The Death of ARM? Don't Bet on It

[Jason Dunn]

<div class='os_post_top_link'><a href='http://news.designtechnica.com/featured_article29.html' target='_blank'>http://news.designtechnica.com/feat..._article29.html</a><br /><br /></div><i>"One interesting announcement was a .5 watt part that is x86 based. There was no mention of ARM on stage; this suggests Intel is finally responding with a similar strategy to the x86 everywhere initiatives from AMD and VIA. While this doesn’t bode well for ARM in the long term, it does suggest that Intel is moving back to a strategy that they should have been leading since the beginning. While they didn’t showcase any new products using this part, it is in development; you could imagine it wherever ARM currently is, including phones, hand-held computers, portable media players, and navigation systems. It heralds a new age of portable devices that are much more similar to their PC counterparts than they have been; that change should result in lower prices for products that are easier to use. Neither is a bad thing."</i><br /><br />Rob Enderle <a href="http://news.designtechnica.com/featured_article29.html">wrote an interesting piece</a> for DesignTechnica a while back - it covers the Intel's Developer Forum, where Intel is talking about their processor strategy and next-gen parts. Most of the article focuses on Intel's dual-core designs and their new Viiv media vision, but the above quote certainly applies to us in the Pocket PC and Smartphone world. Is ARM really going to fade away? Will future devices be running a slimmed down version of Windows Vista? Pundits have claimed exactly that for several years now - that Windows Mobile has a limited future because everyone wants "real" Windows - but I'm not so sure.<br /><br />When you think about the strength of any platform, for most consumers it's the applications. To developers and geeks, it might be the underlying architecture and APIs, but for most people it's "What can I do with it?" and "Will it run this application?". Let's suppose that two years from now Intel has developed an x86-based CPU that can give us the same performance and power ratio as the ARM CPU powering most of the devices today. I think that's the easy part - the hard part is the software. <br /><br />Let's suppose Microsoft were to somehow hack away at Windows Vista and produce a version that would fit in a reasonably sized Flash ROM chip. So you have the x86 hardware and the x86 operating system. What about the applications? Just because your favourite accounting software is now capable of running on the CPU in your handheld or phone, doesn't mean it will actually work or be an enjoyable experience. Application design for small-screen devices is dramatically different from desktop application design, as any Windows Mobile developer will tell you.<br /><br />Will mainstream x86 software developers be willing to make the investment in designing small-screen user interfaces for their applications? Not unless there are tens of millions of devices out there, and a significant amount of those users want to run desktop accounting software on their phone. And will consumers buy these devices by the tens of millions until they can run their applications on them? Somehow, I don't think so.

[surur]

I would expect it would be pretty easy to bootstrap a new windows platform by running windows mobile software under emulator. The WM5 emulator runs quite nicely (and uses 178Mb ram to do so :eek:) and I'm sure it would run even better if built into the OS. I have no doubts handheld software can be made to run well on a small pc platform. What I do worry about though is the ability of then contemporary software to run on such a device. I dont really see Word Vista edition running on a device with less than 512Ram and 100GB hard drive. I feel the advantages of running desktop software is negated by the realities of running on a device with restricted resources (battery, screen size, size of device etc). This by itself will limit the market for these kind of devices.

With WM moving to .net CF, modern software couild easily be recompiled to work on WM. The only difference is that the IDE would prompt them to think of the actual environment the software will work in, and reduce unnecessary elements. This wont happen when we try and run software designed for 1920x1024 on handheld devices.

We can not currently predict what the future will bring, but we will never see a mobile phone running vista. And arnt mobile phones supposed to take over the desktop?

Surur

[cubed]

Rob Enderle's article stated "that change should result in lower prices for products that are easier to use. Neither is a bad thing."

I don't think that the processor is the main factor driving costs...I think it is the supply and demand that is driving costs. Mobility is always at a premium and therefore there are handhelds that cost more than desktops.

I also disagree that the devices will be significantly "easier to use" for the casual user simply because of the processor or a Vista-like OS.

For instance, give a Tablet PC and a Pocket PC to a casual user. Which do you think they will think is easier to use? In my experience it has been a Pocket PC. In most of our experiences, the users believe the only advantage of Tablet PC apps over Pocker PC apps is the screen size.

Another way to consider this is by running a Terminal Server session on a VGA handheld. In my humble opinion, there are some advantages and reasons for doing this, but it is difficult to do more common tasks through Terminal Server on a handheld than it is to work in apps designed specifically for a mobile form factor.

