ARM, x86 Chip Makers Fight to Ride Mobile Growth

April 30th, 2008

200804292350
Daniel Eran Dilger
Apple’s acquisition of PA Semi appears to fit well into the company’s plans to pioneer the development of a new WiFi mobile platform with the iPhone and iPod Touch. Apple is certainly not the only company to see the vast potential in mobile devices. The market for smartphones and mobile Internet devices is currently broad and diverse, with lots of competition both in the hardware components used and in the operating system and development platforms offered.

Today’s growth in mobile messaging and computing devices bears some similarity with the explosion of desktop personal computing in the early 80s described in the previous segment. The difference is that today there is no big equivalent to IBM threatening to enter the market; all the existing, leading competitors in mobile devices are already large and established companies.

Unlike the 1981 IBM PC, which pushed the unremarkable x86 processor and Microsoft’s copycat software ahead of superior technology, mobile devices today are being sold on their actual merits in terms of hardware and software. Of course, that doesn’t mean there isn’t anyone working to shoehorn the square pegs of x86 processors and Microsoft Windows software into the round hole of mobile devices. Here’s a look at the state of chips in mobiles, and how PA Semi expands the options for Apple in the mobile market.

How Apple’s PA Semi Acquisition Fits Into Its Chip History
Why Did Apple Buy PA Semi?

The Glut of Chip Designs Without Buyers.
Apple isn’t the only company buying up fabless chip designers. Texas Instruments is rumored to be acquiring Wolfson Microelectronics, the Scottish company that developed most of the audio processing chips in the iPods until just recently. Apple was Wolfson’s largest customer in MP3 player chips, which made up 24% of its business. The loss of iPod-related revenue was a major hit to the company. TI already holds a minority stake in Wolfson.

Sun also just bought up Montalvo Systems, a startup run by Matt Perry, the former CEO of Transmeta. Montalvo is working to develop low power x86 compatible chips for mobile devices. Transmeta attempted to do the same thing nearly ten years ago, but its attempt to break into the x86 market dominated by Intel and AMD was stymied by an inability to find enough customers for its highly efficient, x86 compatible Crusoe processor, despite the initial flurry of excitement around it.

Both Transmeta and Montalvo worked to emulate x86 code on top of more advanced processor architectures, with Transmeta using VLIW technology common to Intel’s failed Itanium processor, and Montalvo (packed with Trasnmeta alumni and the developer of AMD’s K6 processor) seeking to develop a design similar to the fabled Cell processor, jointly created by Sony, IBM, and Toshiba.

Add in the x86 compatible Intel Atom (used in the few UMPCs sold annually) and AMD Geode (used in the OLPC XO), and you have an awful lot of products in a small space, as nobody seems to be clamoring for x86 compatibility. Incidentally, both Intel and AMD have strapped the overhead of x86 compatibility on top of their core RISC processor designs to keep up with technology advancements while dragging along the legacy of the original PC.

The real problem for mobile devices has been finding a way to make them productive enough to be marketable in serious quantity. Microsoft hasn’t been able to do that in a decade of trying with WinCE, and its parallel efforts to pack the regular desktop version of Windows XP/Vista into Tablet PC and UMPC designs hasn’t been successful either.

CES: Fear and Loathing in Las Vegas
The Spectacular Failure of WinCE and Windows Mobile
Innovation: Apple at Macworld vs Microsoft at CES

Follow the Money.
In contrast to all the speculatively designed, low power x86 mobile contenders, the vast majority (around 70%) of today’s mobile devices use ARM processors built by TI, Samsung, Marvell, and the various other ARM licensees.

In January, ARM announced the shipment milestone of ten billion ARM processors, with three billion now being shipped every year by the hundreds of ARM licensees. In comparison, UMPC devices and the XO combined didn’t sell more than a million units last year. It’s going to take more than a few miracle sales to turn the mobile tide in favor of any one of the various x86 designs, all of which are less mature and offer less flexibility in part sourcing than the wide range of ARM components available.

ARM is of course not x86 compatible, a significant problem for Microsoft considering that Windows XP/Vista is not designed with processor portability in mind. In contrast, Apple had no problem delivering its OS X operating system on the ARM-based iPhone. It’s also had no problem selling millions of the iPhones and iPod Touch units within their first year on the market.

Will Apple Rescue Intel’s Silverthorne?
ARM Achieves 10 Billion Processor Milestone
Origins: Why the iPhone is ARM, and isn’t Symbian

Apple’s Mobile Strategy.
Apple has paired its advanced Cocoa development platform and UNIX-based operating system with the ubiquitous ARM architecture in a bid to ride the vast economies of scale of that processor in the same way that Microsoft happened to be in the right place at the right time to get its clone of CP/M installed on IBM’s PC. Again, the difference is that Apple’s strategy has technical merits.

Additionally, Apple’s mobile hardware will be far more difficult to clone, in part because it develops its own software platform rather than having delegated that task to Microsoft as IBM did, but also because it is far more difficult to do the tight integration work needed to develop a functional mobile device than it is to bang together a desktop PC clone.

That makes Apple something like the new IBM in mobile computing devices, except that Apple lacks any dependence upon a third party to maintain its software platform (which in IBM’s case meant that Microsoft could license the same DOS software to competing hardware clone vendors), and it doesn’t face the cloning problem that rapidly destroyed IBM’s PC business. Rather than losing hold of the market it originated as IBM did with the PC, Apple appears poised to maintain control over its WiFi mobile platform and plot out its future with a lot of flexibility.

