Where It's AT – Assistive Technology Blog

HTML5 and longdesc

The longdesc attribute, which is used to reference a page that contains a longer description of a complex image (such as a chart or graph), is currently not present in the HTML5 specification. It was dropped due to poor implementation by authors (it was often coded incorrectly or didn’t provide a true alternative to the complex image), poor browser support, and because it really only provided utility for some screen reader users. Unfortunately, there is not yet a suitable alternative attribute or functionality in HTML5 or ARIA for referencing a long description page. While longdesc will likely have continued support in browsers and screen readers for some time, it has long been recommended that for optimal usability and accessibility that authors provide a standard link to the long description page instead of or in addition to using longdesc.

HTML5 and Optional Alt Attribute

HTML5 currently allows the alt attribute of the image element to be optional. If an image is given alt=””, it indicates that the image is decorative and does not convey content, or that the content of the image is conveyed elsewhere, such as through an image caption or adjacent text. In HTML5, omitting the alt attribute indicates that an alternative for the image was not provided or cannot be determined. An instance where no alt attribute might be necessary is when a user uploads hundreds of photos to a photo sharing site and will not provide alternative text for each of them. Because the alt attribute is required in previous versions of HTML and XHTML, this would require the photo sharing site to give the images improper alternative text (either alt=””, which is not accurate because the image does convey content, or generic, inaccurate alternative text such as alt=”photo087″ or similar) in order to be valid HTML. In HTML5, the alt attribute may be omitted in this case, and while this does not make the image accessible, this does provide an indication that the image is not accessible and might allow screen readers to then attempt to find or present additional potentially useful information about the image (such as the image file name, results from a related image search, etc.).