Hacker News Reader

Comment by shingen | Hacker News Reader

The endless parade of fragmentation discussions regarding Android all prove one thing: a lot of people don't understand what wins a market. Customers lacking in technological savvy make up the radical majority of Android owners, they are never going to care about this issue so long as their smart phone functions at 85% good enough.

85% good enough, is more than enough to win a market if you have substantial other advantages in your favor.

Windows has always suffered fragmentation, and has rarely been better than 85% good enough. Circa 2002 they had '95, 98, ME, and XP spread across over half a billion units. With further varied patch configurations and service packs. The total global varied combinations are in the hundreds in that context.

But, you say, that's just so horrible! Welcome to reality, where markets are never dominated by perfection and fairy tales. You don't have to like it, you can rail against it the rest of your life, and fight the good fight; and Android will continue to dominate, and it will have absolutely nothing to do with issues related to fragmentation.

View on HN

> Windows has always suffered fragmentation, and has rarely been better than 85% good enough. Circa 2002 they had '95, 98, ME, and XP spread across over half a billion units.

Let's ignore the heroic lengths Microsoft went to ensure compatibility between Windows versions.

You're wrong, anyway.

The cell phone industry is nothing like the PC industry. Whenever you find yourself using a PC analogy to explain some part of Android it means you're getting it completely wrong because it demonstrates an utter lack of understanding of how companies like RIM, Nokia, and Samsung have come to dominate the cell phone industry over the years.

Of course fragmentation is not a problem for the cell phone industry. Fragmentation is a crucial property of the cell phone industry. Samsung became the largest smartphone manufacturer in the world by providing a large, fragmented, outdated array of devices to the carriers.

How could Android fragmentation possibly be a problem for Samsung or the carriers?

The industry has never been about providing up-to-date software. The older your software, the faster you'll update your phone and sign a new 2-year contract.

However, fragmentation is a big problem for people like you and me, who think the smartphone industry should be more like the PC industry, and it's a problem for Google and anyone else who wants to compete with Apple on Apple's terms.

> Circa 2002 they had '95, 98, ME, and XP spread across over half a billion units. With further varied patch configurations and service packs.

It's worth pointing out that this comparison isn't the same. You could, in 2002, run the latest version of Office on all of their operating systems. With Android, there are situations where even Google's own software (see Chrome on Android) only runs on the absolute latest version.

Further, in 2002, all the big name vendors were generally shipping Windows XP on their machines. With Android, vendors are still shipping devices with 2.3. It's still a good OS, but it has to be frustrating as a non-tech user to understand why your brand new Android device can't run software released last week.

And Microsoft was pushing updates and security patches to all those XP machines. Still are, over 10 years after initial release.

Whereas Android users are lucky if they're still getting updates and security patches after 10 months.

Handset vendors and carriers push security patches routinely (though not commonly, the sandbox architecture makes it a lot less needed than on a desktop OS). It's OS upgrades that are slow. Holes get closed.

The article links another article which shows, graphically, how long a number of popular Android phones were getting updates - including security patches. According to it, Android users are lucky to be getting updates after a mere 10 months. Including security patches. And that's assuming you buy the phone immediately after it hits the market.

Not sure which article you mean. You're saying you're aware of an unpatched vulnerability on an android handset?

http://theunderstatement.com/post/11982112928/android-orphan...

There are many, many unpatched vulnerabilities on Motorola Milestone XT720 (and many others). Google doesn't even release security patches for old releases (until recently)--their opinion seems to be that all manufacturers should update to Gingerbread or whatever the latest OS happens to be. Google finally seems to have pulled their heads out of the sand or stopped humming with their fingers in their ears recently and made security point release for Froyo a few weeks ago that backported some security fixes that had been in Gingerbread for a long time, though.

Can you flesh out your statement a bit?

You distinguish between security updates and OS updates. I'm not aware of any difference, an update is an update. I've never seen an OTA update that wasn't a point release. How are these deployed differently?

My old SE phone got random OTA updates regularly, but non of them changed the android version number. I have no idea what those updates where though.