Don't get me wrong, better processors are a good thing. I'm excited to see where things will go in the future. But, I think the success of mobile devices will be to view them seperately from full-blown pc's.

[Felix Torres]

Microsoft is in the process of bootstrapping a new generation of Windows Apps around a new API set (codenamed Avalon) that will scale nicely to both Smartphones and PDAs.
In other words, by the time the new architecture is ready, the apps will be there...
...with a new six-inch interface to go along with them...
(Just as MS cooked up a whole new interface for MCE PCs and Apps, they can do the same for the x86 PDAs and SmartPhones).

Will Office 2003 run well on such a device? no.
But Office 12 or 13 probably will.
OneNote definitely will.
Ditto for full Outlook.

Just keep in mind that any product based on this tech won't be hitting the PocketPC space for another five years. And five years from now the typical PDA, insofar as traditional PDAs will then exist, will more resemble a shrunken version of last year's Tablet PC (1Ghz cpu, 20 gb HD, 512 Mb RAM) than this year's Axim 51v.

A few weeks back there was a big flap over in the console gaming sites over a quote attributed to a game developer who was asked what was the preferred game development platform; XBOX360 or PS3. The answer? x86 PCs. Because the architecture is well known and the development tools are readily available and familiar and you don't have to learn new styles of coding to properly utilize the resources, the way you have to when moving to the Cell or Xenon multi-threaded architectures.

By 2010 the Avalon+Indigo WPF APIs from Vista will be second nature to developers and there will be abundant apps and applets (from the sidebar and secondary-display 'gadget' toolbox) to jumpstart any new platform.

On the other hand, we'll all still be griping over battery life. :-)

[bitbank]

The ARM processor won't be going away any time soon because of the cost. With many manufacturers each offering low cost solutions, Intel cannot compete. Intel currently makes the highest performing ARM chip (XScale), but there are other companies' ARM7 and ARM9 solutions that cost as little as $2 in quantity. The ARM instruction set is also a whole lot more powerful than the x86 at the same clock rate. X86 low-power makes sense for certain applications, but it will take a really good design/price to unseat the current dominance of ARM.

L.B.

[glaggle]

It's been a while, but when I last built an application for a Windows Mobile device, it compiled as both an x86 version and as an ARM version. The x86 version ran on the emulator and allowed me to debug the application. Wouldn't it be rather simple to use an x86 based chip in newer Windows Mobile devices (assuming they reach appropriate power and performance goals) and just distribute the x86 version instead of the ARM version of all those applications?

[frankenbike]

Let's suppose Microsoft were to somehow hack away at Windows Vista and produce a version that would fit in a reasonably sized Flash ROM chip.

Application design for small-screen devices is dramatically different from desktop application design, as any Windows Mobile developer will tell you.

Perhaps. But something like this drives the market. More demand in a PPC for greater sized flash in stock handhelds will mean an end to the absurdly small amount that comes in PPCs today.

And while app design is very different for mobile devices, it could easily be included in GUI design and sense which device it's being used on. Just as some web pages today do ;)

It always amazes me how people tend to fall into three categories on these issues: those who see the future as being exactly like it is right now, those who see it in the context of science fiction (which it is), and those who actually have to do the work to make the future happen.

As a consumer, I prefer the science fiction view, which is that we'll have more of everything.

I also like the idea of ending the practice of upgrading the OS being in the hands of the manufacturers, instead of the open architecture model of current PCs.

[Joff]

I certainly wouldn't want to run XP on my PPC, no thanks!

I have a laptop with XP and even putting XP into hibernate still give me unacceptable start-up time.
Also, my laptop (WinXP's fault I believe) is uncapable of reconnecting to my wireless modem/router most times. Seems like XP gets confused especially when I plug it back to the Ethernet port.... truly mobile unfriendly :evil:

The resources needed to run XP is so ludicrous. What a contrast with Linux, it has already been ported to several PPCs, a real scalable OS (nice to have a full fledge OS, not a cut-down incarnation).

On the other hand, my brave PPC never fails to connect over WiFi and it takes about 1 sec to switch on. What a contrast!

Only downside is that many Websites are unusable on a PPC but I blame website designers for making the wrong implementation choices :lol:

Until the days we get Non Volatile RAM, I can't see XP running on a PPC or a phone. Do we really need all this complex power &amp; resource angry OSes? I would feel much happier with an OS simple enough to use that doesn't need a 4GHz Quadruple-core Pentium 7 to run :wink: