Semantic Web and Accessibility

Automating the Web for everyone - Charles McCathieNevile

This presentation was done for the Association of Internet Researchers Conference, 18 October 2003. It was presented through a chat, so you can read the presentation discussion as you go...

These slides are also available in a one-page version and an SVG version. All slides were auto-generated from the one-page version with a version of the W3C slidemaker linked to JackSVG

The one slide Semantic Web...

Something has some relationship to something else

• The information is on the Web
• And can be aggregated
• You can query the collection
• You can refine definitions

Wouldn't it be cool to...

• note an illustration of an idea?
• find descriptions of a painting?
• get a useful version of a document?
• find people like us?
• make this chat work better?

Let me paint you a picture...

General relativity is easy... (for physicists)

Einstein laid it all out nicely (Oh! In German!)

and the rest of us watch the movie

- (We'll come back to the movie)

How did the Semantic Web help?

A work

• has title "Theory of Relativity"
• has creator Albert Einstein
• has language "German"
• has a version - At bartleby
• has a version that
  • has format of "a movie"
  • has creator "NCAM"
  • is an alternative for the bartleby version

  ....etc

Err, the picture?

Some key features

• Defined vocabularies

  - (which are then mixed)

• Everything is on the Web
• Each "fact" can be anywhere

Finding a description

Easy - we just make one

Well, one each, so people can choose:

The Mona Lisa has

• A gallery catalogue entry
• A description of the paints used
  • A description of the blue
  • An ingredients list for the red
  • "what is special" about the black

So we can

• Deal with different levels
  • Overview
  • "sub-parts"

Put together pictures without seeing them.

(A network diagram isn't a Mona Lisa)

Have our "alt text" without the drudgery...

So, a useful version of the theory

I am

• In a noisy place
• A french speaker
• Lacking concentration

I want something

• About relativity
• Appropriate for junior high

Some key standards

ACCLip

From the IMS global project - describing a learner's needs

Dublin Core

basic cataloguing from DCMI

EARL

Conformance to requirements, from W3C

CC/PP

W3C - what kind of browser features are available

WCAG

W3C - accessibility features like captions, ...

Leading us to...

A captioned, multilingual, described movie version by an authority...

Find people like me

An important real world case:

Maps are hard to read, but map information is ideal for the Semantic Web

Friend of a Friend (FOAF): about people

• Who they know
• Where they are
• What interests they have
• Photos, Documents, etc
• What language they speak

Improving realtime chat

Chatrooms can be useful multimedia collaborative tools:

chaalsNCE BLURB: Improving accessibility in real time

dc_rdfigD: Improving accessibility in real time from chaalsNCE

-->|JibberJim (SJ@dsl-217-155-143-66.zen.co.uk) has joined #rdfig

chaalsNCE D:

dc_rdfig blurb

dc_rdfig Improving accessibility in real time

chaalsNCE D: A bot could pick up a command to add a description to an image

dc_rdfig Added comment D1.

chaalsNCE D: Or a part of an image, as in JibberJim's Shared SVG Whiteboard

dc_rdfigAdded comment D2.

chaalsNCE D: This would allow collaborative annotation rather than require one person do all the work

dc_rdfig Added comment D3.

chaalsNCE D2:

dc_rdfig (chaalsNCE) Or a part of an image, as in JibberJim's Shared SVG Whiteboard

chaalsNCE D2: Or part of an image, as in [JibberJim's Shared SVG Whiteboard| http://jibbering.com/svg/whiteboard/]

dc_rdfig Replaced comment D2.

Semantic Web programs can hang out in chatrooms and provide interesting support...

Thank you

Questions?

Question: please advise on visualization strategies for mapping semantic relationships - can you provide references or web links for research programs in this area?

There are some important things. it turns out that people can't cope with too many things in one diagram, so you need ways of collapsing out the stuff that isn't immmediately relevant. So you are looking at something like how to understand that you say zoo and I say menagerie, and the meanings overlap. I can't provide good references for visualisation research in this area, but most of them would be from traditional graphic communication augmented with the ability to do dynamic presentation. There is certainly work in this area in teh semantic web - tools such as IsaViz with its Graphic Style Sheets, and GraphViz which represents such information. People who studied sets, in maths (I did it in early high school, and at a simplistic level in lower grade school as a prelude to multiplication) will recall ways of showing that some things are in one or other set, and some are in a couple at a time. And that the diagram gets very complicted very quickly. Fortunately on the Web we have dynamic formats like SVG, so you can concentrate on the thing immediately at hand...

Question: is PICS a good role model for metadata?

response: A good role model... Hmmm. In the sense of being there before it is. PICS developed into RDF - the language that W3C now uses for the semantic Web. PICS (and in another way Dublin Core) demonstrate the problem with "flat" models - with RDF you can build up layers of complexity as you realise the complexity of the problems you deal with. It is all very well to call a page of pictures and text a text document (as Dublin Core currently does) but for accessibility that isn't enough information. Similarly PICS is really just name-value pairs. It is very hard to describe the relations between relationships in PICS. But it did have the feature of demonstrating that allowing anyone to make the statements is an important step forwards.

Bye - thanks again