One of the main themes in NoTube is connecting the Web and TV through the use of Linked Data to enhance the TV experience. In the BBC’s part of the project we are specifically interested in exploring how Linked Data can be used to help people:
1. Decide what to watch
With so much choice available, it’s becoming harder to decide what to watch. We are therefore looking at ways of filtering programmes of interest and bringing them to the user’s attention.
2. Find out more about a programme
Currently this involves having to remember to look for the relevant information once the programme has finished, or interrupting the programme to look for it now, with the risk of missing something else of interest in the process.
3. Have smarter online conversations about TV
There is no shortage of online conversation about TV: witness how frequently Twitter trending topics are TV-related during peak-time viewing (at least this is often the case in the UK). However, Web-based conversations, about TV or anything else, cannot be held without URLs that allow people to refer to the thing they are talking about.
The need for dereferencable URLs to identify programmes
Key to all of these goals is the disambiguation of specific programme episodes with unique IDs. For example, the BBC’s /programmes website provides persistent links to programme episodes from which machine-processible programmes metadata can be accessed. Resolving a /programmes URL (such as the URL for an episode of Eastenders: http://www.bbc.co.uk/programmes/b00vzxms) gets you data about the programme, available in various formats including RDF, such as when it’s on next, whether it’s available to watch on iPlayer, a description of it, whether it was part of a series or not, what version it was (signed, shortened), and sometimes who was in it, what role they played, who the director and writer were, and so on.
These URLs provide the starting points for linking out from the programme episode to related information from the Linked Open Data cloud, adding background knowledge and context to the programme and making new, and sometimes surprising, connections.
In helping people to decide what to watch, we anticipate that some of these types of connections will be of particular interest to people. We therefore propose that filtering and presenting the links that led to these connections will provide useful information about whether a programme is worth watching or not. For example: “This programme stars [your favourite actor] and/or is about [your town] and/or is about [a topic of particular interest]“. This approach can also surface programmes of interest buried in the long-tail (i.e. in a large collection of on-demand video content), because it does not rely on lots of people having watched something before it can be suggested.
In helping people to find out more, using the programme URL to automatically link out to related information on the Web from the Linked Open Data cloud allows relevant background information about a programme to be ‘pushed’ to the user in a convenient and unobtrusive way, without the user having to search for anything. For example, an early NoTube demo showed the Wikipedia page about a costume drama, The Forsyte Saga, being sent to the user’s smartphone whilst they were watching the programme on TV. The Wikipedia page provides detailed information about the Forsyte Saga books, the author, TV and film adaptations, main characters, plot, and so on. The user could quickly bookmark the information for later if they didn’t want to interrupt the programme by reading it now.
In supporting online conversations about TV, the URLs create linkage points into social network discussions, as well as shared bookmarking and commenting systems, enabling the video to be referenced permanently as a focal point that can be annotated. The aim is to extend this by providing unique identifiers (based on URLs) for particular events within an episode. This would allow people to reference specific sequences or moments in a programme, for example when discussing “that goal”, “that outrageous comment” or “that hilarious sketch”.
We are currently in the process of planning some user evaluation studies which we hope will validate our hypotheses concerning the advantages of using Linked Data to support these user goals. In the meantime, much of the technical effort is going into setting up the required infrastructure by using URLs to uniquely identify programmes. This is crucial for creating greater coherence for machines, and thereby for TV viewers, who we hope will benefit by enjoying an improved experience.