Please wait for the page to load fully

 

This short presentation uses web technology but works just like other presentation software. To advance to the next slide you can click your mouse anywhere on the page or use the Page up/down or cursor keys. See the Help link at the bottom of each slide for details

A basic statement

i.e. Madonna has a son called Rocco

Let's write that in XML so machines can read it

But here's some more XML

Possible conclusions: Madonna has two sons There are two people called Madonna, each of whom have a son Madonna's son has two names The word "son" has two meanings

Actual conclusions: A machine can make no meaningful conclusions from this data. XML on its own is not enough

We need to identify uniquely the subject matter and the relationships between them. This is done using URIs*

* Uniform Resource Identifier - a techie's name for a URL

Using URIs allows machines to draw logical deductions about the pop star Madonna's son and the biblical character's son

This is the basis of the Semantic Web

We can draw the following conclusions

• There are two different people called Madonna
• They both have sons, as defined at http://family-relationshsips.org#hasSon
• Rocco and Jesus are different people

And here's the crucial bit...

• If I find more information about http://www.popstars.com/madonna I can add it to what I already know whilst being sure that the information is about the same person.

What we can do with it...

This is a typical website as seen in a typical browser

Positive labelling...

After the page has loaded, the browser or helper application fetches the label data and alerts the user by showing an icon.

Positive labelling...

Clicking the icon reveals the information. This is one of the ways machine-readable (i.e. Semantic Web) trust marks are envisaged in the QUATRO project

Positive labelling...

Quatro envisages a similar system applied to search engines. In this case, the icon and extra information are shown alongside the search results so a user can see the trust mark when deciding which sites to visit.

Social filtering...

The Semantic Web offers a truly scaleable method of building trust in websites and, by extension, the labelling information that may be present. When looking at a website, a user may wonder what his or her friends think of a particular site. Have they been there before? Is it good for homework? Can the information be trusted? It would be good to know!

Social filtering...

Shared bookmarks (what Internet Explorer calls Favourites) offer such an option. If a user makes their bookmarks accessible by other user's browsers, either in a community of known people or, on a larger scale, anonymised through a recommender system, data can effectively be sought and obtained automatically.

This has particular potential if it is a teacher making his/her bookmarks available on a school server that can then be accessed by students when they research their homework.

Find out more...

The article that first laid out the vision of the Semantic Web was published in Scientific American in May 2001.

There's a lot of material on the W3C's website, both technical and non-technical.

Find out more about Quatro at the project website.