Phil Fearnley (my boss) has blogged about 2012 plans has a paragraph about the business justification for the use of the underlying semantic data-driven publishing stack:
“We’re developing a new publishing platform that delivers pages that are dynamically and automatically created. Content can be tagged with an identifier that can be automatically pulled into the relevant page to provide a real-time, extensive, and trusted companion to events. We delivered a page for each country, squad and player during the World Cup in 2010 using this model and we’re scaling this up for next year to deliver unparalleled up to the minute detail on each athlete, country and event. Delivering such a detailed and broad service via traditional editorial curation would be cost prohibitive. “
There’s more detail about the underlying technology stack (bigowlim holding ontologies and instance data, canned sparql queries behind a rest API) and how it fits with the existing Journalism CMS in Jem’s post here.