Insight BBC Olympics Website and Web 3.0

The BBC Olympics 2012 website represents a significant undertaking. It's planned to have more than 12,000 pages with individual pages for athletes, teams and event, and thousands of hours of on-demand video. With a dynamic website of this size, the management and “orchestration” of the content is a huge challenge. Particularly when it covers reports, results and statistics, all with live updates.

The approach they are taking is using “web 3.0” or “semantic web” techniques to semi-automate tasks like linking between articles on similar topics and tagging what the article is about.

BBC World Cup 2010 Site

The BBC trialled this a approach on the 2010 World Cup site. For the World Cup the BBC had nearly 800 pages on players, teams and groups alone. This would have been very difficult to do manually without a considerable amount of effort.

The basis of their semantic approach was to implement background knowledge on the players, groups and matches. For example that Frank Lampard plays_for England and England is_in Group C. These are represented as an “ontology” using RDF triples: ["Frank Lampard", plays_for, England], etc. This then allowed them scan for terms to automatically link any articles mentioning Frank Lampard to England, Group C, etc. and infer relationships between players mentioned in a report with their teams and competition without the journalist having to explicitly add them.

The site worked well through the World Cup with over 2 million unique page requests being served per day and more than 100 RDF updates being generated per second at peak times.

Recent Developments at the BBC and Gearing Up for the Olympics

Since the World Cup, the approach has been rolled out across the rest of the BBC's sports coverage with a major refresh of the BBC Sport website. The refresh has introduced a sports ontology using RDF and linked data covering all sports, events and competitions.

This is the underlying architecture that now gives the automatically updating football league tables for example and lies beneath all the BBC's sports content.

The Olympics website be a further development of this and will have pages for over 10,000 athletes, over 200 countries and over 400 disciplines. There will also be nearly 60,000 hours of on demand video content, all of which will be time coded and annotated with appropriate meta data.

Running this amount of content without the automation provided by the semantic web approach would be impractical.

In other areas, the BBC has introduced ontologies for programmes and wildlife.

What does Web 3.0 mean for You?

The BBC is obviously a very specialist web user with very particular needs. So what does the semantic web mean for other organisations?

The advantages that ontologies, RDF and linked data can give are:

• Content can be re-purposed or reused more easily
• Content can be tagged more easily, more consistently, and potentially semi-automatically
• It allows content to be dynamic
• It enables content to be more findable for users and for search engines

It doesn't have to be too difficult to make the initial steps. An ontology can be viewed as an approach to taxonomy with more rigour - it is more formal and includes the relationships between terms.

This approach is likely to be most useful to organisations that are specialists or have large content sets, particularly those with significant sized product catalogues or knowledge content. This would include specialists across all market sectors and many NFP organisations, particularly membership organisations and medical charities.