Berners-Lee: Semantic Web's success lies in cooperation

Speaking at a conference in Southampton, England, on Tuesday, Berners-Lee said all that was needed to build the Semantic Web was for existing databases to be exposed in standard formats.

The Semantic Web project, headed by Berners-Lee, aims to add machine-readable content to address the fact that the Web is largely still a mass of unstructured data with little to link groups of documents together, and no way for computers to manipulate the information in pages.

Berners-Lee explained that the power of the Semantic Web comes not from a single source of data, but from when multiple data sources are combined.

"It's the serendipitous reuse that's exciting," explained Berners-Lee.

Citing his conference hosts, Ordnance Survey, as a prime example, Berners-Lee explained that the organization--which makes maps and guides--has huge amounts of useful data that could benefit the concept of a joined-up, intelligent Web.

Berners-Lee added that the subject of the conference, "Terra Future," was particularly appropriate given how location-based information was currently being integrated into Web sites. "Geospatial information is being seen to be exciting by the Web 2.0 crowd, with things like geotagging and Google maps," he said.

Taking photos of the audience with a digital camera, Berners-Lee explained that while the camera he was using didn't have any GPS capability, it still recorded the time and date when each picture was taken. Combining this with data from his calendar, which contained details of the conference, would allow a computer to infer where the picture was taken and add this information to the photo's metadata.

However, none of this is possible if specific industries or interest groups only use their own languages to describe the information they hold. Instead, standards such as Resource Description Framework (RDF) are needed to provide a common vocabulary.

Berners-Lee gave the example that Ordnance Survey records the location of a church, but not its denomination, but that this information was held elsewhere. "When you start querying properties of (geographic) features, it's no longer just a geographic system. It becomes a generic Semantic Web query," he said.

Berners-Lee believes that this commonality can be achieved without having to start from scratch or modify systems. "The most important message is: Don't change how you're working," he said.

The RDF necessary to make the sharing of geographic information systems (GIS) data possible can be dynamically generated, he explained, in the same way as the HTML for large Web sites is dynamically generated by a content-management system.