Apple has proven its ability to rapidly innovate and maintain extremely high popularity in consumer electronics with the iPod over the last seven years, a business naysayers repeatedly insisted would quickly fall to commodity MP3 makers. Not only has the iPod remained unchallenged as the market leader, but new models are rapidly leaving behind the simple MP3 feature set to become sophisticated handheld computers.

The company also has a successful online community and store in iTunes that has transcended the barriers of the Windows PC platform on the desktop level to become the world’s largest retailer of digital media. That infrastructure is now being devoted to deploy mobile software for a secured platform where piracy and malware will be much easier to contain and prevent.

Apple’s iPhone vs Smartphone Software Makers

iPhone 2.0 SDK: How Signing Certificates Work
Apple’s iPhone vs Smartphone Software Makers
Five Factors Shifting the Future of Malware and Platform Security

Problems for Microsoft in Mobiles.
Microsoft is currently working furiously just to strip down Windows XP to the point where it can run on small laptop devices like the XO and the Asus EEE PC, which uses a conventional low power Intel Celeron chip. For mobile devices smaller than a diminutive laptop, Windows XP is far less ideal and Vista is even worse.

In addition, the company is completely unprepared to further port its software to run on an entirely different processor architecture. Its previous efforts to deploy Windows NT cross platform failed miserably and were discontinued with Windows 2000. WinCE can run on ARM-based devices, but it isn’t compatible with Windows applications, nor is it gaining broad adoption in the market as a mobile OS after ten years of trying. Microsoft even shunned WinCE in its own UMPC plans, the very application WinCE was expressly designed to fill.

The portability of OS X (inherited from its NeXTSTEP legacy) gives Apple the opportunity to ride the economies of scale in the ARM world, possibly partnering with TI (an investor in PA Semi) or its existing ARM suppliers Marvell and Samsung to design and build new ARM-based processors that push the state of the art in mobile efficiency while Microsoft, Intel, AMD, Sun, and Transmeta struggle with maintaining the x86 legacy required to support Windows compatibility on a class of chips that have seen only minimal interest in the marketplace.

Microsoft’s Hardware Handicap.
Microsoft’s repeated failures in mobile devices, from Handheld PC to Pocket PC to Windows Mobile Smartphone to PlaysForSure and Zune have demonstrated that without the direct leverage of a monopoly position it enjoys in the PC world, the company has no special capacity to sell products, particularly if there is any existing competition in play.

Even Microsoft’s brightest star, the Xbox game console, has been a black hole in terms of profitability and a demonstration that that company’s past success in desktop software doesn’t necessarily transfer into hardware sales, even where there is some product tie in to Windows.

Ironically, the most successful applications of x86 mobile chips so far (the XO and EEE PC) have been designed to run Linux. That paints out an ugly future for Windows in mobile sales just as Microsoft’s current monopoly on the desktop is crumbles as the growth in PCs shifts toward mobiles both here and in developing countries.

Video Game Consoles 2007: Wii, PS3 and the Death of Microsoft’s Xbox 360
Why Microsoft’s Zune is Still Failing
Microsoft’s Unwinnable War on Linux and Open Source

Microsoft Behind Linux and OS X in the Battle for Mobiles.
And so the tables have turned. Microsoft’s fortuitous rise on the back of one of the least suitable processor architectures isn’t going to be duplicated, at least not by Microsoft. Mobile devices are exploding on top of the ARM architecture, and Apple’s software is on two iconic brands, the iPod and the iPhone, which are both accounting for a significant chunk of mobile devices and currently leading the market in WiFi web savvy.

The front running alternative in mobile phones outside the US is Symbian, a convoluted software platform that’s difficult to develop for and fractured among the three incompatible flavors specific to Nokia, NTT DoCoMo, and Sony Ericsson. Symbian frontrunner Nokia is itself investing in Linux as a replacement to Symbian while Sony Ericsson is dabbling with Windows Mobile to differentiate its high end phones. Those moves expose the great potential available to the iPhone’s far more modern and capable OS and development platform, which unlike Linux is already polished and mature in the advanced smartphone and handheld WiFi web browser market.

The use of Linux and free open source software is being promoted by other hardware vendors and by Google, although Apple has an edge in platform control that will relate to security, interface sophistication and ease of use, as well as the tight integration supporting mobile software sales and distribution.

While the merits of the various FOSS mobile platforms and Apple’s OS X can be debated, there will simply be no room in the mobile business for a proprietary monopolist to dominate the market with an inferior product launched by a large outside player the way IBM pushed the DOS PC. That’s good for everyone, apart from, of course, Microsoft.

Readers Write About Symbian, OS X and the iPhone
Canalys, Symbian: Apple iPhone Already Leads Windows Mobile in US
Apple’s iPhone Vs. Other Mobile Hardware Makers: 5 Revenue Engines
The Great Google gPhone Myth
Apple iPhone vs the FIC Neo1973 OpenMoko Linux Smartphone

I really like to hear from readers. Comment in the Forum or email me with your ideas.

Like reading RoughlyDrafted? Share articles with your friends, link from your blog, and subscribe to my podcast! Submit to Reddit or Slashdot, or consider making a small donation supporting this site. Thanks!

