Annotating Podcasts?

O’Reilly’s Web 2.0 conference has landed in the midst of Adam Curry and Dave Winer‘s exploding podcasting meme. With Web 2.0 providing MP3’s of the conference proceedings, it becomes readily apparent what is missing in the podcast world: a way of representing and distributing podcast clippings and annotations.

Think about it – if everyone’s recording content all the time, especially conference content, there comes a point of saturation. There’s a theoretical limit to the amount of data you can enjoy. Enabling audio syndication through RSS and iPodder doesn’t just open a flow of information – it opens a fire hose. Nobody has that much time on their hands: RSS, regular web sites, newspapers, radio, magazines, books, conversations, yadda yadda yadda, and now you want to add audio to the mix? And then video (which is the inevitable next step, given the impending release of cheap personal video players)? Something’s got to give. If there’s going to be an abundance of readily available content, there going to need to be an easier way to navigate it.

I envision a hybrid solution to the problem:

  1. One part of the solution is already provided by the existing podcasting solution: RSS feeds use enclosures to provide media files either directly or via BitTorrent to subscribers. This gives the subscribers the raw source material to work with.
  2. The second part of the solution is an annotation format to mark up the source file – a way bloggers to create a simple blog entry that identifies the source of the media file on which they are commenting, the section of the media file on which they are commenting (start time, stop time, duration, etc.), and the comment itself. Of course, these annotations would be distributed via RSS syndication.
  3. The final part of the solution would require an additional user interface element on the iPod (or other media player) itself – a way for the user to peruse the annotations and for a particular source file and jump immediately to the point in the source media to which the annotation is related. This may be the trickiest part to achieve – then again, it might be achievable in other ways. Perhaps the podcast aggregator could divvy up the media files according to the comments it had aggregated for a given media file, embed the comments in the media file’s metadata, and only load the clippings onto the device?

Doing this would eliminate the need for users to listen to an entire audio blog entry in order to get the information they want. It would also put the “conversation” back into podcasting – otherwise, what is podcasting, but radio (a broadcast, one-way media) in disguise? With this kind of annotation system, the user experience would start to resemble the real world a little more – it would be akin to saying, “Hey, I know you caught the latest Gillmor Gang, and you know what I think? I really liked what Dan was saying about X, but in addition I feel <insert commentary>”.

So, who’s going to build it? Or does it already exist?