I'm going to make a prediction here that many of you won't like, but it's how I think things are going. To give some background, consider that back in around the year 2000, most people still had corded wireline telephones. Cell phones were just starting to become affordable to the masses, and VoIP wouldn't start to become a thing for another year or two. In short, for most people, the idea of having a home without a wireline phone was still unthinkable. Nowadays your grandparents may still have a wireline phone, but even if they do there's a good chance it's connected to a cable phone service or VoIP device, not to the old wireline phone company. Still, there are a small but significant number of people who continue to hang onto wireline service, while their kids and grandkids snicker at them behind their backs. And that transformation took place over the span of 10-15 years.
I think it's going to be similar with TV. My belief is that traditional over the air television is going away, for a simple reason: More and more people are refusing to watch "appointment TV". People want to watch when they want to watch, not when a network or a local TV station decides they should watch. And in some markets where the local TV stations use excessive graphics overlays during network shows (weather radar and similar crap), people just want to be able to see their shows without the annoying overlays. So my belief is that over the next few years, TV as we know it is going away, and it will be replaced by devices that let you watch the shows you want to watch on demand. The other major move is that people more and more resent paying for a bundle of channels they don't care about to get maybe a few shows a week that they really want to watch. If I watch only a few shows a week, I don't need 200 channels, or 20. I just need access to those particular shows.
Most diginets are kind of the antithesis of what people really want. They are appointment television, they are loaded with commercials, and in most cases their shows are standard def. Nostalgia can only carry a network so far, and while grandmas and grandpas with failing eyesight might not care as much about seeing shows in high def, I think younger people certainly do, and will. And then there are older people like me, I'm technically a senior citizen but I still hate watching anything in standard def and I loathe 4x3 content that is "stretched" to 16x9. So my feeling is that if they are going to survive, they will need to transform themselves into a service that offers shows digitally.
The great unknown in all this, at least in the United States, is what kind of Internet access people are going to have a few years from now. If the greedy broadband providers keep trying to impose overly restrictive usage caps, or start to price broadband out of reach for lower-income Americans, it will delay the migration to viewing everything online, because nothing beats free. I can record my shows from an antenna or dish, skip the commercials, and watch when I want without paying a dime, provided they're on one of the channels I can receive.
But, what I have thought for a while now is that some smart company ought to buy a crapload of satellite capacity on a strong Ku satellite (so you can use a smaller dish) and start sending programs as data. The way I envision it, it would work like this. You buy or build a satellite/data receiver and hook it to a Ku-band satellite dish pointed at a specific spot in the sky, and also to an Internet connection. You then request in advance of the normal airing time specific shows you want to see, using an online app or the receiver itself, and that information is transmitted (via the Internet) back to the provider. At or shortly before the normal airing time, the company sends a data stream containing the entire show, compressed using RAR format using parity checking. If any part of the transmission is corrupted by bad or missing data, the receiver requests (via the Internet) as many PAR files as are necessary to fix the compressed file*. At the appointed time the receiver uncompresses the RAR file and makes the program available to watch. It contains short (no more that 20 seconds) ad spots, so that for most viewers it will be more trouble to find the remote and skip the ad than just let it play, although that could certainly be done.
The receiver would optionally try to learn your preferences and record shows you might like, even if you forget to schedule them for recording. There would also be a way to play a limited amount of "live" content in near real time (mainly news and sports) which would be similar to the way TV is transmitted today. The advantages are that very little Internet data is used (only to request PAR files for fixing momentary signal dropouts) and if no one schedules a show, then they don't need to transmit it (and they know what's not popular). If someone forgets to record a show, and the receiver hasn't yet "learned" that they usually do watch that show, it can still request that it be resent via satellite during a low-usage overnight period, so it will be available the next day, or immediately over the Internet if you don't have overly-restrictive usage caps. In this type of scheme, all providers could participate; they would essentially act as sources for various programs and then make arrangements to have those programs sent over the data stream. Although, it would be equally doable if a big company such as Netflix, Amazon, or Google wanted to start such a service and only offer programs they can obtain the rights to.
The only way this would work is if the receiver were fairly low cost and if most programs were free or very cheap. I realize that the program providers have to make money somehow but let's face it, sooner or later advertisers are going to wake up and realize that a whole lot of people are skipping their ads. Once that happens, I think a lot of over the air and cable services are going to see the waterfall of money become more like a dripping faucet. I think you can get away with ONE short ad during an ad break (which of course would command a premium price) but if it's too long or if there are more than one, people will skip them. You kind of have to use the old Bill Drake-style radio philosophy - get back to what people really want to hear (or see in this case) as quickly as possible, so they don't tune out. And that is something I think some of the Diginet operators won't get, and they will find themselves with an ever-shrinking audience.
Put it this way, let's say you have five long ad spots during a break and you get $10,000 a piece but no one watches them, or you pare it down to one short break of 15-20 seconds and charge $50,000, but 50% to 75% of your viewers don't bother to skip it. Which is really the better value to the advertiser?
In any case, it is my belief that local broadcasters are on borrowed time. Like the big phone companies, they will be slow to completely die out, but they are going to have to change and adapt, and they probably will only really be relevant for local programming in several years. People are not going to watch network shows (assuming the networks even still exist) over local TV stations. What a lot of folks really hate is local stations and cable broadcasters messing with the show in any way, from pre-empting it for some kind of local or paid content to those stupid graphic overlays I mentioned earlier. I even hate it when you go to watch something on a network web site and they display the local station's logo in the corner - all I can think is, "why, exactly, do we still need local stations for non-local content, and why are they messing with the network video?"
I didn't mean to write a book here, but the bottom line here is that I think 15 years from now the way people view video will be considerably different than the way they do it today. Devices like the Roku are the precursor but I suspect we will see even more developments in technology before the dust settles.
* If you don't know what RAR and PAR files are, you probably never downloaded any big files from Usenet back in the day. See this Wikipedia article:
https://en.wikipedia.org/wiki/Parchive - the beauty of the system was that for each damaged "chunk" of data, you only needed a similar number of PAR files to fix the problem, and they could be ANY of the PAR files in the set. So if, for example, you had five bad "chunks" of data, you would just grab PAR files equating to five or more chunks, didn't matter which PAR files as long as you had the at least the same number of chunks as you had bad chunks in the original. I never understood how that worked; it always seemed a little like black magic to me. But for example if you sent programs as data streams, you could then also send a few of the PAR files afterward and using the proper software, reconstruct the original data stream even if you had a small number of transmission errors.