Crowdsourcing ALT tags

IBM had a news article yesterday about a project to increase screenreader accessibility for HTML sites by a third-party metadata overlay… if someone notices the screenreader representation of an HTML page is inadequate, they can request that a sighted participant add appropriate tags to a centralized database, and this data is then pulled and added to the site the next time a program participant visits that site.

Simple idea, but hard to talk about… Jacqui Cheng has another article at Ars Technica: “The Social Accessibility Project hopes to circumvent the entire problem of dealing with developers by allowing users with screen readers to automatically report problems with various web pages back to IBM. Volunteers can then sort through IBM’s database of accessibility issues and create their own metadata for each element. When users with screen readers return to that site, or go to other sites visited by project participants, they will simply load the latest information from the database and be able to navigate the web with greater ease.”

Neat idea, but I wonder what the social repercussions will be. Website owners have typically rebelled against third-party annotation networks, whether Third Voice or Smart Tags or any of various other efforts. One hot issue is when people see their work mashed-up without their explicit consent… another hot issue is the confirmation of metadata accuracy within such a voluntary network… I guess a third hot issue could be when trying to determine whether the HTML page you experience on one machine is the same as the same address on another machine. Lots of issues to iron out when the canonical page representation becomes mutable.

Still, this is intrinsic to the HTML model… when you can’t predict the user-agent, you’re never quite sure how others will experience the media you create. Greasemonkey is likely the posterboy example of not knowing how your HTML will be rendered. Third Voice was ultimately rejected by the webdev arbiters, but I think it will be harder to argue against this Social Accessibility Project.

No conclusion on my part… this collaboration on improving the webpages of others seems like a very delicate problem, and I can’t predict how it will be received by website designers and website owners in general. Interesting.