WebAIM - Web Accessibility In Mind

E-mail List Archives

Re: feasibility

for

From: Greg Kraus
Date: Aug 17, 2011 11:42AM

One of my colleagues here at NC State, Sina Bahram, is doing research
into using artificial intelligence to assess Web page accessibility.

http://dl.acm.org/citation.cfm?doid=1969289.1969329

He has some followup research forthcoming as well.

In regards to crowd sourcing accessibility changes for Web sites we
have developed a system at NC State that lets us do just that.
Currently it is only used for internal purposes, but it basically lets
me modify a live Web page I don't control in order to correct
accessibility issues and, more importantly, communicate issues back to
developers by leaving comments within the context of the live Web
page. While I don't have a live demo that I can easily show you yet
(although that is actually in development right now) I can tease you
with a presentation I am giving on this.

http://www.colorado.edu/ATconference/SessDesc2011.html#Don%E2%80%99t%20As

Greg

--
Greg Kraus
University IT Accessibility Coordinator
Office of Information Technology
North Carolina State University
919.513.4087


On Tue, Aug 16, 2011 at 6:31 PM, Allen Hoffman < <EMAIL REMOVED> > wrote:
> I know this list focuses on getting web pages to be accessible,
> following good guidelines and standards, but, I'd be interested in
> feasibility thoughts on the following concept.
>
> Can a FireFox or IE plug-in be written that:
> 1.  Assesses Web pages for high priority inaccessibility issues, and
> allows end-user, and crowd-sourced remediation.  for example, a page
> appears and has three unlabeled images.  tool allows end-user to label
> them, and then share via cloud with all other users.  WebVism is more or
> less this in a nutshell.
>
> Then take that and:
>
> 2.  Develop pattern matching and sharing of pages to automatically
> remediate other inaccessible things, and do them without end user inputs
> so much--or rely upon patterns of end user remediation for similar
> patterned content.
>
> I think this could work by in-browser DOM change on the client, and
> would not require a proxy in the middle.
>
>
>
>
>
>
> Allen Hoffman