It does not change his point. Many Windows users are not regularly applying security patches or upgrading to the latest Windows version, making the Windows ecosystem fragmented, yet this did not prevent it from becoming a market leader on desktops, with a large Windows development economy flourishing and producing apps compatible with many different Windows versions and service pack levels.

For the same reason, Android fragmentation is a non-issue.

or upgrading to the latest Windows version, making the Windows ecosystem fragmented

No, this did not made the Window ecosystem "fragmented" in the sense that Android is. Windows always had excellent backwards and forward compatibility, which meant you could develop for XP and still run on Vista and vice versa. You just couldn't take advantage of the latest and greatest APIs, but most apps used the classic APIs anyway --you still had access to the whole of the OS and hardware across all Windows versions, even with a custom built C API or MFC.

On Android devices you are cutoff from using tons of features if you can't deploy to the latest versions. And you have to walk a very fine line to make it compatible with all versions/form factors, and mostly cut corners.

For the same reason, Android fragmentation is a non-issue.

A, the classic "lalalala, fingers in the ears" argument.

     which meant you could develop for XP and still
     run on Vista and vice versa

That's not the case. A lot of apps and drivers broke on Vista.

You should really go and read The Old New Thing, as it's really enlightening: http://amzn.to/wW0Okn

Android too offers excellent forward compatibility! It is easy, as API changes are generally additive. I challenge you to find a few different apps written for, say, Android 1.6 that don't run on 4.0. I am sure some examples exist, but they are rare. In fact, I would even argue that Windows apps fail more often to be forward compatible than Android apps, because of the sheer complexity of the Windows software stack compared to Android.

As to backward compatibility, just like on Windows, it is the developers' responsibility to ensure he refrains from using APIs that are too recent. For example, if you write a Windows Direct3D 10 game, it won't run on XP because this API is only available on Vista and up. Same thing for Android: select the right API level when developing your app! http://developer.android.com/guide/appendix/api-levels.html How is Windows better/different than Android in this respect?

I speak from experience. I have developed on both Windows and Android, contrary to most people in this thread.

My understanding is that Chrome on Android is still very much in Beta and they are intentionally doing a very limited release while the major bugs are being worked on. Isn't IE9 only available for Vista and 7 while XP still has neck and neck market share with 7?

To be fair, when iMovie launched (an Apple app) it only worked on the brand new iPhone 4. If you had a 3GS (i.e., hadn't upgraded yet), you were out of luck. It didn't matter that you had the latest iOS. You didn't have the latest hardware so you couldn't run it, although oddly you accidentally buy it through the app store.

The 4 was 50% of iPhones after a year or so :)

Chrome is a beta released three weeks ago. And while I like it, it's a long way from being production quality. Are there any other examples of ICS-only software? Anywhere? Really the only serious bump in the Android API has been Froyo. There are lots of apps that require Froyo. Honestly I can't think of a single one (other than trial stuff like Chrome or niche things like handset-specific root tools, etc...) that requires even Gingerbread, much less ICS.

I thought that was part of what Frasier is getting at, fragmentation gets you both coming and going.

Both ICS and iOS 5 were released at about the same time, iOS 5 is now the majority OS for iOS devices. On the other hand, ICS is available on about minuscule percentage of Android devices.

Software written for ICS won't work on the vast majority of devices and there is little incentive to write for ICS (which will keep hurting Android tablets).

Developers can start taking advantage of iOS 5 APIs today and incentive is growing. At this rate, iOS 6 will be the majority OS on iOS before ICS gets that title for Android.

Which is a valid point, I think. New features don't get adopted quickly in Android apps because of the need to support legacy handsets (one big one was the gesture support that landed in Gingerbread -- Froyo-compatible apps can't use it, or must use it only for non-critical functionality).

But no, that was clearly not Frasier's point at all. "With Android [...] even Google's own software [...] only runs on the absolute latest version." is a complaint that recent software isn't backwards compatible. And that's just wrong.

