Link-addressable streams, revisited

Peter van Dijck wrote to tell me about his tool for converting the URL of a Real stream, plus start/stop times, into a link to the specified segment. A while ago, I mentioned Rich Persaud's version of the same idea, which works with Windows Media and QuickTime as well as Real. Using either of these, you can do what I did the other day -- namely, link to a segment within a video stream -- without hacking URLs and wrapper files.

As helpful as these tools are, I've come to see that the hassles they alleviate are only part of the reason why we're as yet unable to weave video effectively into blog conversations. In the case of yesterday's clip, for example, there's probably a 50-50 chance that my carefully-prepared link actually worked for you. C-SPAN's streaming setup is amazingly robust, but invariably the content that's most likely to attract links occurs at times of peak load. If I really wanted to make sure you could see that 30-second clip, I might have done better to capture it and post a downloadable version.

That, of course, would raise all sorts of questions. First of all, how? It's doable, but not easily and not (to my knowledge) with free tools. Second, in which format? Third, does fair use cover these kinds of quotations? (I think it should, and will be testing that hypothesis.)

Despite these issues, the overriding consideration may be that streams require specialized servers, whereas downloadable clips (which nowadays play progressively) do not. Downloadable clips are, of course, inherently link-addressable, and since they're short, it's not imperative to be able to point to locations within them.

What we're left with, though, is an asymmetry. Big media organizations, for now, still have the advantage over small independents, because the big organizations are more able to deploy and manage streaming infrastructure. Bloggers can link into those streams, and/or capture and post quotes from them, but can't yet easily produce streams. What we can do easily is produce short downloadable clips.

All this could change, of course, if a hypothetical video-oriented version of were to emerge. For $X per month, I'd be able to send streams from my iSight camera to this hypothetical service, which would support X concurrent viewers of the stream. Hmm.

Former URL: