By Magda Hercheui

The conference presents case studies on how the concept of semantic web is changing the way we understand the management of information, from vertical (hierarchical) forms of classifying content, to horizontal solutions which work through tagging systems and association of content from different documents. The more content suppliers adopt such semantic solutions, the more any Internet user will be available to find the right answer to their questions. The semantic web supports the building of search platforms that permit people to have more precise answers to their questions, even when the elements of each answer come from different sources. In breaking documents into pieces of classified information, all sorts of combinations become possible, overcoming the limits of hierarchical classifications and document-based search (page-oriented architecture).

Main challenges

Although this revolution is coming, and will change forever our perception of information management, there is much to be done, and some obstacles cannot be overcome in the short term. A first problem is the standards for classification of pieces of information, the ontology. In order to ‘give life’ to the semantic web, each document or database needs to follow a standardised ontology that will permit this document or database to be related to others.

The creation of standard ontologies is not easy. We can look forward to a battle of standards in ontologies, as documents framed by one ontology will not talk easily to documents that follow other standards. Companies can create translators to permit communication, but we anticipate problems. Classifications are intrinsically cultural objects – not scientific ones, although some feel more comfortable in believing in right and wrong classifications. So we can expect as well a lot of differences in ontologies created by people from different cultures, a big obstacle to the development of a truly robust and pervasive semantic web.

A second problem is related to the quality of information. The vertical/hierarchical organisations of documents and pages permit us to decide that one source has more legitimacy than others. So although Google brings one million documents, I am able to prioritise those that come from legitimate sources in accordance with my knowledge and perspective. The semantic web permits the putting together of pieces of information from different sources. The question is: should I trust all the sources that have contributed to the answer to my question? The answer is definitely not. To overcome this problem, the documents generated by semantic web solutions need to make clear the sources of different pieces of information in the same document. They also need to allow users to define which sources are legitimate in their opinion. To choose for example that the BBC may enter the search process, but not the Sun.

So much yet is to be done in terms of developing the semantic web infrastructures, but even after having consolidated standardised ontologies, we still will need a lot of work to provide transparency and credibility. Indeed we can forecast an emergent business model, of aggregators that will provide good standards to specific niches of information users, and that will care for transparency and information quality by selecting relevant sector-appropriate sources.

Let’s prepare for a new life on the Internet, and new business.

See below other links to relevant seminars on semantic web during the Online Information Conference:

http://www.online-information.co.uk/online09/conference_presentation_2009.html?presentation_id=606

http://www.online-information.co.uk/online09/conference_presentation_2009.html?presentation_id=637

http://www.online-information.co.uk/online09/conference_presentation_2009.html?presentation_id=646