The Smartphone Wars: No bump, no glory

The April 2011 figures from comScore are out, and I’ve added them to my time-series plots. There are no surprises here, which is very bad news for anybody not an Android fan.

Android’s market-share and userbase growth is bucketing along at the same headlong pace it’s been hitting since June 2010. Bullets can’t stop it! Flamethrowers don’t faze it! History shows again and again how nature points up the folly of men…

Apple is still gaining users slowly with market-share trend looking eerily flat. We now get to laugh, again, at the Apple cultists who thought the Verizon iPhone release in February was going to cut off Android’s oxygen. The very best case they could extract from this data is that growth trends have since shifted about a half-percent per month in Apple’s favor, but that’s within statistical noise and the baseline rates of growth are so wildly different in Android’s favor that it can’t be much consolation.

RIM is toast. Well, that’s not breaking news, but the only bit of drama in this month’s data is that it has managed to drop past Apple in both market share and userbase. The lead-balloon landing of the much-touted Playbook surely didn’t help.

Microsoft continues to shed both market-share and users, with this month’s decline (remember, comScore does running three-month averages to smooth jitter) looking a bit steeper than normal.

HP (relabeled, I used to list this as Palm) hangs in there with amazing persistence, holding on to a nearly static userbase and losing share only as the overall market grows.

50% market-share crossover for Android still looks like being sometime in October 2011.

As before, the most interesting thing about these plots is that the trends for different platforms look as though they’re being driven by qualitatively different processes. The largest difference is between Android and everybody else; Android userbase growth looks uniquely close to exponential/logistic. One of my commenters suggested that the data reflects microbursts of exponential growth being flattened to linearity by product-availability constraints; I concur that this seems the most plausible explanation.

Apple’s essentially flat share trend continues to be mysterious. With another month’s data the slight positive change in userbase growth we saw after January looks like it’s sustaining, but this may just be an artifact of faster overall market growth. The overall picture still indicates a product that has found and near saturated its market, with no rapid growth in the offing or any realistic prospect of catching up to Android. But at least Apple is actually seeing some userbase growth; the other three players aren’t getting even that much.

RIM, besides being in a crash dive, looks the most perturbed by short-term changes in the business environment. I don’t have any theory about this.

HP and Microsoft, on the other hand…”gradual inexorable decline” is the phrase that leaps to mind, if very gradually in HP’s case. The CEO of HP is now talking about licensing the platform out to other hardware makers, but since he has failed to acquire the vital clue that closed source will not cut it any more I doubt this maneuver will lead much of anywhere.

We’ve learned recently that Microsoft is making more money from junk-patent extortion against Android handset makers than it is from its languishing smartphone line. Insert obvious joke about bloodsucking parasite here…it means that the revenue consequences of not having a viable smartphone play will be a long time coming home, and the decline may not matter in more than the long-term strategic sense.

Another good month for Android growth, another sucky one for everybody else’s. Situation pretty much normal.

200 comments

  1. I was tempted to post a spoof comment about the significance of the minute squiggles within the plotted data. But it probably would have just been annoying, so I’ll play it straight. What strikes me is the seeming inexorable nature of the market share curves. Well, lines would probably be more apt, because they seem to be on pretty well established trajectories. Treading water doesn’t work too well as a business strategy when your competitor is flying past like Michael Phelps.

  2. >What strikes me is the seeming inexorable nature of the market share curves.

    Yes. This suggests to me that the drivers underneath these curves are at some scale a lot larger than individual product announcements or quarter-by-quarter business news, and not likely to be deflected by events on that level.

  3. The introduction of iPhone on Verizon roughly doubled Apple’s available domestic market for their phone, which resulted in a certain amount of pent-up demand being released, as well as some sustaining share increase.

    The iPhone month-by-month unit change graph historically reflects 3 phases –post-new-product, pre-new-product, no-new-product-imminent, and we were basically in that latter mid-growth mode when the iPhone 4V hit.

    It seems that the new-product announcement of iPhone 4V caused about one month’s worth of AT&T monthly equivalent iPhone spike, in conjunction with a month-by-month doubling of demand.

    In other words, iPhone 4V was placed into a seemingly mature market of roughly the same size as the AT&T market. The Android offerings available on Verizon are deemed “good enough” by enough customers that iPhone 4V is just another competitive offering, not the “must-have” phone that it was when it first appeared on AT&T.

    Of course, we can’t tease out the by carrier numbers all that easily. We know that, in preparation for iPhone 4V, AT&T aggressively discounted older models, and we can assume that AT&T suffered a bit of churn (but it was less than the analysts expected last quarter) due to availability on Verizon.

