Think about a blind person using the Internet; most of the time
most sites provide a bearable service: information can be
retrieved, catalogues searched, and forms filled in; but every so
often they break: important information is contained in an image
without any alt text to describe it, a link has no useful
description, the reading order is illogical, or the submit button
is inaccessible unless you use a mouse. The 10% of problems make
the experience extremely frustrating and time-consuming and in
the worst cases mean that the user fails to complete what they
had intended to. To avoid this, users would really like to know
before using a website whether or not it has been designed to
meet their particular needs.
Some sites may help a bit by including an icon saying WCAG AA
compliant. This is only of limited use because it is not easy to
search for sites that include this icon, it's not always easy to
find the icon on the page, and it is only an assertion by the
owner of the page that it is in fact accessible.
What is needed is some formal machine-readable methods of being
able to assert the accessibility of a page, have this certified
by a third party, and be able to report any non-compliance.
Just as a blind user would appreciate this information, a website
owner who has gone to the trouble of making their site
accessible, wants to be able to advertise that in a usable way.
If the assertions and certifications were machine readable then
this information could be used by:
- Search engines to prioritise page and sites that support the
- Browser plug-ins to announce that a page supports their
requirements (or that the site specifically does not support the
- Agencies that monitor accessibility of websites as they could
quickly discover sites that have not made assertions and
therefore may not be compliant.
Making assertions about accessibility of a web site or page is
just part of a bigger requirement of making assertions such as:
child friendly, contains adult content, scientific claims have
been peer reviewed, etc.
This requirement has been known for many years and some attempts
have been made to address it (notably PICS Platform for Internet
Content Selection) but nothing has been sufficiently general and
robust to really catch on, except in very specific areas, such as
The biggest stumbling block to implementation was the lack of a
standard format for encoding the assertions and certifications.
This has now been removed by the publication of the W3C
Recommendation POWDER (Protocol for Web Description Resources)
in September 2009.
The heart of POWDER is the Description Resource (DR) which is an
XML document that defines:
- Who it is issued by.
- When it was issued.
- Which resources it covers (for example web pages).
- What it asserts about the resources.
- Who certifies the assertion.
Resources that are included in the DR will typically include a
link to the DR, this will enable a user to know what has been
asserted about the resource and what trust they can have in the
assertion. If the issuer of the DR is unknown to the user and the
certifier is also unknown then the level of trust will be low;
but if either or both are trusted by the user then the level of
trust will be much higher.
A company called Segala was one of the organisations involved in
the development of the protocol. Segala span off a new
company—MetaCert—that is now working (according to
its web site in 'stealth mode') to take POWDER and build the
necessary technology to implement the protocol and exploit its
When these technologies start to become commercially available
the next hurdle will be to create an impetus for their use. If
few websites use the technology then search engines will not
provide the specialise search facilities that take advantage of
them; and conversely if search engines and browsers do not
support the technology then there is no incentive for web site
owner and developers to add the extra function.
An ideal way way to create the impetus would be through
legislation. For example Section 508 could be extended to say
that companies wishing to do business with the US government
authorities must have compliant websites and this must be
asserted by the use of POWDER. This may happen in the future but
I suspect it is unlikely to be the early impetus that is
Another option is a grass roots drive. In this scenario an
organisation representing people with disabilities could create a
DR repository which their members could search to find accessible
sites. In the first instance this might be created by a reporting
system, the users would report non-conforming sites or identify
A third route is through commercial benefit. The first example of
this will probably be in the adult entertainment arena where ICM
Registry are planning to set up a new Sponsored Top Level Domain
(sTLD) .xxx ; MetaCert are in advanced discussions with ICM
Registry to be the labeller of choice. The labelling will assert
that the web site abides by a code of conduct for adult content,
this should help to protect people from accessing inappropriate
My idea for the promotion of accessibility labelling based around
POWDER is as follows:
- Reporting web site accessibility could be done through a
browser plug-in (or by code integrated into the browser that
supports POWDER). Whilst viewing a page the plug-in could be
invoked and the user could report an issue or identify a page in
a positive way. Firefox has a rudimentary version of this concept
in the Report Broken Web Sites function under Help. If the user
was willing, the plug-in could identify all the pages viewed and
assume they are accessible if not reported otherwise.
- All the information collected could be collated and
synthesised to create the DR repository. This method might
quickly produce a useful and significant body of knowledge which
could then be attractive to search engines to support. Its
slightly arbitrary nature (those sites that the users had
reported on) might encourage website owners to create their own
more complete and reliable DR's.
- I believe that the ability to report inaccessible pages in a
consistent manner is fundamental requirement. If it is not
available and a page claims to be accessible what is the user
meant to do? Some sites might include a feedback option and I
would recommend that sites do provide this facility. However, it
is inevitable that each site feedback system will be subtly, or
even grossly, different from other sites and therefore the user
is less likely to use them. If the plug-in was used the central
site could pass this information on to the relevant site owner,
this would be easier if the site used POWDER.
There are several organisations that might be interested in
pursuing this idea and I would be very interested in hearing from
them if they do.