TV Un-Festival: Chris Jackson – A community effort to improve metadata

I’ve been MCing the TV Un-Festival all day, and it’s been fun so far. Right now they are recording a podcast, which I’m not going to blog because at some point you’ll be able to listen to it yourself. Meantime, here’s a short burst of blog posts that I’ve put together throughout the day for your entertainment. (Note: There was no official schedule, so if I’ve misspelt names, please accept my apologies.)

Chris Jackson – A community effort to improve metadata
Chris is a freelance broadcast tech and strategy consultant, geek at heart, ideas for things that are more community based than big companies. At the TV Festival [of which this is the fringe event] hearing about Joost, wasn’t saying anything anyone in this room would be surprised about but it was news to the TV people. Big disconnect between us here and them there, who don’t know much about tech but do know about audiences.

Technically elegant ways that, say, torrents, work doesn’t make sense for the audience.

Two ways to watch TV – either watch what’s on, or you can dereferencing a pointer, i.e. look something up and make sure you are there. Bit torrent is not that simple for people to use, it’ snot something that works well after a long day. How can we make that process easier, that would turn it from looking through a long list of sites to find the torrent, to something that’s as simple as turning it on and see.

Would love to see:
– Permanent URLs
– List of locations for individual programmes, whether TV schedule, bit torrent, iPlayer, and gives the data as to what DRM there is on it, what sort of format it’s in.
– Wants that info to be flexibly improved, so if broadcaster wants to say “I have the definitive information” that it references the canonical.
– Wants the metadata to be simple, and standardised.

TV Anytime is comprehensive, but difficult to use.

Broadcasters should, ideally, be providing comprehensive information. But some broadcasters have different unique identifiers, e.g. the BBC has three for each programme. But a broadcaster might tell you the metadata but would never tell you where the torrent was. Community could step in and do this.

Need to:
– create a standard extensible format
– with an API
– data licensed liberally
– crowd sourced improvements

If this data was better, could make better clients, that could give you all the official locations, times etc. but would also give you all the other locations, and tie them together with a single URL. So people who have seen a programme could send a URL to someone who could then choose how they wanted to watch it, whether on BT, or iPlayer or old-fashioned TV.

Would be interesting then to gather information on how people like to access programmes, so you could see if they prefer to watch TV or use iPlayer or BT.

Risk with current systems is that you only ever get, say, the link to the RSS feed of Heroes.

Q: Broadcasters don’t see it on their interests, because the first thing that people do is tag where the adverts are and cut it out. And broadcasters don’t want to do anything that makes it easier. From our point of view, an extra person who watches it is an extra person, but they see it as a person that they couldn’t make money from.

CJ: Agree, but can do all sorts of other things.

Q: But this is the same as the Freeview programme scheduler.

CJ: What I’m saying is, why don’t we take that info, plus the torrent sites, and iPlayer, and put it all together.

Q: BBC say that “It’s illegal to do this”, but they have never prosecuted, and never will prosecute, but it’s illegal. The problem is that it’s technically possible, and no one has ever been prosecuted, so until the broadcasters either have a day in court and see whether it is illegal, no system will have any support from the BBC or any other broadcasters. EPG data is copyrights, sharing a programme onto torrent is illegal, so no one has been prosecuted. PACT, who represent non-BBC producers, and say “This is out content, so the BBC can only show it once and that’s all they can do”, and we all have a right to record and store on VHS, but transfer it over hte net and PACT say it’s illegal. So it’s not technical it’s a lawyer.

CJ: But there’s a distinction between content and metadata. My understanding is that you can republish the BBC metadata if it’s non-commercial, and have only been threatened by ITV.

Q: There are all these legal arguments, so why do have to bring them together as a service, because that creates a legal target for litigation. How about a client that pulls together different sources and presents it, differentiating the sources, and lets people choose.

CJ: Yes, we shouldn’t keep it all in one place, but we should have a standard.

Q: So what we need is a common identifier for each programme.

CJ: Or multiple identifiers that are cross-linked. But yes, the identifier.

Q: So you could do it the barcode way, there isn’t a global organisation that organises barcodes, so that would be an easily distributable system.

CJ: i presume the names are URLs. But there are a whole bunch of existing systems, and we should be able to make it better. TVAnytime has programme groups (series), programmes, and segments of programmes, and programme locations (like a URI). If the data format addressed these types of ID (possibly except programme segments), should be able to take the URI, and use that to reverse look-up to get the metadata, and then pass around the URL that describes a specific programme, and then others can use that URL to find the programme itself. Not the only way of doing it that, but doesn’t seem to need permission, or to modify streams, etc. If we did this it might help the broadcasters change their minds.

Q: Are there parallels with the music industry and iTunes. Do we instinctively favour solutions that are too complex.

CJ: This is like an equivalent of MusicBrainz, but with links to all the places you can get the programme, not just a link to one source – Amazon in the case of MusicBrainz.

Technorati Tags:

Comments are closed.