Except that you can build apps that are optimized for ICS and that degrade nicely for older versions / less capable devices.

That's how smart developers always operated. I remember when Id's Quake 1 first appeared ... I could play it on a 2 year old Pentium processor on MS-DOS 6.22, even though Windows 95 was capturing the headlights.

You can start taking advantage of ICS APIs today. It's more work because you have to test and workaround older versions too, but the thing that differentiates Android from other frameworks is that it was built with degradation in mind.

Office is Microsoft's own software. And 2002 is still one year later compared to when XP was launched. Manufacturers generally start using the new version of Android 6-12 months later, too.

The difference between Windows and Android is that even if Microsoft finished XP in 2000, they waited until all the manufacturers put it on new hardware to release it. Google doesn't do that. They release it as soon as it's ready, and then manufacturers have to put it on hardware. It's just a matter of perception of who has it "faster".

But I suppose Google could do the same - finish Android 5.0 this spring, and then only release it in winter with a bunch of manufacturers at once.

As for Chrome, this is really an exception, and I don't think you can point any other example where an app only works on Android 4.0. This happened because Android 4.0 is a pretty big overhaul, and Google didn't want to bother with the legacy, when they know that within a year 50% of Android devices will have the 4.0 version. They just wanted to take advantage of all the new API's and not have to find workarounds for the other versions. It's kind of like Microsoft not wanting to make IE9 work on XP anymore, because it's a much different version than Vista and 7.

>The difference between Windows and Android is that even if Microsoft finished XP in 2000, they waited until all the manufacturers put it on new hardware to release it. Google doesn't do that. They release it as soon as it's ready, and then manufacturers have to put it on hardware. It's just a matter of perception of who has it "faster".

Is the XP timeline an idle speculation on your part(as seems to happen with your posts about MS) or do you have references to back that up? You couldn't be wronger.

Microsoft regularly shares builds with OEMs so that they can start optimizing their hardware and even publishes beta versions to the public so even small developers can start making their software compatible.

What do we get with Android? For something that's supposed to be open source, the source gets thrown over a wall at release. One OEM does get a head start at the expense of others though.

Microsoft just released Windows 8 developer and consumer previews, did Google do that with ICS?

>It's kind of like Microsoft not wanting to make IE9 work on XP anymore, because it's a much different version than Vista and 7.

IE8 came up almost 10 years after XP. Not to mention that Vista and Windows 7 were on atleast 40% of mahines then and selling on almost 100% of machines. With ICS the numbers are 1%.

> With Android, vendors are still shipping devices with 2.3.

How is this bad? Do you realize that 4 months ago, 2.3 was the very latest Android version available for phones? Its successor, 4.0, was barely released in Nov 2011 (3.0 was for tablets only). A few months delay is perfectly understandable because vendors need time to customize, build, test, ship, and clear inventory levels of devices running the previous Android version.

> You could, in 2002, run the latest version of Office on all of their operating systems.

This statement does not prove that forward/backward compatibility is easy in Windows. How much time did Microsoft waste into making Office run flawlessly across a wide range of Windows versions? With an application the size of Office (30M lines of code), it is almost certain that it has hundreds, if not thousands, of workarounds for limitations or bugs affecting the oldest Windows versions that Office needed to support.

If the only goal were market share then yes, Android is virtually guaranteed to dominate the market for the same reason Windows has been, and that would be a trump card in this conversation.

However, I don't really think the criticisms about Android fragmentation amount to nothing but beating around the "market share" bush. I think that, believe it or not, folks who are criticizing Android mean what they're saying. With all due respect to Groucho, sometimes a complaint that the Android platform's combination of failure to deliver software updates to existing users and lack of a decent wall around the garden raises serious security and privacy concerns is just a complaint that the Android platform's combination of failure to deliver software updates to existing users and lack of a decent wall around the garden raises serious security and privacy concerns.

It's irritating that people blame Android, Google have been steadily releasing new editions of the OS.

Surely the actual problem is the carriers/manufacturers?

