add-ins let users test single pages. Cloud services either inherently test entire domains or have APIs to process URL lists. Testing whole websites produces data mountains. Individual web page tests can generate dozens or even hundreds of accessibility results. Entire websites can produce hundreds of thousands of data points to evaluate and prioritise. We’ve limited the list to tools/services that appear to be under continuous development or have been brought to market within the last five years. Tools testing to current standards with regular update releases are of more value to time-pressed higher education web teams than dated, but free, open source projects.