Eventual goal:

• Provide a democratised narrowcast-capable media marketplace
• Content producers must be paid (likely in proportion to the popularity of their work, though a patronage system or the like may also be present to encourage unpopular expression)
• As much as possible, no choke points should be built in to allow control over the forms of expression
• As much as possible, the system should run automatically, preferably without any control from a centralised authority
  
• Content is to be made available as widely as possible to provide the largest audience for those expressing themselves
• This may require getting the current players to play ball with you (cable companies and the like)
  
• You eventually want to support "traditional" content (i.e. professionally produced tv shows/movies/music/books)
  
• Which requires bringing the content producers on board in large numbers
• Which means you need money sufficient to support their budgets
  

Okay, so given that, what could such a system look like:

• You need a way to publish content (potentially very large content with large numbers of downloads) in an efficient way such that a garage band can host their album or video with only minimal (or no) investment
• There are lots of such mechanisms, almost any P2P system would serve as the infrastructure
• You need a format that allows an editor/critic to form playlists of content that a user can trust (channels sell a trust relationship with the viewer (and with the cable company, wrt liability))
  
• You need a way for users to subscribe to a playlist
• Paying the editor *and* the artists according to satisfaction with the content
• You need a way to tally popularity (people will want this for selling in-show advertising, for choosing shows etceteras)
• You may want a way to process the profile to generate similarity and suggestion prompts (both at the track and the playlist level (think Tivo Suggests))
• You may want to allow for playlists which are paid for by advertising content rather than by direct fees
  
• You *may* want to allow for exclusive deals (which is how channels get you to buy them)?
• If so, you'll have some significant changes required
• If not, you need to deal with the obvious desire, potentially by providing automated selection channels so that *popular* shows are automatically available and the editors are given an incentive to find *new* and innovative content (that is, they become the cool-seekers, not just yahoos choosing the 12 most popular shows)  Obviously you'll need the way to process the profile information to make this work.
• You need an inexpensive advertising method for channels to attract users, effectively a search mechanism, a browsing mechanism, and possibly an automated or meta-channel suggestion approach
• i.e. you can browse a meta-channel and see advertising clips from the various editors about the kind of shows they program, their quality controls etc.
• Your resellers then can provide these mechanisms to their customers, allowing the customers to subscribe to the channel for whatever period (likely with trial periods and the like)
• You need a way for meta-editors to moderate channels down e.g. for not conforming to their own standards

Now, almost all of that code is already built for your slashdot or similar system.  What's missing is the commerce aspects, you need a way to inject the end-user's money into this thing in order to pay the content creators.  Okay, so basically you need a micro-payment system that can recompense an individual who has posted a piece of content in a wide-open manner.  To make that work, it needs to be so convenient to use that getting around it is more annoying that just letting the charge go through.

How would you do it?

Okay, first approach would be to try to create a gated environment instead of an open one (you have to pay to get at the content), but then fair-use goes to heck (because to make it work you'd need some form of DRM), and you're in the position of trying to control everyone... you get a choke-point that makes you capable of censoring.  Everything's centralised, so payment setup is easy, but again, you've lost the whole "democratised" aspect, so you still get "labels" able to squeeze creators for access to the system.

Second approach, gate the pay-content only.  You're creating a ghetto for the free content, the channels may like picking it up because it costs their customers less, but the professional creators won't flock to it immediately (they're working with the traditional media) so your entire system has the potential to become a ghetto.  Note, however, that being a "virtuous ghetto" could actually be a selling point (think the better independent film festivals).  If your meritocracy really does discover gems in the rough quite frequently you could succeed in your goals of democratising the media; acting as a springboard to professional status, though that doesn't meet the "supplant the traditional (distribution) channels (networks)" goal.

Third approach, keep all of the content openly available and work on an honour system.  Problem is that you have about 1/2 of the populace whose philosophy says that they can and should do whatever they can get away with.  To make them play along you'd have to make it annoying enough to do an end-run that they'd prefer to just pay.  (Yes, I realise that just came full circle).  You could try to make enough just on the "virtuous" people, but then the cost is, almost by definition, almost 2x what it would be with a system where everyone pays (i.e. assuming same level of compensation to the creators).

Fourth approach would be adopting a slightly more "leftist" approach, namely eliminating the *individual* paying for the content, and instead merely *track* popularity and editorial popularity, having a government(s), or some other group with the ability to collect "taxes" from everyone (other candidates being ISPs), pay for the content en-masse (similar to the copyright levy on blank media in Canada).  This reduces the incentive to work around the system on the part of the individual to virtually nil (they're going to pay the same amount regardless).  It would be a *huge* political issue, but if you start the system as a "virtuous ghetto", you may find many governments willing to support the project as a cultural "good".