Technorati Tags: Apple, History, iPhone, iPod, Mac, Microsoft, Software

Pingback: Iphone » ARM, x86 Chip Makers Fight to Ride Mobile Growth()
Rich

“Symbian frontrunner Nokia is itself investing in Linux as a replacement to Symbian”

Do you have a source for that? I’ve only ever seen Linux running on Nokia’s web tablets. I haven’t heard any rumours of Nokia writing or porting a telephony stack.
http://www.roughlydrafted.com danieleran

Nokia is not only using Linux for its Internet Tablets, but it also just bought Trolltech, the vendor of Qt tools for Linux development. The merger PR stated:

“The acquisition of Trolltech will enable Nokia to accelerate the cross-platform software strategy for mobile devices and desktop applications, and to develop its Internet services business.”

Nokia executives speak of Symbian using less than complementary language, but Symbian is really both a decent kernel OS and a series of incompatible personalities. Nokia will either port Qt to the Symbian kernel and ween itself off its current s60 & s40 Symbian platforms, or kill s60/s40 along with the Symbian kernel and standardize on Qt on a Linux kernel. Either way, “Symbian” as it currently stands is going away.

Neither Symbian nor Qt nor Linux are in the same realm as OS X/Cocoa, as Linux is not ideal nor optimized for use in mobiles, nor are there any great development frameworks that are easy to use. Linux is ideal as a PC server, it’s approaching usefulness as a mainstream desktop OS, but its still early to be positioning it as a mobile OS, and there are a half dozen or more main competing development toolsets in the mobile phone space.

Symbian is similarly constrained from the other direction: it makes a great feature phone, an okay smartphone, but rapidly falls apart as it gets stretched into service as a platform for running desktop style apps like the iPhone’s browser and full email client.
Berend Schotanus

The article title is about ARM versus x86 competition in mobile markets, which makes an interesting promise. x86 is the PC processor standard where in the end even Apple “capitulated” for the economies of scale. The article describes how this same economies of scale apparently does not work for mobile processors where ARM prevails even when energy efficient x86 compatibles appear. But why?

Then of course Apple is able to bridge the gap between ARM and x86 while Microsoft is not. And Apple is way ahead of other competitors who still have to build processor portability upon Linux or other flavours of Unix. Sure…

But this big mystery that the title is referring to, how come two different processors have rock solid market positions in PC market and mobile market, remains unsolved.
Realtosh

Good stuff Daniel.

Thanks for the detailed research and analysis of the articles over the past several days. This combination of work has done the best job of dealing with my concern that Android will become the platform of choice for non-iPhone smartphones.

I agree with you that Symbian is on its’ way out; as it is too splintered to make an ideal development platform. Microsoft has been unable to get WindowsCE to go anywhere because of shortcomings in the technology.

So, it looks like Apple will have little real competition in the high-end smart-phone/ internet appliance market.

With Google as sponsor, Android will fill the vacuum left by all the other non-competitors, over the next 3-5 years. Android will become the platform of choice for non-Apple mobile equipment. Most other vendors will eventually standardize on Android, if only in a desperate attempt to try to catch up to the Apple advance the mobile space.

These last several articles detail the value-add that Apple brings to the table in terms of integrated design of both software and hardware. This increased value is seen by the end user as the ease of use. Your articles have shown all the work behind the scene of designing and incorporating specialized custom chips that Apple uses to help bring all the elements together and make its’ offerings stand out from the crowd. These hard to duplicate custom designs will make it harder for other manufacturers to catch up. This is exactly the kind of analysis that I was calling for earlier this year.

Nokia, and possibly others like Samsung, would have the talent and resources to try to copy Apples moves, even in this custom chip area, as they have been building phones and their innards for years. Nokia products lack the overall completeness and polish of Apple gear. They are especially weak in software and will benefit from help from Google through the gift of Android if they ever choose to adopt it across its’ smart phone lines.

BlackBerry is a great 1 trick pony, but RIM lacks the resources or talent to challenge Apple on all the fronts at which Apple excels. It is both the great number of areas in which Apple has talent and also Apple ability to bring all those different parts together into spectacular products that set Apple apart. No other company has either talent in all of the separate aspects that Apple has built up over decades (especially not even Microsoft). Plus, no other company has the skills in bringing the parts together into brilliant and humanly accessible products in the way only Apple seems to be able to do. Apple creates iconic products that are beautiful and easy to use. But under the skin, there is much that goes into the secret sauce. None but the most capable of competitors will ever have a chance to be able to compete in this mobile space that Apple will dominate. And even so, it will take years for any such competitor to build up the abilities to try to compete with Apple.

Again, thanks Dan for your valuable contributions to these analyses.
http://johnsessays.blogspot.com John Muir

@ Berend

ARM was too late to the game to succeed as a desktop processor. Just ask its originator: Acorn. It has proven however to be very well suited to mobile use as its power saving is legendary.

x86 in the form of Atom is coming very late to the mobile space. It’s an interesting development and I wouldn’t rule it out entirely in the long term, but right now even Intel aren’t pushing their chips as phone ready … yet. Still too many watts required. Short battery life is a deal breaker in mobiles.

The Time Capsule and AirPort Extreme can do heavy file lifting for hours on end without even being equipped with processor cooling fans, as they are ARM based. Nice but essentially unnoticeable by the end user. Just imagine the opposite though: a smartphone which kicks on a blower every ten minutes…

x86 brings Windows backwards compatability with it wherever it goes. In PC’s that is essential. In phones: even Microsoft have discovered that it’s not.