The actual problem is the carriers/manufacturers, but Google could have put in requirements regarding upgrade policies. However, they're apparently not in a strong enough position to do that, after all they need to get the manufacturers on board on the promise that they have more freedom with Android.

They could also have made it easier on the manufacturers to do upgrades (ie. giving out continuous updates during development instead of just a source dump at launch time).

On a technical level, yes, but when discussing Android "the product" (aka "the experience") it's the correct way to refer to it.

Android isn't just Google. It's also the carriers and manufacturers.

Spot on! Look at the two big market share holders, Nokia (worldwide & India) and Motorola (US & China). They achieved this by operating on razor thin margins (both, but mostly Nokia) and bending to carriers' every whim (Motorola). See how quickly that market share argument evaporated.

This has nothing to do with his post. Did you read it beyond "What's wrong with Android?" and "Fragmentation"?

He's saying that for his purposes Android is not up to snuff, and then lays out the reasons -- fragmentation was one of them.

There can be an discussions about topics besides market domination. Such as what sucks about Android and how it could be better. An "endless parade" of such discussions doesn't prove that people don't understand what wins a market at all.

Except that the refresh cycle for a mobile device is around two years, and word spreads more quickly than ever that you can't run the latest stuff. If that Android phone can't run the latest apps, lags, or can't do all the things that your friend's iPhone [1] can do, easily, and reliably, you're a lost customer. On the flip side, if you're a developer with limited resources and your app can't run on a vast majority of deployed devices without extensive and expensive testing, you're a lost developer. Neither of these things is good for a computing platform. Never was, never will be. These are first principles. Sure, there might be other factors, (which you haven't specified by the way), but all else equal users and developers want consistency in a platform for the sake of compatibility.

1) Top rated review for Path for Android: http://cl.ly/1P082D1L2D1k1O0i3K0J

> Neither of these things is good for a computing platform. Never was, never will be. These are first principles.

Right, I don't understand why this is so hard to understand.

The nontech users may not know or care what fragmentation means, but they do wonder why each phone the buy has different icons, interface, and even apps, and they do get frustrated when thy can't figure out their new phone because it's so different from the last one.

Windows had tremendous hardware variety, but it never han the inconsistency - even in the same version of the os - that android has.

This article is not talking about what wins a market. In fact, it explicitly disclaims any attempt to apply these lessons to the wider world. It's simply a discussion of where Android falls short as relates to his particular needs. As such, this comment seems awfully off topic.

Fans of the platform seem to hammer on about market share an awful lot.

Stockholm Syndrome?

I really cannot fathom why a consumer, assuming they have no stock in the platform outside their own purchase should be so concerned about something as trivial as market share.

Were the platform in danger of disappearing due to low market share (WebOS) I'd understand that but puffing up your chest because the OS on your smartphone has the best market share seems very strange to me. It's really not a feature of the device.

a) it attracts developers and other content providers, and b) it is a reflection of device popularity (i.e. if a lot of people think it's a good device, there must be something to it).

Except that a) iOS has a dramatically healthier app ecosystem, and b) the poster even points out that most people who buy Android phones don't care about the platform.

Sorry, but the 'Android is dominating' meme took a sever credibility shock with the last round of Android handset manufacturer profit slides. But you know, they'll make it up in volume. Or something.

What does market dominance have to do with profit margins? The two are pretty much orthogonal.

Nor do I understand why fans of Apple products tout Apple's high profit? Isn't that a form of Stockholm syndrom, too? Being happy that the company you buy from makes a lot of money from you?

I wouldn't call it Stockholm Syndrome, but it's something weird. Seems to me that it's the same basic phenomenon as hardcore sports fans. They tie up their identity with this external entity so hard that they almost treat it as an extension of their own ego.

One important difference, of course, is that most local sports teams care about their fans.

I'm assuming that wasn't directed at me-I don't recall having indulged in any-boi-ism of any sort. I also agree with your comment... Being egotistical about something so irrelevant is definitely a bit weird.

