The Online News Association scheduled its semantic web session for a sunny San Francisco afternoon and still drew a decent crowd, which said something about how much appetite there was for working out what the idea meant for journalism.
The three panellists defined it three different ways. Paul Berry, CTO of the Huffington Post, described it as the ability to pull context out of any body of text and structure or categorise it — whether a piece is angry, whether it is insightful. Kurt Cagle of MetaphoricalWeb.org offered the deliberately unhelpful formulation that the semantic web is about gaining meaning from meaning, then explained that it concerns assertions: whether something is true or not true, and how assertions relate to one another. Peter Offringa of CBS Interactive brought it back down by describing it as the ability to analyse text and find content.
On practice, Berry’s first step was tagging — with the caveat that correct tagging is not always the valuable kind. The Huffington Post had tagged its story on Heath Ledger’s death as “Keith Ledger”, because that was the misspelling people were searching for, and the misspelled tag drew ten times the traffic of the correct one. HuffPost also used Adaptive Semantics to moderate comments and manage its community.
Cagle’s larger argument was that audiences were moving away from brands: “We don’t have a music industry, we have thousands and thousands of industries.” If a category can be organised hundreds of thousands of ways, each micro-category is also a micro-market. Offringa gave the commercial version: CBS Interactive used semantic technology to build topic pages across its CNET properties, gathering related content in one place, and to associate users with particular interests and products — which is what made them worth something to advertisers.