@ Realtosh

I can understand your enthusiasm, especially given the weight of Daniel’s arguments. I don’t think that Apple will dominate the phone market however – in the way they have with portable audio – but rather that they’ll dominate particular segments and prove a strong player overall.

Mobile phones are already firmly entrenched far more so than desktop computers worldwide. The low end dominates the market. Feature phones, simple things, bought on price and carrier plan … indeed, in many places “pay as you go” top up cards overwhelm the competition.

Apple are establishing themselves as the principal player at the top of the market as we speak. Then, as they did with the iPod mini and nano, they will surely expand down the way, in the time and manner of their choice. There’s only really so far they can go in that direction though before phones hit the equivalent of the iPod shuffle. Bringing their brand, the iTunes ecosystem and their flair for tiny designs, they may well even be able to compete down there … but sooner or later Apple’s strengths are overcome by the hard and fast price directive deep down in the depths of the trade. The iPod did well even down there because it was still an iPod, with or without a screen, and iTunes facilitated its core function. With cheap, cheap phones though: music playing is a secondary afterthought, and the vast market for simple receptacles of pre-paid call cards is already established and thriving.

It is indeed amusing that the people in the industry who were quickest to scoff at Apple’s entry were the ones easiest to beat: Palm and their like at the top. It’s the depth of Nokia’s low end business which Apple will be competing with in years to come though. And truth be told they won’t have to outsell it even to turn a higher total profit on the venture. So it’s really up to them.

I somehow doubt Apple will be the leading brand of cellphones in Kenya, Laos and Colombia. But then they scarcely need to be. Just keep it in perspective.
Urian

The other day I was thinking about this:

Imagine that you can record video in a future version of the iPhone and when you have recorded all the clip you can modify it with a custom version of iMovie in any place, after this imagine this with any type of application.

The problem is that for this change you need a powerful mobile hardware and this could be the reason why Apple has bought PA-Semi.
MikieV

@ John Muir

“There’s only really so far they can go in that direction though before phones hit the equivalent of the iPod shuffle.”

Agreed.

The old arguments about Apple’s market-share for computers, compared to BMW’s or MB’s market-share in automobiles comes to mind…

As long as the margins are good, they don’t have to dominate the mobile-phone market – percentage wise…

—

@ Dan

“That infrastructure is now being devoted to deploy mobile software for a secured platform where piracy and malware will be much easier to contain and prevent.”

Has always amazed me how iTMS is the elephant in the room that all the cheerleaders for each successive “iPod Killer” fail to make any mention of…

iTunes and iTMS are almost always dismissed as “Why do people use them? They suck…”

Don’t they understand how that trifecta is what has made the iPod “un-killable”?

And, with Apple now in the smartphone market, “they” make no mention of how Nokia, Samsung, or whoever will create a competing “ecosystem” – along with whatever mobile device & OS they put together.

—

With the “3G” iPhone supposedly already upon us, I’m curious to see if the PA Semi purchase helps Apple design chips which take advantage of the new spectrum recently actioned-off by the US Government…
kovacm

@John Muir

“The Time Capsule and AirPort Extreme can do heavy file lifting for hours on end without even being equipped with processor cooling fans, as they are ARM based.”

they are Intel based ;)

[The modern AirPorts actually use ARM processors from Marvell, as my take-apart in AppleInsider revealed. – Dan]
Realtosh

@ John

Yes, exactly what you said word from word.

Note that I did reference “high-end smart-phone/ internet appliance market” as the specific segment that Apple will dominate. So, in other words they’ll have the best phones/appliances. As such their brand will be quite valuable in the cell phone space. And as you say, over time they will have Shuffle-type cell phones; not so much in form factor but in comparable (ie lower cost) cell phone market segment.

I agree that Nokia will give them the most competition, especially at the low end. But everything has its’ limits. If the competitor (Apple) has consistently better phones and higher customer satisfaction ratings; even at the low end their will be a preference for Apple-branded phones.

Remember that it was not that long ago that Motorola was at the top of the cell phone manufacturing business. Now, they are just an afterthought. Motorola got beat in the transition to digital, with their slow response to competitive threats from Nokia and others. Now the shoe is on the other foot. Apple is the competitive threat. Furthermore, at this time, I can’t see any other company that is capable of competing effectively against Apple in the high end mobile space.

I still believe strongly that the biggest competitive threat to Apple in the long-term in the mobile space comes from Google in the form of the Android platform. At the moment there is not as much interest in Android software development, as there is in iPhone development; if only because Apple is moving much product, and as Apple will soon move much more, as version 2.0 comes out and more geographies are added. But Android is open-source, so there is not the same fear of becoming beholden to Microsoft. Plus Google is willing to shepherd Android and provide it resources and will allow Android to catch on as a platform in a way that most other open-source projects are usually unable to do effectively without such support. So, if Android does ever catch on, which I entirely expect it to do so, it would provide a platform onto which other less sophisticated manufacturers could build upon to try to compete with Apple’s complete solution.
beanie

UMPC is entering the middle of the S-curve product cycle where expansion and growth accelerate from the early adopters.

