kafsemo.org

Mivvi is about television

2006-01-16

Television watching is shifting, from broadcast TV to DVDs and TiVo, and from DVDs to TV-on-demand and downloads. Odds are, it’ll shift again.

But here’s the thing: I just want to watch my shows. I want to say what I’m interested in, see what’s available, click some buttons and watch. I’ll wait until after broadcast time and I’ll pay for DVDs and downloads but, when the technology exists, I don’t want to have to arrange my time around arbitrary broadcast schedules, and I don’t want to suffer a wilfully standalone system.

Mivvi is a project to help integrate different ways of getting at television programmes. It’s still in development, and not quite yet my ideal system but, right now, it Does Stuff. There’s a guide to getting started that will take you through what it does now. It’ll show you upcoming UK TV schedules by series and episode (great for catching repeats with obscure scheduling), it’ll help you to manage media files (perfect for archiving and DAP shows) and, under the assumption that you’re not going to infringe copyright, it will make it easy for you to track download sites for new shows.

Subscribe to a series, or just a specific episode, and only see the schedules you want. Want to see if that season is out on DVD before waiting for the broadcast to start? The framework’s all there.

Under all this is the common concept of an episode, linked to web pages in online guides. If you’re used to listings guides with tiny, poorly-written plot synopses, a double-click to get to a full web page is a real improvement. Most of the current data uses TV.com, but there are also links to HBO’s fantastic episode guides and even wikis.

So yeah, I think it’s neat.

Technology

It’s RDF, right the way through. All the episode guides are RDF, the media management is RDF, and subscriptions are defined in terms of URIs.

A few predicates are defined by the http://mivvi.net/rdf# namespace, and everything else is Dublin Core. A strong design goal was that series data could be defined by anyone: Mivvi is simply the framework.

There’s a guide to producing your own data, as well as some samples of what happens when you start to build on that RDF investment by bringing in new metadata you’d never have included in a standalone format. Take a look in mivvi-data.zip: it’s totally view-source friendly.

The source isn’t open yet. It needs a little bit of tidying, then I intend to go for LGPL.

Thanks

Thanks to everyone who gave their feedback – it’s made for a noticably better product. Thanks also for OpenRDF’s Sesame (great response to bug reports) and Redland RDF, without which I couldn’t even attempt this.

Feedback

It’s easier to write code than it is to document it, and evangelism is even harder. If you’re interested in the possibilities, I hope you take a look and that you get something out of it.

If you’re a semantic web booster or cynic, I’m really interested in discussing the viability of desktop RDF, and about making it as accessible as possible.

There are a few specific projects that would be great to integrate with: the TV IV for data, MythTV and any content repositories (especially those based around content-addressed storage).

So please, send mail if you have any questions or suggestions.

(Music: Pavement, “Price Yeah!”)
(More from this year, or the front page? [K])