While I was on my holidays there was a storm(ette) about
rev=canonical and how it isn’t possible in HTML 5 because
rev isn’t in the spec. (Apparently, the answer is to use
Mark Pilgrim published an article about link relations in HTML 5 with more information about the
rel attribute, which I found interesting; I had no idea that relations such as
rel=author were available to allow auto-discovery of license information, and author details.
So I want to float the idea of
rel=accessibility that would allow assistive technologies to discover and offer shortcuts to accessibility information, such as a WCAG 2 conformance claim, or a form to request content in alternate formats (for example).
The reason this would be useful is that links to such pages are generally right down in the footer of the web pages. This means that, for screenreader users, they have to navigate to the end of the page to find the link, or not know it exists.
Ironically, on sites that really do need a link to accessibility help (because of lack of structure to navigate with or huge amounts of content before the footer), those who need it are unlikely to find the link to the help.
In the “bad old days”, helpful developers would give an
accesskey attribute to that link (which are generally undiscoverable to the human or to a parser, and which often conflict with assistive technologies’ command keystrokes).
A standardised way of indicating the related accessibility information would be better and not rely on arbitrary keys chosen by a developer.
So, should I propose that
rel=accessibility be added to the list of values? It looks to be an arduous process; although you don’t need to prove your worth to the HTML 5 gatekeepers, you do have to prove your worth to the microformats gatekeepers.
I thought I’d ask you guys first—is this a good idea?