Atom is also great for fanless set-top media devices. I believe Apple TV under-clocks the 1GHZ CPU to around 350MHZ and has a fan. With the Atom, it could run at full speed, 1GHZ, and could go fanless.
addicted44

I am not as confident about Apple’s strategy. I think they have the best one, but what makes Mac vs PC in the 80’s different from iPhone vs. Android phones in 2008? Isn’t there a real concern that Google can step in and wipe the iphone out like MS/IBM did to the Mac? Assuming that the Android does pan out, any thoughts on why this will not happen?
labrats5

I’ve actually given much thought as to whether or not Apple would create a lower end feature phone to compete with LG’s Chocolate and it’s ilk. If we were talking about anyone other than Apple it would be a no brain assumption. It is easy money. However, Apple has said on many an occasion that they are just as proud of the products they haven’t released as the ones that they have.

With the ipod, Apple quickly realized that they had the opportunity to corner the market, so they allowed some software disparity between the ipod and the mini, and a huge disparity between the ipod and the shuffle. However, I think Apple, particularly in the wake of the huge outcry to open the platform, now see the iphone/ipod touch as being far more similar to the mac (a unified software platform) than the ipod. And they are (probably) realizing that they now have a problem they would never have foreseen just a year ago: the ipod brand is actually holding back the touch! Think about it; how often do you hear about the touch on forums or sites like this? almost never, even though it is the same software platform as the iphone. We always talk about the iphone processor, or the iphone SDK, because the iphone has it’s own identity separate from the ipod. Now, the iphone is still in its infancy, so how much damage do you think would be done if a iphone mini was created right now, while the platform is still struggling (on the ipod touch side) with it’s identity? A lot of damage, that’s what.

What Apple needs to do is a big rebranding. My suggestion is that they should adopt the touch suffix for both the ipod touch and the iphone, and from now on refer to the iphone SDK as the Touch SDK, and the iphone OS as the Touch OS. They need to create a meta brand above the iphone and ipod before they can even think about introducing lower iphone models.

Just my 2 cents.
http://coderad.net StrictNon-Conformist

You know, Dan, it seems you repeatedly harp on one “fact” that you seem to believe is true, that you use in your logic, but… it just ain’t so. What’s that “fact” ? That’s the one where you claim Windows XP/Vista isn’t portable to other hardware architectures. Just assuming that Microsoft stopped it for Windows 2000 doesn’t mean that it’s actually true: it means something else entirely, and something you mention, but fail to assess correctly for Windows: the reality that all the other potential Windows platform CPUs simply have never sold enough systems where people wanted Windows of any kind to make it work.

The fact of the matter is that Windows (32 bit and I’d wager 64 bit, based on the fact that you can boot Vista 64 on EFI like a Mac has) that’s based on the NT kernel (Win9x need not apply) has *ALWAYS* had the HAL (Hardware Abstraction Layer) which makes it mostly a matter of writing a new HAL for whatever platform, and the rest of the kernel and user-space is otherwise very portable to any CPU with paging hardware and a supervisor/user-space level of execution (or more than one: x86 has ring 0, which is usually the only kernel-level ring used, and user-space is typically ring 3, but there’s 1 and 2 as well: PPC and most other processors only have user/privileged levels) and if Microsoft wanted (as in, they saw enough money in it to possibly be worth it) they can port it to any processor in a very short time.

However, that last point has the sticky-wicket of the chicken/egg dilemma of every new proprietary OS: an OS for some hardware has zero practical value without externally-written applications, and externally-written applications have no money value to third-parties if there’s not enough financial interest (volume of machines with that OS installed and used by would-be customers) for them to bother. Also, a lot of third-party applications often fail the portability test themselves, making it that much less likely third-parties will bother, because that’s an expensive testing and support proposition for the potential return value.

In addition to that, the biggest reason XP (I won’t mention Vista: it is simply *HUGE* in comparison) isn’t currently on cell phones has nothing to do with portability, and, oh, were you aware there’s Windows XP Embedded? That’s used in quite a few embedded situations, and is quite readily fit into a relatively small footprint: if Microsoft wanted to, the iPhone and iPod Touch actually have more than enough resources to run the OS itself, quite comfortably, without having to strip much out that doesn’t make sense to include in such an embedded device: things like a lot of the Services and MMC, which don’t make sense for a phone. Windows XP Embedded is setup such that you can build a system with only as many of the Windows components (even at the kernel level) as you want and need: I’ve used it in the job before my current one, for Coinstar kiosk software.

There are quite a few things I mentioned that Microsoft can trim out of XP to put it in a rather small system, especially considering how portable it really is, at least as far as executing on any processor and display device. Windows CE/Mobile/whatever-they-currently-call-it has some differences in some curious things that keep the size/complexity quite a bit lower than XP/Vista, that makes it not possible to port some older types of Windows applications, but most of these don’t matter much on the cell phone: those things (it’s been a long time since I looked) might be security API’s, and definitely MDI (Multiple Document Interface) support is gone, not that you could use that on a 320*240 screen anyway to any effect, though it is still used in Visual Studio (again, you can’t use that on a cell phone!) and I could guess that the registry stuff is stripped (I’ve not checked). But, still, I’ve only alluded to the biggest reason Windows doesn’t work well on the cell phone realm.

The biggest reason? Well, Apple clearly looked at all the stuff that’s come before, and learned what hasn’t worked, at the expense of others, and then went and did something that works rather well, for the most part. What did Apple get right that Microsoft and the others have that’s miserable? GUI!

