Just over a month back we were told that Google has cumulatively activated 200 million Android devices, rapidly closing in on the 250 million+ iOS units Apple has reported.

How many of those remain in play?

I strongly suspect that many more early Apple devices remain "on the road", and the operational current count heavily favours iOS.

Rapid Obsolescence

I currently have four Android smartphones. The original HTC G1 and Magic+, a Nexus One, and a Galaxy S II. Add two Android tablets (a viewSonic gTablet and an Acer Iconia A500).

The HTC G1 and Magic+ inhabit the bowels of some drawer somewhere. The Nexus One sees periodic use only when my oldest son wants to watch some YouTube videos on the couch and can't find my S2. The gTablet is derelict, while the A500 sits broken.

I have six Android devices, five legitimately activated through Google, of which one sees real use. Many early Android devices were, in many regards, the Hyundai Pony's of the technology world, quickly rendered unusable or with limited resale value.

Speaking of which, resale value is a good indicator. Early in the era of the Magic+ I considered selling it to get a better device, my hopes dashed seeing Craigslist jam packed with people trying to do the same, pricing down to "not worth the trouble" prices. At the same time the few used 3GS units were barely discounted from retail price.

The same was true of the Nexus One, the Galaxy S, and various other Android devices. The resale value just isn't there, and I suspect that rather than supporting a robust second-hand device market, most simply got abandoned.

The same certainly isn't true of the Apple devices. I would wager that the overwhelming majority of Apple devices from the 3GS on are still in use, whether by their original, second, or third owner.

Like a Toyota, the value (and corresponding resale value) is still there.

Supporting Metrics

All of this is just speculation. Further, it's from someone who loves what Android has become, and what it represents. Android is big enough, and successful enough, that we don't need to hide from this reality.

There are some metrics that support the hypothesis. Look at the Facebook metrics for iOS versus Android - 79 million monthly average users on the former, versus 45 million on the latter. Despite having some 80% of the reported activated base of iOS, Android only sees just 56% of the Facebook use.

Vast differences between the demographics might explain the usage difference, but if that were the case then it should actually skew things in favour of Android: it is the platform that is preferred by younger adults where Facebook use is saturated, while iOS has primarily taken off in the 35+ realm.

It isn't demographics. Any nonsense about people getting Android phones and using them as dumb-phones ring ridiculous.

Among devices still in use I would guess that Facebook usage rates are similar. It's simply that tens of millions of Android devices are no longer in play.

Which is a number far more relevant than the gross sum of ever activated devices. Tell us how many devices are still in use, and strive to keep that number as high as possible.

So what gives? So many pundits tell us that it's a great device. That it's the best option....if for some reason you don't want an iPhone.

Which, as they say, is the rub: The people who love Windows Phone 7 are almost universally iPhone boosters, and they love WP7 as the non-threatening, adorable little scramp that they hope will distract the Android legions. The regular cheerleading squad is at play: Marco Arment, John Gruber, MG Siegler, and others of the "Apple profits are righteous" gang.

They love Windows Phone 7 because it legitimized the iPhone by making an almost identical set of philosophical platform choices. The few areas where it differs are often facile. Windows Phone 7 comes from multiple manufacturers, for instance, but the hardware is so strictly defined that the differences are minor, the advantages of the competitive spirit extinguished.

It turns out that if you want an iPhone, there is a company that already does a really good job making one.

Whatever misdirected soul has demolished all of Microsoft's once legendary advantages, desperately trying to out-Apple Apple, needs to be punted from the company. This strategy will not and cannot work. When the people who supposedly love your product, giving it endless praise, are people who would never buy it, you're doing something wrong.

From Gizmodo's email blast last eve-

If you have any decently modern Android phone, everything you do is being recorded by hidden software lurking inside. It even circumvents web encryption and grabs everything—including your passwords and Google queries.

So I went checking, actually hoping to find this villainous software.

My Galaxy S II - No Carrier IQ.

My wife's Captivate Glide - No Carrier IQ.

My Nexus One - No Carrier IQ.

My peer's new Razr - No Carrier IQ.

I expect little from Gizmodo, but such a fundamentally unproven claim — so trivial to dismiss — really sets a new low.

But they're hardly alone. This whole story has been an exercise of hysterics and technical ignorance. Even on self-purportedly enlightened sites like Hacker News, the dominant opinion is superficial, lacking any real curiousity or discernment at all.

