Wed

May 17
2006

Nat Torkington

Nat Torkington

BBC R&D Mega-TiVo

When I visited BBC R&D labs last year after EuroOSCON, I saw an absolutely astonishing system. Macro is a few PCs with video capture cards recording a pile of domestic British channels and storing the encoded files for later playback. Tom Loosemore talked about this today at XTech so at long last the cat's out of the bag and I can talk about it. You can see it in operation here but actual show downloads are password-protected (and you can't get a password, for obvious reasons).

Macro is gorgeous. It consumes the BBC's electronic program guide to determine when shows begin and end, who's in them, etc. There was a way sexy PHP interface (written, I think, by the fantastic Phil Gyford) to the archived programs that let you subscribe to shows, actors, keywords, etc. (I don't recognize the current UI, so I think the cool social software is still inside the R&D firewall). It's social software, so you can see what your friends are watching, your friends can recommend shows, and shows can simply bubble up to your attention because a lot of your friends are watching them. When I saw it I immediately thought, "this is the way my TiVo should be."

Because there are so few interesting channels, this is the sort of system that it's perfectly possible for civilians like you and me to build. They use one machine per channel, relying on approximately 1GHz boxes. The on-the-fly capturing and encoding uses open source software. The video pipeline is built using Kamaelia, a BBC open source project that's a Python distributed pipeline system. If a dozen of us got together, each bought one PC, and installed them in a machine room, we could record all the channels we need (e.g., NBC, CBS, ABC, UPN (I gots to have my Veronica Mars), CNN, Fox, Comedy Central, Cartoon Network, HBO, ESPN, PBS, and BBC America for maximum amusement value).

The best part of Macro is that it is all legal. It's just time-shifting, a protected fair use.

And because (apparently) talking about the conferences I'm organizing is a nervous tic for me, I have to say it: I was so excited by the promise of Macro (or "BBC Archive Testbed" as it was then known) that I've asked them to speak on the subject at this year's EuroOSCON.


tags:   | comments: 4   | Sphere It
submit:

 
Previous  |  Next

0 TrackBacks

TrackBack URL for this entry: http://blogs.oreilly.com/cgi-bin/mt/mt-t.cgi/4675

Comments: 4

  Justin Mason [05.17.06 11:04 AM]

'There was a way sexy PHP interface (written, I think, by the fantastic Phil Gyford) to the archived programs that let you subscribe to shows, actors, keywords, etc.'

It's worth noting that MythTV's MythWeb interface offers a fair bit of that, iirc.; I doubt there's any relation, but if you *do* want your Tivo to look like that, you could replace it with a Mythbox ;)

  Tim O'Reilly [05.17.06 01:03 PM]

I agree with Nat. I saw Macro too, and it's awesome. I also think kamaelia is a very interesting toolkit. Would love to hear of other interesting kamaelia applications.

  Phil Gyford [05.18.06 02:51 AM]

It's very kind of you to describe me as "fantastic", but I'm a little confused about the rest of this...

I did indeed write a PHP front-end to a system called "BBC Macro" or "BBC Archive Testbed", an internal-BBC-only prototype website that was built on a database of TV programmes encoded by the clever Ludlams (who also built Promise.TV). You might have seen a demo of this system somewhere -- the site is white, yellow and purple if that helps jog your memory! This did indeed have all the social features you mention -- social networks, rating TV shows, tagging, etc.

I know nothing about this Kamaelia system and have no idea how, if at all, it's related to the work the Ludlams and I did for Tom Loosemore at the BBC.

  Michael [05.18.06 04:06 AM]

I'd agree - Phil's front end was fantastic.

The relationship between the Kamaelia based system and the original macro is that this is a replacement backend system built from the ground up, designed to provide an API for different front ends so we can experiment and see what's possible.

ie Kamaelia's role is to provide that flexible backend - the raw data & information in an accessible way. Kamaelia is designed after all from the ground up to allow systems the kind of high level flexibility desired. The current front end is just some simple blog code of mine to allow people to browse content as it comes in.

Without that back end, any front end is just pretty. Without a cool front end, the backend is pretty boring. Hopefully we'll see the re-emergence of Phil's front end, or something even more fun :)

Regarding more Kamaelia based systems, I'll try to write up something about that on the kamaelia project blog :)

Post A Comment:

 (please be patient, comments may take awhile to post)






Type the characters you see in the picture above.