Semantic Web and Metadata

W3C has worked on making a single resource accessible to the maximum number of people. Unfortunately, as Jutta Treviranus points out, this means resources are, in fact, optimized for no-one. Sadly, we learn that less than 3% of resources are likely to satisfy W3C guidelines. So - after more than 10 years of WCAG, there has been relatively little progress in making the Web accessible.

We talk about 'user-centred' design but it is often 'users-centred' - what is best for a lot of people or will help a lot of people. Essentially, 'single user-centred' design needs to be concerned with what an individual needs. That might not be something that is accessible to a majority of other people but rather what is useful to them. Labelling a resource as 'universally accessible' (likely to be useful to lots of people), does not ensure it is good for the individual but, more, how does the individual user know if the resource will be good for them?

Nancy White talks about three kinds of communities of people:

• -------------me --------- we (communities) - the network (distributed, open connections)
• I do stuff for myself - we share stuff ------- stuff is there for everyone

Question is: how do/should we manage stuff?

• as 'me' - if it's not accessible to me I change it so it is
• as 'we' - if I find it's not accessible to my friends, I fix it so it is
• as a network - how do I know what people will need? who should choose for them? how do I provide for them to make the choices?

• as 'me' - I use my desktop
• as 'we' - we use fileMaker pro
• as 'the network' - for them, we depend on metadata

Also, if the focus is on 'the network', we get the benefit of what is called 'the network effect' (Metcalfe's law). If Metcalfe is right, we will make more progress than before because when we share the making of alternatives, we end up with a lot more of them. We 'distribute' the effort but need to find the alternatives in a 'distributed' environment. So the idea is that we use more people to make alternatives and we use the technology to deliver what we need.

How we do this is perhaps described by the term 'linked' data' which is often understood to be a new name for the 'Semantic Web' (Tom Heath). The idea is that the 'Web' is not just a huge collection of data but a web of data in which the data is interwoven. TBL wrote:

The first step is putting data on the Web in a form that machines can naturally understand, or converting it to that form. This creates what I call a Semantic Web – a web of data that can be processed directly or indirectly by machines. (Weaving the Web, 2000, ISBN-13: 9781587990182)

So what does he mean?

OK, but we know this means two levels of attention for those of us working on accessibility:

• making the content sufficiently flexible so that it can be handled by assistive technologies

and

• making the content that a user needs, discoverable by them.

(Actually, the same works for all users because there is effectively no difference between people with assistive technologies, people with phones, people in noisy places, or whatever...)

So -

• how do we make more flexible content? WCAG, ATAG, UAAG, etc
• how does the data get linked? by being described in a way that machines can understand, by 'metadata'. Remembering that computers are basically not good at speaking natural languages, we have to constrain ourselves in our descriptions so they can understand.

I think of the descriptions we need as follows (SWeb-wise and then Metadata-wise):

I even go so far as to use the following metaphor:

I have used the octopus model to help design the AccessForAll metadata.
The image below, for example, shows that the 'octopus' model is not being applied:

What I like is a set of Dublin Core Compliant Accessibility Terms that is also Semantic Web Interoperable

See http://dublincore.org/accessibilitywiki/

For convenience, I recommend the single DC Term accessibility as it will make a huge difference to a lot of people but where possible, the full set is desired.