Gregg Vanderheiden 3/28/2003
Proposed is an entirely new way to look at Web content accessibility. It includes:
NOTE: At this point I do not know if this will work out. It is proposed in rough form to get peoples ideas and to raise it for discussion because of the above advantages. So this is not finished and not necessarily the way I want to go unless it all pans out. So raise questions. Make suggestions.
With the proposed approach, the primary goal would be to ensure that material is machine perceivable, operable and understandable. That is that special user agents (either on user's computer or on a server) are able to successfully identify the words, phrases and other semantic content to operate the controls and to be able to decode the underlying information and functionality.
A second level priority would be to make the material directly perceivable, operable and understandable using standard mass-market user agents without any assistive technology.
This approach is chosen for a number of reasons:
First, individuals who have disabilities have a very wide range of needs. If the content can be perceived, operated and understood, then the information can be re-presented to individuals in a form and format which best meets their needs and constraints. Individuals who are blind can use screen readers or talking user agents, individuals with low vision can have the text scaled to fit their limitations with regard to field of view acuity, color perception, etc. Individuals with physical disabilities can have the content and functionality rendered in a form which is easy for them to navigate, manipulate and control. Individuals with language disabilities can have content presented visually and/or auditorally with highlighting and or and other "scaffolding" that would assist their reading and comprehension. Individual with language disabilities or for whom the content is not in their primary language could have it translated into their primary language including sign language. Individuals with cognitive disabilities could have the content re-presented in simpler forms or have it supplemented with non-text content which was familiar to them or even custom to them which would make it easier for them to comprehend. Content from different sites could also be presented in a similar fashion if this was necessary or of great value to the users.
A key aspect of this however is that all of the above could be done without requiring the author to present the default presentation of the material in a form to accommodate any one of these groups. That is, if the author met the "special agent perceivable/operable/comprehensible" criteria, then they could have their default presentation be whatever they thought it should be to match the bulk or sweet spot of their audience or to achieve whatever effect they desired.
A level 2 priority would be to make the default presentation accessible across disabilities without requiring AT. That is, sites wishing to be accessible to people using standard browsers/agents (without any AT) would make their sites accessible to as many different disabilities in its default form or in one of multiple forms available directly from the site.
A level 3 priority would be to make sites even more accessible to particular popultions by tuning the sites to those populations
In discussing this, there are a number of definitions or terms that are important. They are:
Free Baseline Agent (FBA) – an agent defined by W3C as the baseline agent and is available free to users. [note that the ability to localize the agent for different countries or locations is important and effects conformance]
Public Available Transcoding Server (PATS) – These again would be servers which would be free to the public and which could accept content and could carry out transformation s. PATS must be implemented in a way that allow them to be localized to different languages.
Special Agent Software (SAS) - This includes both Free Baseline Agents and Publicly Available Transcoding Servers (FBAs and PATSs).
Reference Transcoding Server (RTS) – A server that can be used to check content for conformance. The RTS reflects the abilities currently available in the (free) Publicly Available Transcoding Servers (PATSs and the FBAs). RT Servers are updated over time to reflect the current state of FBAs and PATS. RT servers are always backward compatible. That is, content that passed with earlier RTs would pass with later RTs. This is necessary so that Web pages that were "conformant" yesterday do not become non-conformant when an RT is updated. As PATs and FABs are improved, reference transcoding servers are updated to handle new types of internet technologies.
RTSs would separately report the ability for FBAs and PATSs to handle content and could also report their availability in different languages and locations.