Which suggests the fifth approach; tokens or the like which can be presented by the individual saying, effectively "I've paid for this month's access (to recognised group X)", with signature by the user's key of the group's key for them for each piece you want to download.  Group X is then obligated only to pay, according to the popularity of a given track, that amount they collected.  You can build some fairly straightforward checks and balances so that on demand Group X can ask for verification that their users really did access content from the producers (those signed requests by the users), both producers and payment clearing groups would be able to deny each other if there's a problem (creating a marketplace for the payment handling groups).

Again, the key here is that anyone who has paid has little incentive to try to get around the system (their money has been paid, they may care in the abstract that it get to those they like, but they have no incentive not to let something they want to watch not get paid).  The Groups have little incentive to play favourites among content producers (all of their money is cleared out (minus whatever fees they collect)), they do want to make sure that their customer's content continues to be produced, and they want to keep their relations with the content producer's brokers (we'd have to assume the actual technology would be handled by someone else for the garage band, likely by two or three clearers per band).  The content producers *do* have an incentive to falsely claim that they've been viewed by the consumer, which is why you've had the consumer sign a chit saying they've accessed the content and provided a way for the Group to verify the existence of those chits.  The group *would* have an incentive to tell each content creator that they were a smaller fraction of the total (keeping the remainder), which is why you'd likely want a trusted third party collecting chits from all content creators and reporting to all involved on the final counts and respective amounts.

Now, you'll note that the fifth approach is actually just the first approach, and winds up with a fully-gated environment, except that now you have a signature from the content producer saying effectively "I'm satisfied that they've paid", so your software can share that certificate with the other embedded bittorrent clients.  Those who have contributed have a vested interest in *not* sharing the content with those who haven't, so they will likely run a client that checks for those certificates (otherwise they have to foot the bill for the production of the content for all those who share the content).

You still have lots of failure cases there.  Most obvious is the question of someone who uploads someone else's content without a price attached, or worse, with a slightly lower price (i.e. they step in to collect the real producer's money for a popular show).  That's where you would have the channels stepping in, but again, you have a failure case where someone can set up a very popular "pirate" channel. However, again, for the legitimate user, there's no advantage going to the pirate channel (their investment is fixed, and they want the content to continue).

So what about the blaggards?  The ones not willing to pay anything, or rather, willing to set up a "pirate" operation that lets others view the content they are pulling from the network (i.e. one person who pulls the content from the network for thousands or millions to share without paying anything at all to the user)?

You *don't* want to make it impossible to create channels of all-free content (that's where you're growing the network from).  You don't want to have networks have central registration.  What you need is a way to *police* the network for the bootleg copies, some way to discourage those who are using without contributing.  Obviously you could go for a legal operation (target those running the channel), but if you make it possible to target someone running a free channel (it's always possible to target someone running a for-pay channel, you just follow the money trail), you've just made it possible to target someone running a free political channel.  So, how do you discourage them (other than denying them access to the population running in payment-required mode)?

If the "pirate" network gets too large, the whole system collapses (not enough money paid to the producers to keep them producing, and beyond a certain level people who would normally pay for inertial[1] reasons start to see the increased fees to compensate for the piracy and switch to pirating themselves).  So the pirates *should* (but probably won't, being immediately selfish/shortsighted almost by definition) be trying to limit their own numbers, save that they want enough people in the network to make the P2P downloading fast.

How do you combat the threat?  Well, I think it comes back to that same issue: you have to make it more bother to be part of the pirate network.  You have to make it so convenient and inexpensive to belong to the legitimate network that the bulk of the populace will simply choose the legitimate network because it pays the artist without creating an undue burden on the user.

Obviously this doesn't even begin to address the question of determining fair-market value for content (i.e. preventing the case where the content producer makes a 2 second spot with an advertisted cost of $10,000 then signs into a collective and views it themselves, resulting in most of the fees for the collective being passed directly to them (because it outweighs all of the reasonably-priced content)).  Maybe you have to have it all be single-price-or-free, but then how do you distinguish between the 15s short and the 3h feature film?  Maybe there's a set of blocks, free, low, and medium, where the distribution is automatic, but a higher-cost area (think Pay-Per-View) adds to the user's tab, negotiated between the user's clearing group and the source wrt getting the certificates signed (avoid the 'I would *never* watch "Hot Sheep Girls"' effect).

Which, I think, is where I'll stop rambling on this subject, I don't think we've solved the problem yet.

[1] Note quite correct usage, what I'm trying to imply is those who do the "right thing" because it's what they've been taught and its easy enough to do, but who are not actually thinking through the ethical ramifications of their actions.