    But the fact remains — the potential market size doubled (actually slightly more than doubled) instantaneously, and the only instantaneous bump was about an extra month’s worth of supply, and the sustaining bump is merely the previous run rate adjusted for the size of the new total market.

    I think the muted instantaneous bump is entirely consistent with my theory that all the people who absolutely had to have an iPhone had already switched to AT&T and gotten the iPhone. It would have been very bad news for Apple indeed if there wasn’t at least this increase in the run rate, because most of the new acquirers on AT&T, like the new acquirers on Verizon, don’t fall into the category of “must have an iPhone,” else they would have already had one.

    So, when iPhone becomes available on Sprint and T-Mobile, I think we can expect a corresponding increase (or, eventually, slowing of the decrease) in Apple’s run rate, but no noticeable pent-up demand bump, because these carriers represent even smaller markets.

  4. Actually, since the iPhone 4V became available mid-month, we could probably say that the instantaneous bump was about 50% higher — around a month and a half’s worth of the iPhone on AT&T run rate.

  5. I do have an explanation for the death of RIM in corporates. Exchange 2007/2010 is finally taking off.

    If you’re a corporate, your mail server is running Exchange. Until very recently, that meant Exchange 2003. The 2007 release required a major hardware upgrade, and the point where that would get spent was the beginning of the recession.

    There are a bunch of security features in EAS* in Exchange 2007/2010 that bring it very close to parity with RIM/Blackberry for security (most notable, for me, is device authorization which means that a specific device is authorized to connect to a specific user account and no other device can connect to that user account even with the user’s password – this makes the phone effectively a token in two-factor authentication). I’ve just been doing a bunch of security reviews as my employer sells services to banks, and they won’t accept EAS before Exchange 2007, but will accept Blackberry.

    With three years of Moore’s Law, and the relaxation of budgetary restraints as the recession ends, Exchange 2010 is much more appealing, and a lot of people are rolling it out. The security case that was the dominant case for all-RIM policies at corporates is dying because of the improved security in EAS. Add in the price differential (EAS is included in Exchange but you need to buy BES licences for your Blackberry users) and corporates are moving towards a policy of letting users pick whatever phone they like (within a fairly large selection).

    EAS was originally the protocol used by the early Windows Mobile phones to sync with Exchange, but was soon licensed (and later published in a way compatible with a GPL client) by Microsoft, and is now used by just about every smartphone on the market.

    * Exchange ActiveSync, the protocol that phones use to sync email from Exchange servers

  6. @Richard Gadsen –

    This is very good info about phone sync and enterprise-level security – thank you. My University had been stuck on Exchange 2003 until this spring; I had bought a used Windoze Mobile 6 device from a colleague because I had heard that we couldn’t support correct / secure / reliable sync with Android devices. Said “smart´´phone was one of the most useless, clumsy, finicky personal computing devices I have ever owned in my life. (I had only given up on my trusty Palm Pilot when some combination of Exchange / Outlook / Windows 7 upgrades finally murdered it’s ability to sync.)

    Now, with a pure heart and a relieved conscience, I will $#1tcan that item and get a real (‘droid) phone!

    Anybody want a portable Solitaire player – going cheap!!

  7. @John – There are some features of EAS that are only supported by WM6.5 (even Windows Phone 7 doesn’t support everything) but no admin in their right mind would require them. The base feature-set is supported by Android.

    Of course, your University admin may not be in his right mind.

  8. @Eric – “There are no surprises here, which is very bad news for anybody not an Android fan.”

    Methinks your bias and motivations are showing. I really doubt that the end-user experience of any other phones are going to suffer dramatically because of these numbers. It’s only “very bad news” for people with a vested interest in the market share of non Android phones.

    I always find a delicious irony in how much of a fanboi you are in your rants against Apple fanbois. You deny it every time, but your words betray you. Your cultism is every bit as strong as theirs, and that include not being able to see it in yourself.

    Conversely, these numbers are indeed “very good news” for people who’s self esteem is tied into their choice of mobile phone operating system.

    1. >Your cultism is every bit as strong as theirs, and that include not being able to see it in yourself.

      Ah, the whining of an embittered loser. It is sweet unto mine ears.

  9. @Richard Gadsden / John D. Bell:

    Exchange ActiveSync (which I currently use on my EVO 4G), is a feature of Outlook Web Access / Outlook Web App. Just to clarify for Jeff Read. :)

    @esr:

    BTW — HP’s webOS does run on the Linux kernel and has some other open source components. If HP had a half a brain, they’d open source the whole thing. It’s not like HP is a total stranger to open source; most of the updates to HP-UX since 11 have come from using open source components such as xorg, OpenWBEM, Samba, etc.

