This page contains thoughts on accessibility and metadata.
Background
The idea for produced a document on accessibility and metadata arrives from the interests
within my organisation (UKOLN) in metadata, my interests in the Web (since 1993)
and my recent involvement in web accessibility issues (member of the DISinHE Steering Group,
attendance at the WAI meeting held at the RNIB, UK in June 1998).
Please note that I am not directly involved in accessibility work myself.
Interested Parties
The following have been invited to comment on this document:
Thoughts on WAI
From an outsider's perspective the Web Accessibility Initiative (WAI) appears to be
focussed primarily on educational issues. Although the WAI Protocols and Formats group
has a remit to address broader issues, from my attendance at their meetings at the RNIB
is seems that the group exists to review the accessibility aspects of other W3C groups
involved in protocol development, and not in being pro-active in developing appropriate
protocols and data formats themselves.
Metadata and Accessibility
Tim Berners-Lee, in various keynote speeches, has promoted the idea of
machine-understandable metadata and has argued that "if a computer can make a decision,
it should." (paraphrase of his keynote talk at WWW6 - I think.). This document
considers the relevance of machine-understandable metadata for accessibility.
Scenarios
A number of possible applications which make use of machine-readable metadata are
given below.
- Personalised formats
- The user fills in user profile details, which enable the server to deliver formats
which are designed for the user's disability.
- Enhanced site navigation
- A server provides a sitemap in an open format (e.g. RDF). A user-agent helper
enables a visually impaired user to navigate quickly through the site.
- Structured information about images
- A Web site containing large numbers of images provides structured information about the
images within the images PNG format. A user-agent helper renders appropriate information.
- Profiles for Authoring Tools
- An intranet is being designed for a particular community (e.g. the xxx). The authoring
tool is configured to provide assistance in authoring for that community.
- Configuration Details for Accessibility Checkers (website areas)
- An accessibility checker (such as Bobby) is configured to avoid certain areas of the
Web site (e.g. areas not the responsibility of the author).
- Configuration Details for Accessibility Checkers (disability)
- An accessibility checker (such as Bobby) is configured to report on named disabilities
(or combinations of disabilities).
- Exploitation of desktop capabilities
- A Pentium III PC enables high quality graphical displays which are needed by users
with a particular disability. The server recognises the client capability through the
Pentium ID.
- Web collections
- A user wishes to navigate quickly through a set of related pages (a document split
into multiple pages). Use of the (unimplemented) <link rel="next"> and
<link rel="previous"> enables this to be done.
I started to give the following examples, but realised these weren't to do with
metadata.
- Exploring a CAD drawing by a visually-impaired student
- A visually-impaired engineering student is accessing an engineering drawing on the
web. Use of Scalable Vector Graphics enables the image to be magnified without loss of
resolution. Attached textual descriptions can be spoken.
- Use of External Link Databases
- The XLink spec is used with an external link database (rather than embedded links).
This enables links to be created tailored for user motor disabilities.
- Use of DOM to manage colour for the colour-blind
- The DOM enables client-side scripting languages to read and update HTML and CSS elements
and attributes. So for a red-green colour blind user the colour of text and the background,
for example, could be amended.
Registries
I think registries (machine-readable stores of information which can be processed
by applications) may be important. For example rather than definitions being hard-coded
into applications, or stored in configuration files which are only accessible to the
application, they could be made globally available in an open way.
Relevant Technologies
The following technologies may be relevant:
- RDF
- W3C's metadata framework.
- CC/PP
- Note submitted to W3C for defining client characteristics.
- PNG
- W3C graphic format, which enables structured metadata to be embedded.
- Transparent Content
Negotiation
- Server feature for transparently serving "best" format when variants available.
- Web Collections
- The note is now obsolete, but the applications described are still relevant.
- XML
- Needed to define structured file formats.
- SVG
- Structured graphical formats should be more accessible than bitmaps (such as GIF)
Additional Comments
Some additional comments:
- Some people within the WAI community feel that approaches based on "universal design
principles" and complete backwards compatibility. I do not necessarily go along with
this. I'm more inclined on use of personalised formats / technologies (which, I think,
is in keeping with W3C's views on use of PDAs to access the web).
- I'm not saying "If your browser can't do x, then upgrade it".
- There are likely to be benefits for all web users in some of the possible solutions
e.g. a site map for use by a visually impaired end user could also be used by an
indexing robot.
What Next?
What next?
- Identify stakeholders
- Software developers
- Researchers
- Dissemination
Accessibility Metadata