There's an excellent line in the middle of the Android Support Visualization article he links to (http://theunderstatement.com/post/11982112928/android-orphan...):

"In other words, Apple’s way of getting you to buy a new phone is to make you really happy with your current one, whereas apparently Android phone makers think they can get you to buy a new phone by making you really unhappy with your current one."

85% is good enough to have someone buy your crap, but it's not enough to make them loyal to your crap.

    >85% is good enough to have someone buy your crap, 
    >but it's not enough to make them loyal to your crap.

I get tired of the fanboistic comments in a lot of these threads. There is a big, strategic, complex, technology game going on and, apparently, a bunch of smart hackers think phones are like simple breakfast cereals (and even those aren't simple).

A big battle is afoot. Act like it. Stop pretending that one side wears white and the other wears black. I always hated Microsoft's products (aging myself: I used OS/2), but they understood the game that was afoot. I'm not anti-Apple. I'm anti-blinders and so many of my friends who use Apple products seem to employ them...

To their credit and only in limited ways are Apple finally acknowledging that network effects are important. But they're doing so in a world quite different than the one in which Microsoft thrived. As I said, I hated Microsoft's strategies, so I hope Apple will not emulate them too closely.

EDIT: I appreciate the lack of downvotes... I was hesitant to write such an opinionated bit, but was frustrated.

> I get tired of the fanboistic comments in a lot of these threads. There is a big, strategic, complex, technology game going on and, apparently, a bunch of smart hackers think phones are like simple breakfast cereals (and even those aren't simple).

I also upvoted you, even though I don't think you're quite correct. I personally don't believe that "hackers (...) think that phones are like simple breakfast cereals". Not that I'm a "hacker" or anything, but most of us HN-ers know how difficult is to "make things just work". Apple does that, ie "making things just work".

And, trust me, I'm not an Apple fanboy. I use a Windows 7 machine for development at work, all my websites are hosted on Ubuntu, and the only Apple products I own are my laptop and my phone.

I know it's only anecdotal, but until yesterday I was the very happy owner of an iPhone 3 (not even 3GS). Everything had been working just fine for the last 2 years, apart from a slightly decreased battery-life and Facebook's native app suddenly not working anymore, which was the moment when I realized that maybe it was time for an upgrade. What did I do then? I went straight into the store and bought an iPhone 4, because that way I knew for a fact that for the next 2-3 years I would have no problems in using my phone. I have no such guarantee from Android phones, even though most of them are similarly priced.

Judging from the amount of people on old iPhones and iPads complaining about how every new version of iOS slows things down or is just plain not supported on their older model, that quote sounds rather... imaginative.

So, strangely, for app testing purposes, I actually own one of each model of iPhone and iPad ever made by Apple. The performance, as you said, certainly does vary. However, from my experience, there's been only one really bad moment: the iPhone 3G on iOS 4. I'd put the next worst as the original iPad on iOS 4.3 (it's better on iOS 5, actually), but even that was nowhere near as bad and still completely usable.

However, this is actually an interesting point about Android. While you might not get updates, you can be reasonably certain that it isn't going to break: if an update is going to run horribly on a device and carriers/manufacturers only update a few devices, you're probably safe. In contrast, on iOS, Apple uses code signing to actually enforce usage of the latest release — you simply cannot downgrade.

If you go through the forum threads, you can see that I am not alone in being frustrated by iOS5 on the original iPad. Even scrolling on the home screen is laggy from time to time. I have already disabled Spotlight for all except apps and most notification BS.

I wonder how much of this is actually down to the OS vs the apps.

For example, I have a very cheap Android phone. It's not the best, but it does a good job. But lately I've noticed that some of the apps are running much slower now after some recent updates - Maps, Twitter, Angry Birds and esp Facebook.

All these need to work in harmony to get the best overall experience, but at the same time work well for the lowest common denominator (me in this case!)

