Audio's been the canary-in-the-coal-mine of digital interactivity, with digital video riding its coattails, and it's no wonder because, done right, video requires lots more digits than audio.
The winding course of the broadband river to the home is straightening bionomically, as Michael Rothschild might put it (http://www.bionomics.orgwww.bionomics.org). What the entertainment ecosystem demands, the network flaura and fauna will supply. We are the beneficiaries of a rising tide of digitization, a bionomic flood, and the global competition for our attention (and wallets).
The transition from acoustic to electric was far more savage than that from electric to digital. Rights holders eventually figure out the customer is always right (whether or not we agree with the customer, no copyright law can force open a wallet), and here too money will find a more efficient path between artist and fan regardless of the rightholder reluctance.
Ultimately, interactivity will be direct to centralized servers. In the meantime, buffers and caches will cover for network inefficiency, whether controlled by the network or the end-user, as with smart buffers like Tivo (now sporting USB ports) or Replay (now shipping with Ethernet built-in).
Several factors need to come into play before VOD can find its market. You need sufficient bandwidth and universality - at least 1Mbps, which that is well beyond what's available today at the consumer level. But then again, consumers may trade quality for convenience.
You need a codec that can deliver a television-like picture at that data rate, but progress has been made.
You need a connection between the servers and televisions - I don't think five million people will want to watch movies on their PC screens while they're checking their email. You need services that offer attractive selection and terms. Finally, the content suppliers have to decide to sell their stuff in bit form. All of these things need to happen.
You should be able to get something like video-on-demand today, using a relatively unobtrusive caching arrangement, on cable, DSL, or even satellite. The fact that you can't - after so many years of imagining it, developing the core technologies, and experimenting with business models - leads me to expect that VOD isn't likely to happen within a decade. Progress has been nil.
I wish I were more optimistic. I watch TV while I exercise, and it annoys the hell out of me. I sit there with my clicker - going between MSNBC, Headline News, NBC, ABC, CBS - and they all show their commercials at the same time! With VOD or even today's PVRs, I wouldn't have to put up with that. But I don't expect to give up my clicker any time soon.