The sub-$200 dollar portable pc-compatible computer is coming, and Intel thinks it can milk the 'toys-for-rich-idiots' market that so notably DOES NOT, AND WILL NOT EXIST for restricted performance PC's. In other words, Intel still does not get it. The cheapest atom still prices itself out of the ultra-cheap (and soon to be massively successful) products.

Intel will change its strategy, but not until the competition has got a good foothold. At the moment, price (of the CPU/chipset) is everything. Performance has only to be good enough to browse at the rez of these devices (other commonly required apps need less processing power than this). However, it would be nice if the coming second-gen products can decode video at their native rez. All other far less important 'power-app' requirements (like decoding hidef video to an external output) can wait til the third gen, at which time Intel will be offering a sensible price range.

Intel may well think it is making hay while the sun is shining, and possibly delaying (or killing) the sub-$200 portable PC. This is the history of this biz. The dinosaurs taking many many months to notice that conditions have changed, and new behaviour is not only desirable, but essential.

Intel, without doubt, has a process, and thus power advantage. However, if I were a manufacturer of these cheap portables, forced to use non-Intel at the mo, I'd use a removal battery, and provide two, so that one could be charging as the other was in use, allowing time-unlimited mobile use in the home, eliminating (in one sense) the power efficiency advantage of Intel's CPU's.

If you want to know how Intel still thinks, look at the previous uses of ultra compact PC-compatible micro-motherboards. Apart from normal price portables, you had all the industry controller cards, selling at the most extreme prices possible. Even VIA fell prey to this thinking, failing to get any success for its rather poor CPU's, by always attempting to milk a specialist market. VIA should ahve been ready and waiting for this emerging marketplace.

Sub-$200 portables need (powerwise)
1) flawless browsing at the coming rez of 1024x600

2)mpeg4 video decoding (simple and advanced) to the device rez using a combo of CPU and GPU acceleration

3)performance enough to handle java etc., online apps

4)errr... that's about it, the other 99% of commonly needed apps need much less CPU power than the above

You can see that other hardware specs of these ultra cheap devices are going to be more important. At home VOIP (and video) may well be a desirable added feature (although I can see this going either way). Having VGA out (and a fullsize keyboard in) would allow the device to be converted to an always ready at-a -pinch desktop (for non-big-tin PC users), but this feature tends to cost nothing anyway.

I'm a big tin owner, but happy that machine doesn't follow me around the house. However, not having net access just cos i'm sitting or standing somewhere else is idiotic. The conventional portable (even in its current 'small' form factor) is laughable for this purpose (let alone the cost). Understand that the vast majority of computer users found in the future hardly want to touch (if at all) a computer today.

Interestingly, PDA like devices in the past have covered all the required bases, but never together. Atari did a PC compatible micro-portable years back. We've had good performance with a good colour rez, and keyboard from Psion. Plenty of types have had wireless connectivity. However, only perfection was going to provide massive success, and all of the previous attempts at the PDA were far from perfect.

The PERFECT PDA is small, VERY cheap, has a proper keyboard, proper hi-rez colour display, wirelessly connects to our hubs, and is FULLY PC-compatible. In other words, a PC in our hand. Funny, but years back, and no-one in the industry would doubt such wisdom. Today, old school industry people can't see the forest for the trees.

Remember, no big industry cheese anticipated the success of the MP3 player, GPS sat nav, or the digital picture frame (usually because they had tried poor versions of these products earlier, and failed to get market success). The ultra cheap sub-portable PC will outsell any digital product yet conceived.

This pricing is ... unexpected. Since an Atom system isn't even close to current ARM systems in terms of performance per watt (even Intel's fabrication advantages still couldn't throw off the boat anchor that is x86) I was expecting Intel to bargin-basement price these systems to push x86. Maybe even sell at below cost.

Unless Intel gives heavy discounts or otherwise "aggressively markets" these systems, I can't see them taking off.

And it's not as if being x86 give you any advantages anyhow. You're not going to be able to run any "normal" flavor of Windows on these things, and at a minimum you're going to have to recompile specifically for the platform unless you want your application to be an absolute dog performance-wise. So why not compile to a better ISA?

The real reason is that Intel owns x86. With ARM dominant in this market, Intel has to pay royalties to ARM. If they can push x86 hard enough to make it dominant, they're the ones who are going to be taking money from everyone else, if they even let anyone else into the market at all.

But they're going to have to do something about the prices if they want to acheive this.