Why I Don't Fear Carrier IQ

If I did find the software on one of my devices, I can't say I'd be overwhelmed with fear.

The "security researcher" — realize that the media invents professions and credentials when they need a stooge to prop up a story without legs: when the WSJ ran with some domain stats I had gathered, they declared me the "world's preeminate domainologist". In reality my interest in domains started and ended with the fact that it was a large set of data usable for an index tutorial — became aware of the daemon on his Sprint phone while using the logcat tool available in the Android SDK. This is a tool run by millions upon millions of devs.

His observation actually wasn't unique, and the service had been noted and dismissed by many other devs using Sprint Android handsets over the prior year. He was the first, it seems, to announce unsupported, hysterical conclusions, the drama escalated when Carrier IQ poured fuel on the fire by sending him a heavy-handed cease and desist, which they later retracted (though I suspect that this whole episode will end in some large libel lawsuits against very sloppy organizations like Gizmodo).

He wasn't using a "packet sniffer". He was using the most rudimentary debugging tool that Android developers use, whereupon he found this software literally announcing its presence and what it was doing. Basically the exact opposite of hiding itself.

"But it records your keystrokes!" you say. There has been zero proof of this thus far. What has been shown is that the daemon software intercepts keystrokes, proudly announcing its achievement into an in-memory ring-buffer log. I suspect — with zero proof — that it uses these events to populate basic aggregate data such as "uses the keyboard heavily". Nor has it been shown that any of this data is actually transmitted to Carrier IQ beyond aggregate statistics.

Where are the logs of this activity? The history of your keypresses and your web activity? This is brutally easy to find, if it exists. Find it and become an internet hero.

But I doubt you will. Know why? Because the legal landmine that Carrier IQ would find itself in, operating in one of the most heavily regulated industries around.

It goes against intuition that they would record, much less transmit to themselves, this information when it would represent enormous liability. Perhaps I'll be proven wrong, but I highly doubt it.

Nothing that has been found contradicts exactly what Carrier IQ stated in their public announcement. All of the purported proof otherwise is a demonstration that most of the people manning the keyboards of the tubes don't have the slightest clue what they're talking about.

EDIT - 10:28PM - I've gotten a few emails deriding my "douchey" tone, in particular that I scarequoted "security researcher". I did so intentionally: Tech news has gone down a path of stupidification where nothing is validated, and where an expert opinion is drawn from the most rudimentary, superficial analysis. Looking at logcat entries does not make a security researcher — it makes a log dabbler.

In any case, Carrier IQ has spoken up regarding the issue, explaining pretty much exactly what I guessed above. Do you have to trust them? Perhaps not. But their explanation is entirely rational and in line with the service they provide.

EDIT - December 5th - Quite a dramatic narrative change has occurred. Now instead of being on every newer Android phone, it's actually on a subset from two US carriers (AT&T and Sprint). Further it isn't "recording" your keystrokes, it is, perhaps insecurely, logging a subset of details to the in-memory ring-buffer log (which is not "recording"). A nefarious app — if it requested and was granted the READ_LOGS manifest permission that startling few apps have — could exploit this sloppy logging.

Even the mainstream media is reporting on Adobe's announcement that they're retrenching from Flash on mobile. While Adobe's statement is wishy washy and half-hearted, the outcome is certain: Flash is dead.

Flash is dead on every platform, including the desktop. Flash on mobile was a zero financial return activity Adobe took to try to extend the best-before date of the technology as a whole. As it fades away the whole stack becomes questionable.

Increasingly there is one web as the days of the mobile romper-room draws to a close. While client specific limitations remain, my smartphone is more capable than some desktop browsers. Binning by the physical size of the device isn't reasonable (and is almost always unwelcome, which is why many smartphone users now spoof their userAgent to avoid such web crippling).

If Flash is a no go on many of the web's consumers, it makes no sense for any of the web's consumers. If you have HTML5 resources and implementations at the ready, why would you ever publish a synonymous Flash resource? The answer in practice today is for economic segmentation: Many sites demand a subscription to access the HTML5 resources, temporarily taxing iOS users. That isn't sustainable.

My Position On Flash

More than a few have emailed me with what amounts to a "nah nah told you so!" victory dance. Apparently some took a prior missive I wrote as pro-Flash. How they came away with such a fundamentally incorrect interpretation is open to serious question.

