(Li) The World Wide Web (WWW) is still comprised of mostly HTML documents. It is made by people for people. Machines do help them surfing but are themselves "stupid". This was of no great importance at the beginning of the Web era in the 1990s. Meanwhile, however, there are thousands of millions of websites and in this jungle–and despite Google–it is becoming increasingly difficult to find, organise and tend to information.

The selfdefining Web

After inventing the WWW in 1990, Tim Berners-Lee turned his attention to the creation of a "Semantic Web", and he has been working on this vision ever since. By the means of an self-describing Web, machines should be able to find and use information faster. For example, via a "semantic travel search engine", a user should be able–using input such as destination, time and budget–to obtain from the Web offers from diverse travel agencies, airlines and hotels in order to find the cheapest possible combination of airline ticket, hotel, car rental and sundry items. Finding the best possible offer with a single click would save time and money.

This is made possible by complementing existing websites with machine-readable descriptions. The basis of the notation is provided by XML (eXtensible Markup Language) and RDF (Ressource Description Framework) provides a uniform vocabulary to describe the content of any given document. Finally, using the Web Ontology Language(OWL), more complex statements and relations about Web resources can be made, allowing reasoning about Web resources.

Recently, in the ETH World Showcase (1) the Web technology expert Erik Wilde from the ETH Computer Engineering and Networks Laboratory (TIK) presented the approaches that a number of research teams and the World Wide Web Consortium (W3C) (2) are taking to develop a Semantic Web. Besides prototypes and examples of application, as yet unsolved problems arising from the generation leap in the global data network are addressed.

When machines go surfing by themselves: The ETH World Showcase poster for the Semantic Web. large

Mostly research, hardly practice

In contrast to the simple–and for that reason, widespread HTML–Semantic Web technologies are considerably more precise, more powerful, but also more complex. Moreover, a globally uniform description of data calls for a common understanding, common values, a common cultural background and ... trust. This seems to work well in a lab, but outside, in an indescribably multi-faceted world, problems abound. This is why the Semantic Web is still primarily a subject of research and, in contrast to the WWW or XML, has hardly been put into practice.

Modesty boostsunderstanding and trust

More modest approaches attempt to reduce the concepts to small, easily understandable and describable areas. This makes it easier to build a basis for trust and a common understanding. One example of such an approach is the ETH World project "ShaRef" (3), which Erik Wilde initiated and which was also presented in the first ETH World Showcase of this year.