There’s two things everyone else has screwed the pooch on when it comes to cell/portable device GUI:

1. There’s no mouse – WIMP relies on the mouse too much
2. It’s low resolution – GUIs that are WIMP rely on a higher resolution than a cell phone provides

Apple got it right by providing the iPhone/iPod Touch with a resolution that’s about as high as feasible (for budget and available tech) for something that can also be put in your pocket while also being put up to your face and ear and also being held without causing too much fatigue: a resolution slightly lower than the original Mac, and actually fairly close to old higher-end arcade video games of the same era. As a result, you can see enough text (though it may be tiny!) to get enough useful context, but not so much that it’s quite the same as a full-blown modern GUI, but by the genius of using multi-touch that’s finger-based and making it as large as it is (humanly-touchable without being too small for most) as well as (I didn’t think this’d be an advantage at first due to lack of tactile feedback) also not having a fixed-function keyboard, which is not 100% localizable (I don’t want to think how a hardware keyboard for Kanji would look, but I’m guessing it wouldn’t work well on a cell phone) but instead having something 100% dynamically-reprogrammable. Oh, having a *smooth* GUI that flows in a way that makes sense helps a lot, too :)
Realtosh

@ addicted44

I’m not worried about Apple’s strategy. Apple will have the best phones in the marketplace. They will be able to extend their cell phone lines to various price points. They just have to be careful of the partners they depend upon.

The current and future iterations of the iPhone will be the best phones in the mobile space. Apple will continually and aggressively update the iPhone, as they did with the iPod. Competitors will not get a chance to easily catch up. In fact, in the mp3 market, the competitors have been ineffectual in trying to outdo the iPod. The same will likely happen in the mobile space, especially the at more profitable top of the mobile market.

The biggest problem with Android would be for Apple to depend on Google for critical apps on iPhone. At some point, Google might feel torn between supporting Apple’s iPhone and their own Android. They may be tempted to make their software better on their own platform, as Microsoft did to Mac years ago with Office apps and Windows, and then repeated with Internet Explorer.

Luckily the Mac outlived these attempts by Microsoft to marginalize this competing platform. Today, the Mac is getting a good lion’s share of profitable computer sales. That’s got to be having an effect on manufacturers of commodity PC boxes, who are having to compete on price since there is little to distinguish one Windows licensee from another. The ever decreasing margins on the PC side are forcing the manufacturers to strip out as much as possible, and creates little incentive for PC makers to spend resources to optimize their PC designs as Apple does. It is no accident that the fastest PCs to run Windows are Macs, as Apple takes the time and money to design the insides to optimize performance.

In the same way, Apple is walking away with most of the mp3 player market, especially the most profitable higher-end. Apple established the iPod’s dominance. Then they extended the iPod line to protect from the iPod cheaper mp3 player competition.

Also, it seems clear that Apple will do the same in the mobile space. Apple will extract much of the profitability at the higher end of mobile market. Then over time, Apple will create phones lower down the spectrum with simpler designs, reliable phones, great interfaces and happy customers. As long as they can make money at it, Apple can create phones at various price points.

But first, Apple needs to establish their dominance at the top end of the cell phone market. Then they’ll be able extend their phone line to any profitable price points where they see they can make money, without sacrificing the value of the brand. — as they do with Mac and iPod.

Android is also geared to smart phones. Since it will be hard for anyone to out do Apple at the top end, some Android licensees will likely try to build lower cost smart phones to carve out a market for themselves. Apple will want to position some simpler phones to outflank these cheaper Android phones and drain any chance they might have to cut into the iPhone dominance in the smart phone mobile space.

I am confidant in Apple strategy. There is no IBM in the cell phone industry to be tricked into giving Microsoft the upper hand and the keys to a monopoly. There will be a scrappy fight based on the quality of the products and the companies that stand behind them. Apple has proved to be a amazing competitor in open all-out fights of this nature. Apple will clean up in the cell phone space. They jut need to insure that their iPhone platform does not depend on critical applications provided by the supporter (Google) of the platform (Android)with the biggest potential to compete against the iPhone.

Google is making great contributions to the value in an iPhone. Apple needs to not rely on Google’s never-ending commitment to the iPhone over Android. Apple needs to be able to fill in any holes, that a change in the current warmth of the relationship with Google, might leave behind.
Rich

“[Symbian] rapidly falls apart as it gets stretched into service as a platform for running desktop style apps like the iPhone’s browser and full email client.”

I don’t buy this argument. S60’s web-browser already contains the same rendering engine as the iPhone. It also supports advanced features which aren’t available on the iPhone, such as Flash 8 (including native YouTube vidoe viewing) and carousel viewing history.

It’s on the hardware front that S60 can’t compete – the large touchscreen with multitouch makes all the difference. That’s the difference between the two browsing experiences, not the software.

The purchase of Qt is interesting but there’s never been mention of Nokia switching to Linux for their phones – neither by the companies involved or any of the competent analysts.
stefn

Forget games; think communication. Adding iChat AV is a great reason to build up the chip structure of the iPhone/iPod Touch.
http://www.roughlydrafted.com danieleran

@ Rich: Flash 8 is really not a feature. Also, I didn’t say Nokia had announced a plan to migrate to Linux, I only pointed out that the company has taken huge steps in that direction.