Nonetheless, technically Flash works on mobile, perhaps simply because the devices finally caught up to its demands. On the Galaxy S II it plays simply brilliantly, without an ounce of hesitation or strain. It really, actually works.

Nothing Steve Jobs said about Flash — his position motivated by the lucrative (in more ways than the obvious) app store — was proven right by Adobe's financial predicament and resource refocusing. Steve Jobs fulfilled his own prophecy, however, helping to hasten the demise of a derelict technology.

False Prophets

Flash sucks. Flash sucks slightly less, however, than exclusive ghettos for single platforms, leaving alternatives and the spirit of competition out in the cold.

And let's be real -- Many of the false prophets of HTML5, cheering the loudest over Adobe's announcement, boost HTML5 half-heartedly as a proxy in a fight they should have little interest in. These usual suspects take every other opportunity to cheerlead native apps over web apps.

Save us your fraudulent cheering.

It's going to fail miserably. It will sell in limited numbers to the few looking for something new and interesting — a conversation piece to bore guests with — but it's a partial solution searching for a problem.

People who have difficulties with programmable thermostats don't care about learning thermostats. The claimed energy advantages are almost certainly overblown, as they generally are with energy efficiency products: A modern insulated home has temperature inertia that sees limited benefit from anything short of vacation mode, and certainly sees no benefit over the hypothetically miniscule advantage this might provide over a standard "business day/night cycle" pre-baked programmable thermostat cycle.

Add that people who have the sort of schedule it could learn are exactly the sort of people that a bog standard programmable thermostat, of the hundreds commercially available, caters to perfectly. The people who don't aren't served by this product.

But it's a very interesting initiative, and the brains behind it will hopefully evolve into more interesting and more commercially viable solutions.

These are the offshoot benefits of the smartphone war: Displays, processors, RAM...it is astonishing how inexpensive incredibly powerful solutions are now, coupled with a low power draw that makes them viable and embeddable in everything. I have derelict smartphones stashed away in drawers that would demolish the expensive embedded solutions we used to drive critical systems just over a decade ago.

Years back a peer and I did the initial steps of a "startup" in exactly this space. We contrived the name "tenom", with our imagined logo being Monet's signature rearranged (not sure of the legality of such a repurposing), the second idea being the bar code representation of ten Ω (get it? tenom...ten ohm? I was so proud of myself). We were in the embedded computing market and it was obvious that the home as it was lacked intelligence. We wanted to give it some IQ.

Our imagined solution included occupancy sensors, switched ducts with area HVAC control, and integrated security. We spoke with some builders and it was fairly obvious that it wasn't going anywhere: the markup on a home to implement the solution had a baseline in the tens of thousands of dollars, but worse most buyers simply didn't care.

Technology has come a long way since then.

Now your HVAC controller can actively monitor current and coming weather conditions to, essentially, "prepare" (flush the home with exterior air during a cool night and then seal it up as the sun rises, or vice versa). Rather than rudimentary light and proximity sensors, it could know where every member of the household is simply by the fact that many carry a smartphone nowadays.

It could monitor our calendars, hook into our Google Latitude accounts, all to proactively prepare for ideal conditions the moment the first of us arrives home.

And of course people are doing things just like this right now with homegrown solutions. Embedded control boards are a pittance now.

The future is wild. Inexpensive, fantastic quality displays/touchscrens and high performance but miniscule power processors ensure that it will be a rich experience.

I've written a lot about smartphones over the past while. It's a topic that I get more passionate about than I need to. I've tried to pull away from the discussion but it keeps pulling me back in.

It is a discussion that matters. It really, really matters. It is the most important technology shift that has happened in years, and it will have, and is already having, a very strong influence on this industry.

But the primary battle is over. The war is won, with users victorious.

This is true regardless of the smartphone platform you choose to embrace, whether iOS, Android, Blackberry, Windows Phone, or other.

There was a brief period where service providers and application makers felt that the market had become uniplatform: If you wanted to watch their movies, listen to their music, do banking, match colours, and so on, they only needed to cater to one closed platform. It's no surprise that many of them were overjoyed with this situation at the outset (before it was turned against them), as it makes life simpler.

It also threatened innovation and competition.

That lock is broken. It is a thing of the past. No longer does one platform reign supreme, and the new reality of a multiplatform world is established and will be going no where anytime soon.

