The tests at GDC were done with dedicated machines in the building (despite what they claimed about external connections, such things are not allowed at GDC in that form) and so CAN work.

This DOES have a future in local places like hotels. In the outside world, no.

Bruce – yes, online multiplayer can work even up to 300ms or so because a) that’s not a lag to your input and b) the computer can predict the paths of your opponents usually quite well.

Here it can’t predict or buffer inputs, it’s only playing a game and trying to stream a 5Mbits video back to you without buffering or the ability to drop a frame. Even youtube with massive, massive servers and video less than 1/10th that size has to buffer ahead and doesn’t always work.

The final killers for the real world are that, if they really have generated encoding this magic, they’d make a billion times more selling it to places like YouTube and even remote trading floors than they ever will here. The price won’t work either, they have to pay for the computers, and the computers to encode, and the massive bandwidth required out of subscriptions, they won’t be price competitive with consoles.

Nice idea, will work nicely in hotels, won’t work worldwide. Even if you ignore the speed of light (which even if everything was perfect is 100ms of lag between here and New York.

“It’s a very ignorant article,” said Mr Perlman, who said Eurogamer had conflated issues of frame rate and latency.

“They are independent factors,” he said.

FYI, the Eurogamer article :

http://www.eurogamer.net/articles/gdc-why-onlive-cant-possibly-work-article

We can’t tell for sure that onlive won’t work, but the article is interesting nevertheless and, at least, seems documented and argumented enough.

We’ll see soon enough if onlive is the revolution which is announced !

“We have nine of the largest game publishers in world signed up.

“They have spent several years in some cases actually going and reviewing our technology before allowing us to associate with their company names and allowing us to have access to their first-tier franchises.”

Some say OnLive aims at casual gaming, notably to answer the question of lag and else on high end gaming.

But unless I missed som’thing, there are elements to consider:

1. Many casual games are not Solitaire. They are REAL TIME applications.
There, no matter the power requirements, lag happens.

2. I wonder how you’ll convince mom and dad that they should play their casual game at a low resolution on their super expensive HD TV because there are “packets being lost”…

3. How are you going to explain to your average casual gamer what lag is?
Why would they care? They’re used to their PC, Wii and DS game. Eventually, you may tell them that it’s the future, and see that mobile phone game that seems to lag simply because it’s not powerful enough? Well, kinda the same here, there will be lag, regardless of the CPU/GPU requirements.
I really think mom is going to like the idea of having her e-pees take intermittent naps and I surely don’t want to be the one to explain her this stuff, no thanks.

By the way, how the hell can this stop piracy pray tell?
Sure, you cannot pirate THEIR system, but what about the current way, you know, buy a machine of $600-800 that does plenty of things, and go torrent some?

The alternative is: the ultimate form of compression, by sending textures and meshes over the client just once, and rendering it there. Sure it needs horsepower, but GPU power has been going up MUCH faster than bandwidth got faster. This is the crux of the matter: to make OnLive work, it needs massive technical progress on the bandwidth front, but in the meantime notebooks will continue to evolve, until in X years your “thin client” runs circles around today’s HD4870 / GF9.

If it competes at all, it’ll likely be with something that’s not 3D sourced 7th guest meets wing commander cutscenes or somesuch? The sims with prepared live segments?

These OnLive people are obviously idiots who’ve never achieved anything like the monumental feats of these anonymous message board posters.

I am of the opinion that this form of delivery WILL work – eventually.

Who knows if OnLive are a little (or a lot) too early, but the idea of a Cloud-based service which handles processing, input and relays an image back to the player feels like the next stage of evolution for this industry.

Just think, these guys may have removed a massive barrier to entry (the $150-$400 console purchase). All they need is to get their technology bundled with a few TV manufacturers and things will get really interesting.

(assuming it all works, of course)

The above + extra unknown lag = too much delay in response and not much you can do to hide it.

There are smoke and mirrors that you can use but it limits the kinds of games that will be able to get away with it.

This is how selling snakeoil works. Make it sound plausible to the layperson by making vague (faulty) comparisons to existing tech. Have a physical prototype to show off, and big-name publishers giving their “backing” (i.e. accepting free money).

I’ve seen a few greedstruck business journalists get excited about this. I’ve seen far more developers comment that it faces insurmountable problems (the main one being latency) that will prevent it from ever working as described.

Digital Foundry (probably the most authoritative source on such matters) are going to be giving their view soon – I expect it won’t be positive.