Nobody believes Symbian is the future. It only makes for good PR to recount how many licensees of Symbian there are worldwide, ignoring that the three Symbian platforms are actually different. Again, this would be like roping Mac OS X, Linux, Solaris, AIX, etc together and calling it the Posix Platform to hide the reality that they’re all really separate, competing platforms.
http://www.roughlydrafted.com danieleran

@ StrictNon-Conformist: talking about the theoretical portability potential of NT is a lot like talking about how the old Apple had plans to modernize the Mac OS. The problem for both wasn’t theoretical possibility, but proven inability.

NT failed to work cross platform because MS did a poor job of developing a cross platform development and deployment system. It also never really supported the vast complications involved with running a fully insecure system across multiple processor architectures. The big problem for Alpha/MIPS/PPC NT users was that MS didn’t release service packs in tandem with x86 versions.

MS now says there “wasn’t demand,” which is BS-speak for “nobody wanted it because it was poorly done.” Perhaps it will take the same historical revisioning to the Zune so that future generations can be assured that it was a great product, but the demand never materialized for it.

As for “Windows XP Embedded,” it’s not mobile at all, its just MS’ “fix it with Windows” solution (ie every thing looks like a nail when you only have a hammer).

Windows-based ATM, register, and Coinstar machines ARE A PC attached to some specialized hardware, not some sophisticated mobile systems running a scaled down version of Windows that can also be used in mobile CE devices. Look at UMPC: its a total joke in usability.

As you point out, Apple’s WiFi mobile platform is purpose-designed. If Apple had tried to scale down an Intel based Mac (with the desktop UI) into a UMPC form factor, it’d do about as well as UMPC, too.
http://www.roughlydrafted.com danieleran

@ beanie

“UMPC is entering the middle of the S-curve product cycle where expansion and growth accelerate from the early adopters.”

Not every product has a growth curve. UMPC is a rebranding of the same failed Tablet PC products Microsoft has been pushing since the early 90s, and there has been absolutely no uptake AT ALL.

“Atom is also great for fanless set-top media devices. I believe Apple TV under-clocks the 1GHZ CPU to around 350MHZ and has a fan. With the Atom, it could run at full speed, 1GHZ, and could go fanless.”

That may be the case, but Apple TV is a pretty low-profit, experimental device. Look at Apple’s high volume AirPort business (I haven’t seen the latest figures from NPD, but last year Apple was the top consumer WiFi base station vendor): fanless, compact, does lots of stuff, and based on ARM processors.

Apple clearly doesn’t need Atom. Intel would love to get Apple’s business however.
Pingback: [Mega Merge] Australia/3G iPhone speculation - Page 57 - MacTalk Forums()
http://coderad.net StrictNon-Conformist

Daniel,

Microsoft proved the OS was very portable: long before x86 went to 64 bit, Windows was 64 bit. It also was deployed on hardware platforms of opposite endianness, which is a major measure of portability. When it comes to large projects, most developers do NOT use an IDE like Visual Studio, and instead use the command line tools: thus, your arguments don’t hold water. I maintain that you claiming that you know Windows (NT-based) was not/is not portable is without merit whatsoever, and I insist that you post your sources from online whereby you get that information, because having taken a careful look at your online resume, I carefully note you’ve not ever programmed in a single systems language: they’ve all been internal languages to content creation products, or some web scripting languages. I also found it entertaining that you claim the University of California claims you’ve got the equivalent of a Masters degree in Computer Science: what you’ve done may accurately depict something similar in Information Technology or Informatics, but is far too weak to qualify as a Masters of Computer Science, or even a Bachelors.

What you’ve done amounts to a lot of content creation and network and machine management: this is technician’s stuff and MIS/IT stuff, outside of whatever web development you’ve done, along with project management in there somewhere, which doesn’t mean squat for knowing how to program. What you need to remember is how many of your readers are qualified, experienced developers that have worked at a low level, who have (like I have at 8-bit, 16-bit, 32-bit and 64-bit with differing byte orders) developed at the OS and driver level, and work on large, complicated applications, and have done so over multiple machine hardware and OS platforms, who actually know what we’re talking about when the term “portability” is bandied about. I find your articles interesting, and I’m definitely on the Apple platform (first on it in the Apple 2 days, now working towards iPhone software development) but stepping outside of your experience is a dangerous thing to do, because you can almost be certain someone that knows what’s what will call you on it, and it detracts greatly from the rest of the article in question.

Oh, as regards to the statement you make about UMPC and usability? I guess that’s all a matter of semantics: is OS X (the kernel) the whole OS, or is that just Darwin? As-is, Darwin is just as “useable” as Linux on the command line: without Aqua, which is made possible to develop for by the Cocoa Frameworks/API, it is just a command line. So, too, you can still technically be running Win32 from a command line, though you’ll obviously be stuck without any GUI apps: you still have the same power of console apps the Linux and Darwin people have. If OS X were also stuck running on a 320*240 screen, its usability would be greatly reduced, too, because that’s just not a sane amount of screen real estate to make a nice GUI with icons, and is similar to the concept of trying to drink an ocean through a straw: you’ll eventually accomplish it (assuming you are immortal), but it’ll really suck! If Apple hadn’t come up with the multitouch touchscreen, it’d be hard to see how they’d do much better with 320*240 resolution, and even at that resolution, it’d be diminished greatly: Apple took the higher practical road with what was currently available (not the highest possible: just the highest practical for all engineering constraints) and saw the weaknesses with being stuck with that (relatively) low resolution, and having to use it with fingers without carrying around a mouse or stylus (yuck, for a phone???) and solved the problems, mostly pretty well, considering that too much larger, they’d have a PDA that’s too big to stick in a pocket, and too much smaller, they’d have something that was “cute” but not suited to the majority of the human population: very much like their simplified product lines that have general purpose machine categories that may not be a perfect fit for everyone, but there’s something that’s close enough to get the majority in that group mostly happy.
http://johnsessays.blogspot.com John Muir

@ StrictNon-Conformist

Right, so what exactly was your point again?

