Why Droid was worth the wait

November 11, 2009 | Anand Shimpi | Comments | |

AnandShimpiThe dirty secret in the smartphone business is that it’s not marketing, but patience that makes a top-selling phone — the patience not to jump in until you’ve managed to couple great software with great hardware.

Consider Apple, which made the insane figure of $1.67 billion last quarter. At a high level, to the uninformed, it would seem like Apple is nothing more than a design house that can order components. It rarely uses hardware that isn’t available to its competitors, yet Apple is able to post incredible earnings in a recession. Instead of relying on exclusive components to give it an advantage, Apple makes good hardware choices and couples it with solid design and software development. Apple may masquerade as a good hardware manufacturer, but it’s really a good hardware chooser.

Now Motorola’s new Droid smartphone shows us Motorola’s patience and careful selection of hardware and software could pay off well, too. It’s new Droid phone is the first Android phone built with good hardware. Every previous Android phone on the market uses an ARM11 processor, originally introduced in 2003. When I wrote about the ARM11 core in my iPhone 3GS review, I compared it to a modern day 486. Remember the 486? It was great, for 1992. Not so great today. What do you pair a slow CPU with? Slow graphics, of course, which again, has plagued every single Android phone on the market thus far.

To be fair, the first versions of the Android OS didn’t natively support anything but these slow platforms. But that’s also where Motorola’s been smart. It waited patiently for the new 2.0 release of Google’s Android platform, a platform that supports a nifty little chip from Texas Instruments called the OMAP3430. The OMAP3430 combines an ARM Cortex A8 CPU core and a PowerVR SGX graphics core. It’s the same chip that’s in the Palm Pre, and comparable to what’s in the iPhone 3GS. If the older Android phones used something akin to a modern day 486, the Droid uses a modern day Pentium. We’re not quite bleeding edge yet, but we’re getting there.

Smartphones are PCs, just slower. That’s not meant to be a controversial statement, it’s just the truth. A modern day PC is made up of a general purpose processor (that Intel processor that all those rockstars are talking about), a graphics processor (that GPU NVIDIA keeps saying is going to take over the world), a chipset to get the two to talk, some memory/storage and a basketful of other chips that enable things like USB, Bluetooth and WiFi. A modern day smartphone is made up of the same components, it’s just that Intel doesn’t make the CPU, NVIDIA doesn’t make the GPU, and many of these processors are integrated onto a single chip. It’s called the SoC, or system-on-a-chip, and it’s what happens when you iterate Moore’s Law enough times that you can build fast-enough hardware in a very small space.

Previous Android phones had SoCs that took an ARM11 CPU core and mated it with some similarly unimpressive graphics core. To put it in perspective, this gave most Android phones a first-generation iPhone level of performance. Remember that the first iPhone came out in 2007. That may not seem like a long time, but in CPU performance terms, it is.

Now, Motorola didn’t do anything special to make the OMAP3430 work. Google and Texas Instruments did everything there. Motorola simply aligned its schedule with the Android 2.0 release and remained patient while its competitors sold slow Android phones. Motorola is the first, but definitely not the last. Next year we’ll see more SoCs based on Cortex A8 (and A9 — oooh bigger numbers) make their way into Android phones.

I will say that my two biggest issues with the first generation of Android phones were the lack of multitouch gestures and poor performance. With phones like the Droid using (more) modern hardware, we’ll finally get to fixing the performance complaint. Better performance means a more attractive Android platform, which means better competition for Windows Mobile, iPhone OS and Web OS based smartphones. Which in the end keeps everyone fighting and innovation churning. Sweet.

Anand Shimpi is Editor in Chief and CEO of IT industry news and technology analysis site AnandTech, where he has been reviewing the latest from PC components to gadgets since 1997.

Next Story: 301Works teams up with Internet Archive to preserve shortened links
Previous Story: Nominate your favorite companies for the Crunchies Awards

Bookmark and Share

  • chukari
    has the writer used latest htc hero? then you may revise your whole blurb. how about doing some significant homework first ........
  • anandshimpi
    The HTC Hero uses a 528MHz ARM11 based SoC, it's still in the same class as the previous Android phones. It's unfortunate because I was looking forward to it being the first Android phone to break the mold.

    Take care,
    Anand
  • Justa Notherguy
    Nice analysis, Anand. Not least of which because - as usual - you stick to insightful commentary spare us the jargon.

    Speaking of hardware, what are your thoughts on how Android phones skimp on the ROM? Do you figure this was just a figment of the FOSS model (like, say, because the manufacturers are pinching pennies)? Maybe its part of Google's Master Plan - pushing the industry toward a secure App2SD model? And, in either case, do you think this will hurt Android's chances for market share, short-term?
  • mesfldds
    Anand from Anandtech!?!? Its neat to put a face to a name. Anand is a stud!
  • HereAndNow
    An ARM Cortex-A8 CPU is definitely better for mid-to-high end smartphones, but what about the low end (i.e. feature phones, appliances)?

    Presumably, the upcoming ARM Cortex-A5 CPU (ARM11 successor) is intended to address this segment. It should make it possible to push Android all the way down the device chain, creating a HUGE market of devices (hundreds of millions?...billions?), for developers to offer their Android apps.
  • benjitek
    The writer of this blog doesn't seem to be all that familiar with the subject matter. While I use a PC, I keep an eye on the industry in general, and if you don't think Apple is a hardware-designer, instead stating them to be a hardware chooser -- you should really take a look at what you're writing about; without doing so you just appear to be attempting to fill a web page to meet a deadline :(
  • Justa Notherguy
    The direct quote from Anand is:

    > Apple may masquerade as a good hardware manufacturer [...]

    Either you missed the point or you don't understand the vagaries of modern product-design. Today, in large part, hardware-design _is_ 'hardware choosing'. Engineers are encouraged (kinda like fish are 'encouraged' into the net) to use off-the-shelf components, sub-assemblies, etc. 'Custom' pieces are kept to a _bare_ minimum.

    That's how companies like Apple - with no manufacturing facilities of their own - minimize costs and meet tight schedules. But, as Anand mentioned, there's no magic to the bits & pieces they choose...Samsung can buy 'em, too. But what Apple does very well is (a) they know when saving money harms the consumer's experience, (b) so they don't.

    As Anand implies, that's an important skill. And its a rare one, too. That's how HTC, Samsung, Dell, etc. all made the same mistake - they picked the wrong chipsets for their first Android handsets. Sure, their choices cut their per-unit costs by a small amount, but the end result is phones that are merely 'decent'. Worse, they make the O/S look bad.

    My dad used to call this foolishness 'false economy'. Somehow, Apple consistently manages to avoid that trap. So, its nice to see Motorola - a company in such poor shape they have plenty of reasons for trying to pinch pennies - take the high road, with Droid. The extra effort shows, in a fine phone that could help save its makers from a sorry end.
  • mobi
    You have no idea what you are talking about. Motorola wasn't patient. They came out with the Cliq before the Droid. Motorola would have come out with an Android device a year ago if they could have. They were incapable of doing it sooner.

    They would be pretty stupid to watch their market share approach zero simply because Android 2.0 wasn't available.

    This is a terrible article
  • When I think Anand, I think Anand Lal Shimpi... I've been reading Anandtech for years! This is so cool, keep writing stuff here!