The project was created by French independent developer Théo Ducreux Select HTML The address content of the 5,000 most popular domains was studied, their code effectiveness analyzed, and almost nine out of ten errors were found.
Photo credit: Mohammad Rahmani / unsplash.com
Among the samples, only 2,656 were targeted at humans, while the rest were used in technical fields, such as Google’s “gstatic”. The vast majority of them, regardless of whether they have a human-readable home page, display incorrect HTML code that violates World Wide Web Consortium (W3C) and Web Hypertext Applications Technical Working Group (WHATWG) standards. These standards are designed to ensure that browsers and screen readers display requested pages in approximately the same way.
A total of 100,305 HTML code violations were found across all sample websites, including the latest HTML5 specification, and 18,863 CSS errors. 87.2% of the sites violated web specs in at least one place, only 12.8% had correct HTML code, and only 2.6% were completely “clean” without throwing any errors or displaying any warnings. More than a third of websites fail accessibility tests, meaning they are not processed correctly by screen readers used by people with visual impairments. 20.4% of websites lack alt text for images (which often contains important information). 41.6% of pages lack ARIA tags in their code that identify page areas.
Experts point to two culprits for the current woes: developers and browsers. Especially when using front-end frameworks, invalid code can occur: the most common error is incorrect nesting of tags. The simplicity of modern browsers prevents developers from paying attention to this: the browser simply ignores any incomprehensible code, and the error is corrected, restoring the desired image. Unfortunately, screen readers don’t have this kind of durability.
If you find an error, select it with your mouse and press CTRL+ENTER.
