Annotating the content of community-edited, or 鈥渨iki鈥�, sites like Wikipedia would allow smarter ways to access the information they contain, researchers say.
Wikipedia is free for anyone to edit and contains millions of articles on a diverse range of topics. The remarkable scale of the voluntary project is, however, sometimes overshadowed by concerns over the quality and accuracy of its pages.
Computer scientists at the University of Karlsruhe in Germany have developed modifications to Wikipedia鈥檚 underlying software that would let editors add extra meaning to the links between pages of the encyclopaedia. The team鈥檚 Semantic MediaWiki system lets authors add 鈥渁nnotations鈥� 鈥� or tags conferring meaning 鈥� to articles and the hypertext links between them.
Advertisement
These annotations are buried within tags but are also displayed on relevant pages. For example, the links between pages on London and those on the UK could contain an explanation of the relationship between the two. This would include the annotation 鈥渃apital city of鈥� in one link direction, and 鈥渃ountry of鈥� in the other.
The researchers have set up a running the Semantic MediaWiki software.
Smarter searching
Adding annotations could lead to smarter ways to search sites like Wikipedia, the researchers claim. Project member Markus Kr枚tzsch, says the first adopters could be niche communities who maintain their own wiki sites on specialised topics.
鈥淚 think early adoption will be led by communities interested in data such as animal species information,鈥� he says. 鈥淪emantic information is most useful in situations where data can be clearly defined.鈥�
Annotating this information could let programmers create applications that organise this information in new ways. For example, software could create a visual map of the links between different species of animal recorded using a wiki.
Some web experts call for similar annotations to be added to all websites. This vision for a 鈥渟emantic web鈥� is driven by scientists including web pioneer Tim Berners-Lee.
Wider web
Nick Gibbins a web expert from the University of Southampton, UK, says the Semantic MediaWiki project builds on this idea. 鈥淚t鈥檚 a very interesting piece of work,鈥� he says. 鈥淚t differs from the Semantic Web effort, as developing the ontology will be a community effort.鈥�
The researchers hope eventually to see their software implemented in Wikipedia itself. 鈥淲e are talking to the Wikimedia foundation, and Wikipedia,鈥� Kr枚tzsch says. 鈥淪ome members are keen, but some are dubious about additional complexity.鈥�
Kr枚tzsch concedes it could be a challenge to ensure that the code will support a site as popular as Wikipedia, which receives about 4000 page requests per second. He also acknowledges that even if semantic annotations are added to Wikipedia, there is no guarantee they will be entirely accurate.
鈥淔actual errors are strongly coupled with the semantic errors,鈥� Markus Kr枚tzsch says. 鈥淲e hope not to introduce any more.鈥�
Concerns over accuracy have dogged Wikipedia as its influence has grown. In December 2005 the site was forced to tighten its editorial rules after a prominent US journalist discovered serious inaccuracies in an entry referring to himself.