We all win. Even if you prefer the previously dominant platform, you win as you can be sure that the innovation and pace of improvements in their product are heavily motivated by the competition.

It is a fantastic world that we live in. We're in an amazing place.

Where do we go from here?

Since it's unveiling, however, quite a few have questioned why Apply didn't bump up the screen size. Some of the popular competitive devices have larger screens (while others have smaller screens, but such goes a market where people choose devices ideal for them), so many expected Apple to respond in kind and grow the device from the size it adopted back when 3.5" was actually a pretty huge display.

They didn't. For whatever reason they stuck with 3.5"(*). Some declared it a weakness.

Never fear, though, the Ex Post Facto squad is here to save the day. They're ready to tell you that a small screen is not a competitive weakness, but instead is actually a universal strength. Once Apple let the cat out of the bag, the hordes started defensively building justifications for why a 3.5" screen is not just good — which it most certainly is for a large class of users — but rather that it is, in fact, superior to other options for all uses for all users.

Larger screens are undesirable, they say, because they project a need to compensate for a small penis (this is seriously one of the arguments, taking a breather from its traditional utility in making jealous people feel better about people having nicer stuff), they're like the tail fins on 1950s cars, or they're too large for the single use case of walking in a parking lot while using your device with one hand.

The last one astounded me not for its irrational position — such ridiculous claims are a dime a dozen, especially among the cottage industry of "I own an iPhone and now I'm trying an Android device and thus I am an expert on mobile devices and my complaints are righteous" — but rather that it actually gained traction across the net: I've seen it linked and held as if it were a scientific truth on countless sites.

Ignoring some technical flaws in the presentation, it opined that because the smaller iPhone is arguably more usable for one-handed on-the-go usage, at least in the pre-voice recognition era, that validates the design as a work of brilliance.

But what about every other usage model. 99% of the time that I use my device I flip it to landscape mode and type with both hands, enjoying a large, clear, sharable-if-I-want display that doubles as a navigation device. Don't those uses count? Doesn't that invalidate that irrational claim right at the outset? Isn't it just as brilliant and considerate for makers to satisfy those needs as well?

A modern smartphone has a lot of uses. You choose the size of device based upon how you use your device, not because someone is desperately trying to rationalize the heavenly design of one single option.

* - the iPhone features a 1.5:1 aspect ratio versus the 1.66:1 screen on most competitor devices. This is how the Galaxy S II, with 7.8 square inches of screen space (38% more than an iPhone 4's 5.65 square inch screen) is just 12% wider than an iPhone 4. What this means in practice is that maintaining that aspect ratio while growing the screen leads to a quickly chubbifying device. To grow the iPhone screen to 4.3" would increase the device width, assuming a similar bezel, to 15% larger than a Galaxy S II (or about 75mm wide). I would guess that Apple would rather gain resolution independence in their API before growing the device, with which they can start to vary the aspect ratios.

This is fantastic news as it discounts all rumors that the device was running ancient versions of Android. That's important as it ensures the device comes out of the gate with the many foundational improvements that Google has added to the OS. It also amply demonstrates that Amazon's fork is minimalist, and is likely limited to the shell.

The hardware itself is missing a few features. It doesn't have a gyroscope, for instance, however such is a hardware facet that has gone almost entirely unused in the few products that feature it (some are confusing a gyroscope for an accelerometer. The device does have full motion sensing, just not of the gyroscope variety). The target itself is generally locked down, but that's okay -- for Android developers you can choose to participate in that curated garden or not, with ample alternatives if you decide not to.

Has capitalism ever seen such a rapid accumulation of corporate wealth before?

I'm less enthused about Steve Jobs the legend, created in the wake of his death. While I appreciate that there's the "too soon" period, during which mass ignorance and rewriting of history are more excusable, the canonization of Jobs, and the overstating of his and his organization's contribution, has reached intolerable levels.

Like most greats, Steve Jobs stood on the shoulders of others. He rightly believed that good artists borrow, great artists steal.

He was but a real human being, with all the limitations and failings that come along with such a reality.

And it runs Android.

"But the Android part doesn't really matter" so many pundits are busy stating, trying to convince each other that this device falls in the "other" category.

Utter nonsense. The Android part is critical.

Instantly the device has countless available applications. Those apps are built on a matured, evolved, well understood coding platform.