An issue I have with his claims:

> A Price Observation All of the even slightly cheaper phones are much worse than the iPhone when it comes to OS support, but it’s interesting to note that most of the phones on this list were actually not cheaper than the iPhone when they were released

The chart really doesn't quote pricing correctly. I'd prefer if he just used the un-subsidized cost. The subsidized costs are meaningless as different carriers charge radically different amounts. I bought a Sprint HTC Hero in Fall 2009. He claims it was a mere $20 cheaper than the iPhone; in reality I was paying at least $15 less per month than an ATT customer with an equivalent plan. Over the contract, that's $380 cheaper.

But that's not $380 that has anything to do with Apple or Android.

That's like saying your new LG TV is $400 dollars cheaper than your last Samsung TV, because you -also- switched your Comcast plan to pay less monthly.

That's not a fair comparison.

Again, that's why I prefer comparing unsubsidized costs. My Hero may have been hundreds cheaper but perhaps some of that cheapness was from being on an inferior carrier.

Still not a fair comparison. I guess the only truly fair comparison would be a Total Cost of Ownership, but those can be difficult to accurately determine. Specifically how much you might lose for uninsured breakage, etc.

What are you responding to? Certainly not to the linked article. Can you read? You are completely off-topic.

This is about deployment in schools, nothing else.

Some of these issues are not immediately apparent but might be factors when you're buying your next device. Only time will tell.

>The endless parade of fragmentation discussions regarding Android all prove one thing: a lot of people don't understand what wins a market.

market share is useless if your users are not spending money.

Android is in the same position as Nokia 10 yrs ago. And we all know how much Nokia profited from their enormous market share.

I don't understand the comparison with Nokia. Many people I know still happily use their dumbphones from around that time. It sucks that the economic system punishes this.

Android phones are smartphones that fall apart after two years. Google doesn't care if these get replaced as long as people view ads, vendors know they can sell another device to the same person two years later. Both sides benefit from market share.

And people don't understand that wining a market means nothing and it can even be a race to the bottom, it's wining the profitable part of the market that matters.

Was this written in 2001 or something? It's like the PC continues to dominate but the industry is dying for profit margin, the cool kids have all moved to OS X and/or Linux (go to any university campus or programming conference), and all the majority of the money and mindshare go to the ecosystem that's not PC based...

>And people don't understand that wining a market means nothing and it can even be a race to the bottom, it's wining the profitable part of the market that matters.

By that metric IIS is whomping Apache and Nginx. But you don't see that discussed around here. Only the metric that suits people's affiliations and biases seem to really matter here.

Android will continue to dominate? You mean, it will continue to win most activations so long as it lets any budget device count?

Because what other metric is Android winning? Apple captures the lionshare of the entire industry's profits. HTC and Moto are barely breaking even on Android devices.

Google can't even get users past 2.3, no one is using 4.0, and they're already on track to have 5.0 out in what, three months?

Some "domination". I wish I understand what metrics people used to determine these things.

I guess it comes down to "activations". Which, when you get down to it, is a rather poor indicator of platform strength.

Downvote and don't reply all y'all want.

Using activations as a metric is bad.

"India is dominating America because they have far more human activations (births) than America!"

There's a good analogy as to why 'activations' do not tell a complete story.

Please reply with what other metrics people use when saying that Android "dominates". I'd honestly love to know. Otherwise, it's nothing more than a mirror of rose-tinted Apple fanboys, it's Android fanboyism.

"The endless parade of fragmentation discussions regarding Android all prove one thing..." All this proves is what an irrational fanboy you are. What does "winning a market" mean? Money made or numbers sold? Customer satisfaction? Ongoing revenue? As with all fanboys you concentrate on the figure you think is in your favour: market share (no mentions of which of which seem to include returns). Apple have the numbers that matter, revenue, profit, customer visibility (show me the queues of people waiting to get the latest - soon to be out-of-date - Android handset). As of right now to claim that Android is winning a market (remember markets exist to make money) is almost pathetically deluded.

I would suggest that for the most part people buy an iDevice because that is specifically what they want, and for the most part most people buy Android devices because the handsets are cheap or are being pushed by the salesman.

You also ignored the rest of his post which detailed other problems with Android and the Android ecosystem.

Reply on news.ycombinator.com