Podcasts Need Fragments for Citation

There has been a lot of discussion on podcasts and blog posts in the last few weeks about the difference in audiences between podcasts and written media. Jason Snell summed it up neatly at Six Colors in "Nobody's Listening," discussing in part how ideas related to Apple's increasing difficulties in keeping its software quality high were talked about for years by John Siracusa, Marco Arment, and Casey Liss on Accidental Tech Podcast (and elsewhere) — but the same ideas only lit a match to become a worldwide firestorm when Marco wrote a blog post about it. Jason wrote:

…podcasts are not easily chunkable, shareable content. At least with YouTube videos, there’s a standard way to link to a specific time code. Podcasts like Serial can go viral, but podcast content doesn’t. Large parts of the Internet are powered by an economy of link sharing. Podcasts sit outside of that.

But there is a standard way, as Kevin Marks pointed out to me this morning. The Media Fragments URI defines precisely how to identify stretches of audio and video through a standard URI. In fact, Jason is experimenting with using such links at The Incomparable. Some embedded audio players support this link.

However, there's no per se standard way to listen to any audio or any podcast. Jason can use those links on his site, because he uses an embedded audio player on The Incomparable that supports them. Although note that when loading the page, because it doesn't auto play, even using the correct URI format doesn't advance to the audio piece in question. Also, the entire audio file loads, even if you're just listening to a piece of it, and you must wait until that piece has loaded for it to begin to play.

None of this makes it easier to take a snippet, whether a 20-second sound bite or a 15-minute topic on a podcast or news broadcast, and use it as the basis of discussion. And even when one can link to a specific piece of audio by reference, there's no way to comment on it or embed just that section on another site. Fair use might allow excerpting the snippets if short enough (even though fair use doesn't have a "length" test, it's often asserted as if it does), but many web hosts and audio sites will not allow you to assert fair use in order to share audio, no matter how well founded your assertion is.

The sharing options for podcasts on Overcast.fm pages

I could envision an opt-in clearinghouse, such as at Marco's Overcast podcast app's corresponding site. Right now, the site is adjunct to the app, allowing users to have minimal but useful options to use their subscriptions and current episode list to listen while away from the app. Marco supports the Media Fragments URI at Overcast; on any podcast page on his site (such as this one for a recent episode of Back to Work), while using Overcast's web player, you can opt to Share Link with Timestamp. Doing so allows another listener to listen starting at that point.

This is a start. Next, one would want stretches of audio identified, and the potential for users to crowdsource transcriptions on a snippet basis. Transcribed snippets, if not entire podcasts, would allow the discussion of ideas more readily. Then those snippets should also be able to be embedded, further allowing a blog or news site to reference a persistent, identifiable fragment that could be played on page—a critical aspect of robust discussion.

However, this starts to get rapidly into concerns about liability for copyright infringement, even though fair-use doctrine would almost surely prevail, because fair use encourages discussion, especially of portions of a work that don't diminish its commercial potential. But fair-use defense is always a defense: you can't affirmatively ensure fair-use rights.

And we don't just want Overcast as the sole repository of podcasts and snippets. Better than one or many commercial sites, would be the Internet Archive supporting this effort, which should be both persistent and robust. To make such a thing happen, the Internet Archive would need to feel it's relevant enough, as some changes would be required on its end to feed just snippets of audio. And podcasters might need to allow the Archive or other organizations to use their audio for this purpose if they don't already license their podcasts (as I do all mine) under a Creative Commons license. (Any CC license should suffice, since even the most restrictive, such as BY-NC-ND [attribution, non-commercial, no derivatives] would allow intact, referenced segments for discussion.)

I wonder if there's any way to move this forward, or it's just a percolating set of ideas in the ether that are too mired in the heterogeneity of audio distribution and playback? It's wonderful to have many options, as it's what has made podcasting thrive. But is there a way to make podcasts more relevant and less insular and demanding of time through better exposure of bits and pieces? I welcome discussion!

(Via Phil Windley, Jon Udell wrote about this a decade ago! And here's Kevin Marks from a year ago about citing points on a web page, rather than referencing the whole.)