ETech – BBC Programme information pags

More BBC chaps presenting innovations in the BBC world
Given something, an identifier makes it addressable – then you can build structures and systems around it. Eg. barcodes, Amazon ASINs, postal codes. The BBC has 40K hours of national television broadcasts across eight TV stations, 76K hours of national radio broadcasts, 42 local radio stations, world service, BBC America.

What are PIPs (program information pages)? It’s a system that represents all BBC programs on the web. They group programs by series (2nd episode of Friends from season 1). Logistical issues: how to you deal with schedule changes; BBC is a huge org ~30K people, legal issues. Episodes are their detail pages (eg. episode 1, friends).

The web product. Page for every episode of every program that the BBC creates. The Web structure is a simplified reflection of the data structure. Wanted to improve navigation between episodes. Every individual episode is uniquely identifiable and addressable forever. Sample was Radio 3. Schedules persistent forever. Over the next few years, expect to see all eighteen BBC national TV and radio networks becoming PIPs. Every program will have it’s own unique identifier. Do I hear “customer who visited this program page also visited…” šŸ™‚

Few architecture tidbits. Electronic program data has no meta structure. Feed processing cleans up the data using various heuristics. SMEF is the data model they created at the BBC (open). They were hand wavy about what heuristics they used to handle authority issues (like actor, directors, etc).

Advertisements