They went the curation route, hooking it into the Amazon App Store (not sure about side-loading), however that is an almost irrelevant bit of fluff, as anyone building apps for Gingerbread or lower is now heavily motivated to go through the trivial process to get their app in that catalog.

No special coding necessary. If you've built an app for Android, including copious use of the NDK, submit it to Amazon's App Store and voila, you're rolling on the Fire, and can benefit from its success.

I think Amazon's "forking" is a tad overstated. Given that they have the entire runtime platform to support the very, very broad functionality of third-party applications, it seems more likely that they essentially replaced the shell while hooking in some media programs. If it's true that they're running 2.3.5, as rumor has it -- and the version really matters because with the version come the many improvements that have happened in the code chain -- then such a superficial variance is proven as it would be unlikely they could merge from the trunk so quickly otherwise.

If you aren't paying for it, you aren't the customer, but instead are the product being sold.

While there is an element of truth there, it's a dangerously simplistic, naive way of viewing customer-business relationships.

Consider that I paid GM a lot of money for a big, Earth-destroying SUV. Further, I pay their OnStar division a not-insignificant monthly fee for their services. In return OnStar recently revised their user agreement in such a manner that make me one of the products they sell.

This is hardly a new behaviour for businesses collecting lots of data. Tom Tom, as one of countless examples, sold user data to police departments, among others, allowing them to more appropriately place speed traps.

Modern businesses comprise many divisions, each of which seeks to maximize revenue however possible. If they can package their customers as products and sell them, in many cases they do. Assuming that the transfer of money makes the relationship more pure or singular is pure sophistry.

In an entry a few weeks back I opined that Apple's success with the iPhone presents an incredible risk to Microsoft: There is a tremendous overlap between people who already owned an iPhone and those who bought an iPad, and now we're seeing the same thing playing out in the PC market (where PC includes Macs).

It's the iPhone halo-effect. People who love their iPhone tend to love the brand, and from that the other products (not discounting that the quality that Apple put into their iPhone is in their other products as well, so it isn't an entirely undeserved carry-over).

Those things I don't love, however, can be traced back to Samsung's infatuation with all things Apple.

A lack of a notification light, for instance. This makes a 10 out of 10 device a 9 out of 10, in my opinion. I have absolutely no doubt that when Samsung rationalized this decision they looked to the iPhone as their model: It doesn't have one, so better without, right? Not even close. It makes it a major nuisance knowing if there are notifications pending, reducing the utility and value of the device.

The Galaxy S II also lacks a trackball of any sort. This makes editing text a PITA as you try to move the cursor around awkwardly by poking at positions on the screen. Track "balls" (preferrably optical) also help for on the go phone use, allowing you to scroll emails, etc, with the phone in one hand.

Both are failings of the iPhone, not features or redeeming qualities. Copying such brokenness doesn't help the platform. Learn to ignore those who will invariably herald whatever Apple does because their rhetoric is not evidence based, it's simply an indoctrination to follow and champion.

Even iOS' rubber-band scrolling (where it lets you exceed the bounds of a list but then bounces back) makes an appearance on the GS II, as it does on HTC's Sense. Why? This is not an intuitive interface element, and is deficient compared to Gingerbread's stock behavior. There is no real life metaphor that correlates with this broken usability.

Stop copying the iPhone. Seriously, just stop. You don't need to, and when you do it makes your products worse.

And just to get this in the wild to avoid the inevitable patents, start thinking up better input techniques. I would like a small touch sensor on the back of the device, for instance (hold your device and you can immediately imagine hundreds of uses). The possibilities are endless. Don't wait for Apple to prove that it can be done.

Update 2011-09-17- Samsung has a simply incredible manufacturing process: They seem to make a million variations of everything, specifically targeting niches, carriers, and industries (amply disproving the garbage "make one that is not quite right for everyone and that will be the best" mantra). For the Sprint variation of the Galaxy S II they did away with the large, rectangular iPhone button and went for the four capacitive buttons, added a notification light, among other not-like-an-iPhone changes. The result is a much better device. Hopefully they learn from this.

Consider that simple joins are continually held as one of the primary weakpoints of many web scale implementations. Add the complexity of joining GUIDs (not even considering the lower data density, slower inserts, etc). It can be brutal.

This is one of those topics that is often incorrectly classified as premature optimization. It isn't. Your data structures will often live longer than any other element of your project. That is simple truth, demonstrated by millions of projects across millions of organizations.