W3C Workshop on Metadata for Content Adaptation: Attaching Metadata to Web Content

Shadi Abou-Zahra, W3C/WAI
Chair and Team contact for ERT WG

Introduction

The Evaluation and Repair Tools Working Group (ERT WG) develops techniques to evaluate and repair Web content for accessibility. The primary focus of the group is to develop the Evaluation and Report Language (EARL), an RDF syntax, for documenting evaluation and repair information. While EARL is intended for reports generated by Web accessibility evaluation tools, much of the design challenges relate to device independence.

Evaluation and Report Language

There are different usages of EARL including post-development conformance reviews of Web content to the Web Content Accessibility Guidelines (WCAG). A more active usage is annotating Web content with EARL meta-data to indicate test results which are not necessarily limited to WCAG checkpoints. In fact, EARL needs to be generic enough in order for it to be forward compatible which makes it also interesting for other test-suite developers.

Web Content Annotation

By annotating Web content with meta-data which describes test results, the content can be matched against user profiles or preferences. For example, Browsers could remove flickering images from Web pages for people with photo sensitive epilepsy, or restructure a page- or even a site-wide navigation. Search engines could also employ similar techniques to deliver pages with specific features, or at least make the meta-data available for the Browsers.

Getting a Grip on Web Content

So far the term Web content remains unclear. While some WCAG checkpoints target specific markup elements or even a smaller scope, sometimes they address whole pages or Web sites. Attaching test results (regardless whether they are post-development reports or active annotations) to Web content remains a challenging issue in the development of EARL, especially in the face of live content which is generated dynamically.

Scenarios and Approaches

Basically there are three fundamental approaches which have been raised within the group. The trade-off is always between the complexity of the solution (i.e. implementation) and the delivered features.

Date

Using the HTTP header, the dates of the result assertion and the Web page are compared. This method is the simplest but a slight change in the document could lead to the obsoleting of the whole set of related tests.

Hash

Hashing the Web content and storing the control in each result will allow Browsers and other tools to determine the validity of the result assertion with the specific section of the Web page.

Digest

Similar to hashing except that the digest is unique within a Web page so that matching content can be searched if it has been moved with the document but not otherwise changed.

Research Questions

The development of EARL will require the ERT WG to address questions which may also be of interest to other groups working with Semantic Web technologies such as the Device Independence WG. Exchanging ideas and best practices during the Workshop on the following items will be the mail focus of this presentation:

• how to point to Web content which may be a piece of a document or a whole site;
• what hashing and digest mechanisms are available and portable across devices;
• are there other directions and approaches which the group did not yet consider;
• do any implementations of any of the discussed potential solutions exists;
• which groups are interested in collaborating in developing a robust solution.

Last changed on $Date: 2004/10/05 20:03:23 $ by $Author: rhanraha $