In the past, I’ve made no bones here about the fact that I consider proprietary console platforms to be a sub-optimal content delivery platform for games. I understand why they’ve got to this stage (desire to seed the market with advanced, standardised tech at less than cost price, requiring lock-in to recoup later), but that doesn’t make them a desirable end-game. Closed systems are by nature market distorting, and can hamper innovation, because when only a chosen group of ‘authorised’ developers have access to deploy on it, you’re not maximising the amount of content innovation available. XNA and the CommunityGames Channel go some way to addressing that, so kudos to MS for trying, but the fact that it mandates a certain technology means that it’s something of a dichotomy - it’s great that it’s open, but it’s significantly hampered by the fact that what you develop will only work in that environment (or on Windows), when in practice a lot more people want to be able to reuse their technology across many platforms, like mobiles, the Mac etc. Also the tools and libraries for XNA might be good and expanding, but they still represent a tiny subset of the breadth and maturity you get access to when using C/C++ so I’d still consider it less desirable as a target platform. Of course, the reason for this restriction is security, which again goes back to the closed platform argument.
Personally, I most love devices where anyone can create on them, using whatever tools they like. All other media follows this pattern - films, books, music can all be created by anyone with the will to do so and will ‘play’ on a standard device in any home; the limitations associated with games are a throwback to a historical position which gets eroded every generation (look at the exclusive / cross-platform ratio this gen compared to previous iterations), and which will eventually disappear. The PC remains king of the hill in my book - but of course commercially it’s a bit tricky. The two main issues are piracy and hardware compatibility - price is sometimes listed here too, but that’s blown out of proportion; the kind of people that say you need to spend £1500 on a gaming PC are talking out of their arse, you can build a good gaming PC that will outclass the main consoles for a third of that price these days, it’ll never be as cheap as the console, but you can do a lot more with it.
OnLive tries to address these problems by running your games in a server farm (the ‘cloud’ if you will), and just having a relatively simple device at the customer end - which can be an existing PC or a small set-top box if you don’t have one. I think it’s a little early, since internet speeds are still a bit rubbish in many parts of the world (here, for example) and not up to the 5Mb/s required for HD delivery, but there’s definitely a good kernel of an idea here. Lag is of course an issue, most concerningly on the input side, but in practice the kind of information going back and forth is actually no more than what you have in a standard multiplayer game, and providing the video streaming can keep up, it might just work. I can’t see it working for very lag-sensitive games like music games, but for most other things if the QoS is acceptable it might be ok.
This deals with hardware requirements, because the set-top box is cheap, and piracy because the games are never actually on your machine anyway. I could imagine games moving more significantly towards a rental / subscription system in this kind of scenario, except where the developers get paid more directly based on play time rather than just the initial box purchase like with current rental schemes - I consider that a good thing for developers, since those who make games which people play for months or even years deserve to be rewarded for that.
The one concern I have is that the service provider still controls the content that you can see. I could imagine that once the technology is mature, other service providers would compete and hopefully content delivery would be more open. However, in practice I can see that initially service providers would want to lock down the service, putting us essentially back in the one-provider situation again. So, even if the technical hurdles can be cleared, this might not lead to open gaming nirvana just yet. A requirement would be an open development package (something like SteamWorks), where anyone could develop content that would work on the platform even if they still had to gain authorisation to be hosted there (and I could imagine the barrier to entry here would be lower anyway).
So, despite the difficulties and unknowns it’s a step in the right direction I think. The closer games get to the state of availability and openness enjoyed by other media, the better we’ll all be for it. I actually think that the next generation of platforms may be a stepping stone in this direction - I would actually not be surprised if Microsoft was looking at this area for the 360’s successor, since they have always been fans of digital delivery. I could imagine that perhaps the next iteration, probably announced in 2010/11 might be a hybrid device, capable of disc play but with a much heavier emphasis on digital delivery, and I don’t think full downloads are necessarily the way that will go for full games, because of the high up-front wait time. I think it’s more likely to be either something like OnLive, or the progressive download technology used by GameTap (where just enough of the game is downloaded to start playing, with the rest downloaded as you play in the background).
Either way, I think the ‘traditional’ console business model is reaching the end of its run, and the next 10 years will see it transform completely from the way it is now. Those who think it won’t - remember that 10 years ago the Internet was still embryonic, Amazon was still small and Google didn’t exist - 10 years is a long time. Technology just doesn’t sell games as much as it used to (see the Wii), and in that environment it makes less sense to have to subsidise cheap cutting-edge (at the time) hardware in the living room and recoup through artificial restrictions - it becomes all about the content and how you sell as much of it as you can to customers; which means re-examining your content delivery approach, going broader and opening up to allow more people to deploy content (because how will you find the next big thing otherwise?), taking away disincentives to buy (such as having to buy 3+ consoles to play everything), getting content to people quicker, and most importantly in the manner they want it. Forget what worked in the past, you won’t discover what will work best in the future by holding on to a model that was developed in the 80’s.