UMPC = epic market fail.
Windows anywhere beyond the x86 desktop = ditto.

Some pretty hokey portability on show there, whether it sounds good in the dev libraries or otherwise.
http://coderad.net StrictNon-Conformist

@John Muir

Dan was trying to state that Windows at the lowest level isn’t portable to other machine processors, which is blatantly false, as it was designed from the start to be portable.

That there was a massive failure in the market due to chicken/egg problems of there not being enough machines of the other processor types to make it profitable for Microsoft for the amount of returns, and also not profitable for the number of potential titles copies sold for Windows developers (who generally rely on the huge market base in numbers of x86 users) is part of that problem.

Microsoft did themselves a disservice, I believe, by attempting to do what they did with all their mutations they’ve called “Windows Mobile, CE, xxx” in that it is far enough away from the main Windows code as to being notably different to develop for, though admittedly, some of the things they stripped out (Multiple Document Interface) had no real reason to be kept at all. But, the stripped down Windows mutations are not the question he answered wrong that I am harping on here: it’s the basic premise that he stated Windows itself was not portable, and he claims this all without having any personal experience at a level that would allow him to be remotely qualified to judge this.

And as I pointed out towards the end of the most recent post, if Apple had tried to do an exact duplication of their desktop interface on a 320*240 screen (and perhaps even on the 480*320 screen without multitouch) while it might have looked nicer, and while Apple may keep it locked down so applications don’t make it as crashy as I hear a lot of Windows Mobile systems are, it’d still be a tough GUI to shoehorn into such a small amount of real estate. But, the GUI thing (interface-wise) is not the original question I answered that was brought up about portability, though it does go directly to usability and thus also to marketability: Microsoft failed to insist that portable devices have the capacity to do a WIMP GUI justice, and I expect that’s at least part of the equation for their failure (but certainly not all by a long stretch) when going for the mobile world.
http://www.roughlydrafted.com danieleran

@StrictNon-Conformist: You should be careful when announcing what somebody else “was trying to state.”

I wasn’t “trying to state that Windows at the lowest level isn’t portable to other machine processors,” that’s just the strawman you found easier to attack that the point I actually made: that Microsoft tried to deliver NT as a portable OS and failed miserably, years after the much smaller NeXT seamlessly accomplished the same task.

All Microsoft had to do was copy. Instead, it did the typical sloppy job where its products were designed by marketing pinheads instead of engineers. Cutler did make great pains to develop a hardware abstraction layer, and that attempted to accomplish what MS planned for NT, but by the time NT was functional (v5 Windows 2000), cross platform efforts were dead.

You can blame chickens and eggs and the market for not having enough Alpha and MIPS machines to deliver a functional NT port, but the obvious question is: Why could the much smaller NeXT deliver functional ports to not only Alpha and MIPS, but also SPARC and HP PA-RISC?

WinCE was a product MS undertook because NT was not optimized to run on anything smaller than a workstation class desktop. It was an attempt to replace Unix with a new wheel, albeit a wheel that didn’t work as well, had a poor process model, had severe security problems, and was slathered with other MS absurdities such as the Registry. NT was not a general purpose, high quality OS.

Meanwhile, once NeXT got some funding, it was able to not only allow Apple to migrate its Mac platform to new hardware, but also scale to mobiles using yet another processor.

MS did fail in trying to put the (horrible) Win95 interface on mobile devices, but it also failed in developing NT as a purportedly portable OS that couldn’t port or scale effectively.

If you blame MS’ problems on the era of technology in the mid 90s, I’ll point out that NeXT developed its OS in the late 80s.

If you blame MS’ problems on difficulties in the hardware market, I’ll point out that MS maintained a strict monopoly over the PC desktop, and every other vendor (again, NeXT) struggled far harder.

What other excuses can one imagine for explaining away why MS’ technical incompetence was not the company’s fault? That it didn’t have enough money? That it lacked access to talent? That it was restrained by an outside monopolist? That it didn’t have political clout? That it faced a tough economy?

As some point, you have to stop making excuses for the most ridiculous company to ever exist in technology, and just admit that MS is a horrible disease that needs to be eradicated, just like its namesake.
Realtosh

@ Daniel, John

http://www.businessweek.com/technology/content/may2008/tc2008056_305255.htm?campaign_id=rss_tech

Here’s one of the first articles in a more mainstream publication, that reflects on some of the conflicts of interest that may exist between iPhone and Android; and also between Apple and Google, especially with Schmidt, CEO of Google sitting on Apple’s board, when the companies at competing against it each other in the mobile phone space.

http://www.businessweek.com/technology/content/may2008/tc2008056_305255.htm?campaign_id=rss_tech
Pingback: Mobile EEE PC, UMPC, and Internet Tablets vs the iPhone — RoughlyDrafted Magazine()

ARM, x86 Chip Makers Fight to Ride Mobile Growth

Navigation

Search this site

Recent Articles

Search This Site

Categories