What Ouya Isn't

> You make it sound like developers need to develop specifically for OUYA, many Android games can be very easily ported to the OUYA platform.

This has been soundly disproven by anyone who's used XNA to try to port between the 360 and WP7. Most non-trivial games need to be significantly re-thought to be comfortable on both touch and controller-based systems. The interfaces are so hilariously divergent that, no, it's not a simple port at all.

At least, if you care about good games. (Many XBLIG developers don't either. That's a ghetto too, like I expect OUYA to be.)

Initially I was excited for Ouya. Not because I was ever going to own one (console/tv gaming sucks for my lifestyle) but because it would mean there would be better games on Android.

Except now I'm not so sure.

Making a game work on a touchscreen 30cm from your face is a completely different proposition from making a game on a controller 4m in front of your face. It's nice that the underlying OS is the same, but it's not that nice.

It's not a case of just adding controller support-- your entire game changes. Fruit Ninja works on touchscreens, it doesn't work on controller. Street Fighter works on controllers, it doesn't work on touch screens. And FPS works well on neither (go go mouse + keyboard).

It's not just the controller either: playing games on a couch in your living room has different motivations to playing a game on a smartphone. Shallow 'toilet games' make sense on smartphones, they do not make sense on consoles. Deep 1hr+ strategy games, or games with consistent network access etc, make sense on consoles, they don't make sense on phones.

So I think one of two things will happen: games will either heavily target one platform or the other, and have either no support or horrible crippled support for other control schemes and mechanics, or games will genericise to the point where your controls and 'motive' is less important.

I'm not a big fan of either result.

Note: Console vs. PC is a good case study. They've had years to get this right, and there are still lots of horrible console ports. I'm not talking about bugs or graphic quality here either, but stuff like the controls on PC being awful, the UI being targeted toward consoles (Play Skyrim of Oblivion to see what I mean), hilarious console-focused messages about not turning off my computer while the game is saving, etc. If Bethesda can't spend the money getting two UIs right I can't imagine an indie dev being able to.

There is a case to be made that mindshare and user culture matters more to long-term success than the actual product, assuming the product meets a minimum threshold of quality. (See: the Mac user-base from around 1995-2005).

If the OUYA succeeds, which admittedly looks like a long shot for all the reasons described here and elsewhere, it will be due to capturing hearts and minds of devs and power users for the years it will take to gestate into a profitable platform, and not because it's the best value prop for either users or devs on day 1.

Author here. I've updated my post to include some words on the current Android games market and how none of it maps to the Ouya. Simply, a game written for a 4.5" diagonal phone screen will not scale down to the 2-3" diagonal touchpad on the controller and then back up to the 52" diagonal of a TV. You have lost too much fidelity in the process; the game will feel wrong if it even works at all.

As for Android being the future of gaming - I don't believe it. Why might a motley collection of devices with disparate specs upend a highly focused and adaptable juggernaut, besides it making a great story? This won't kill the current gaming juggernauts anymore than streaming content killed television ones.

I like the post overall, but the tech analysis is really weak. It makes it look like you spent a few minutes reading specs on wikipedia and decided that was enough to compare and contrast the hardware. It's not, and that comparison doesn't really add anything to the post.

To properly compare the processors you need to note that they are using different instruction sets and that the 360's processors are in-order with hyperthreading. Are the ouya's cpu cores in-order or out of order? do they have hyperthreading? What's the memory latency like? How big are the caches?

To properly compare the GPUs you need to understand the major differences in architecture. The 360 didn't have '512MB of memory and 10MB of video memory'; it had 512MB of memory that was shared between CPU and GPU (which means extremely cheap direct access to memory used by the GPU - something with no analog in modern PC architectures) and then 10MB of extremely, extremely high-speed framebuffer EDRAM on the GPU. These two unusual design decisions meant that overdraw was nearly free on the 360 (because the framebuffer memory was so fast) and that you could use the GPU to help out with CPU computations or have the CPU help out with rendering because both could freely access each other's memory. The Ouya could have double the clock speed and memory of the 360 and still fail to run 360 games if it has no equivalent for those features, because if you have a GPU/CPU memory split, you can end up needing two copies (system memory and GPU memory) of data, and it becomes much more difficult to have the GPU and CPU assist each other.

Someone who's done development for the 360 with the native dev kit could probably provide more detail here, I've only used the XNA dev tools (so GPU access, but no native CPU access) - IIRC there are some other perks the 360 has like a custom vector instruction set that might also give it an advantage over similarly-clocked competitors.