    1. >If HP had a half a brain, they’d open source the whole thing.

      Yeah, tell me about it. By what I hear about the (considerable) technical merits of webOS, if HP had a half a brain they wouldn’t be stuck down around 5% market share either. Sigh. It is what it is.

  10. I thought we decided that Apple was manipulating their market share to remain at exactly 25%

    1. >I thought we decided that Apple was manipulating their market share to remain at exactly 25%

      That was somebody’s guess, but I couldn’t figure out a plausible way for them to do that even if they wanted to. I mean, how would they project total market growth accurately enough to know how many units to ship?

  11. @esr:

    > By what I hear about the (considerable) technical merits of webOS, if HP had a half a brain they wouldn’t be stuck down around 5% market share either. Sigh. It is what it is.

    What are these considerable technical merits ? webOS is an obscure thing to many of us.

    1. >What are these considerable technical merits ? webOS is an obscure thing to many of us.

      I speak under the possibility of correction by someone who has actually written for it. The report I hear is that it’s structured like Chrome – Linux kernel underneath, and the entire user-visible interface is the browser. Thus the apps are all HTML, CSS, and JavaScript, on the OS services are all accessed through the DOM. This tremendously simplifies life for app developers.

      Really, if what I hear is accurate, I would be just as happy in an alternate history where an open-sourced webOS wins as I am in the reality where Android is winning. Sadly, HP fumbled the open-source part along with everything else.

  12. Yeah, tell me about it. By what I hear about the (considerable) technical merits of webOS, if HP had a half a brain they wouldn’t be stuck down around 5% market share either. Sigh. It is what it is.

    Absolutely. I knew about three weeks after I bought my Pre on the launch day that webOS would never take off; the elegant hardware and software couldn’t make up for the series of consistent fumbles by Palm and Sprint (everything from not having plan upgrades handled correctly to not having the SDK out far enough ahead to have many apps available at launch). webOS is fascinating to me in that apparently users who aren’t techies are nevertheless using homebrew apps and even applying patches, but I’m skeptical that even my next phone will be a webOS one, and pretty much resigned that the one after that won’t be.

  13. What are these considerable technical merits ? webOS is an obscure thing to many of us.

    webOS is a stock Linux system with a graphical UI (Luna) that is entirely HTML 5/JavaScript. This allowed Palm to focus on optimizing Web performance and reap the benefits across the board. All app-state type storage is in sqlite3 databases, I believe originally using a custom API but now with standard HTML 5 local storage. “Native”-type functionality (e.g., actual control of physical devices) is implemented via Java services that are exposed as browser plugins.

    When I say the entire UI is HTML, that’s literal. Adding a UI control to the global WiFi/Bluetooth/airplane-mode menu to turn the camera LED on and off consists of applying a patch that inserts an HTML table row and the JS to interact with the LED service, and early on, when the phone app didn’t display call duration, a patch quickly appeared that inserted an additional table cell to the call log.

  14. WebOS as of last year was C++ classes, and had a lot of really neat stuff for games, but there were some huge holes. They had a contest and no barcode readers and I just finished my embedded qr-code program, then I discovered there seemed to be no way to grab a picture (or video, audio, …), at least not to actually access it. I do hope they fixed that, but opening up the whole system would help. Android is better because the API is complete – you can do anything with it (including any hardware). WebOS is stuck being closed and unless something radically changed, incomplete. Apple would have languished if apps couldn’t access the hardware.

  15. @esr + Christopher:

    Thx for the explanation. It sounds beautiful. I can see how that approach would get the upper hand over android if it was open-sourced.

    Fuck HP. Fuckin morons.

    1. >I can see how that approach would get the upper hand over android if it was open-sourced.

      Yeah. That might have been a better outcome. Sigh….

      (The clowns who think I’m an Android fanboy are, of course, too thick to learn anything from this conversation.)

  16. >if HP had a half a brain

    I have a fairly recent HP desktop system. eMachines are cheaper and better.

  17. It’s been a while since HP was well managed- risk averse, with the bean counters seemingly in charge. The previous CEO seemed to hold the Dilbertian belief that, if only he could cut costs enough he could make a profit without having to sell anything. The only new products or services or expansion all came through acquisition… and the acquired weren’t necessarily pleased about it.

  18. In this case, I’m pointing the finger at Palm specifically. HP hasn’t open-sourced the Luna stack, but otherwise, it hasn’t been the one fumbling the ball here. *sigh*

  19. I have a fairly recent HP desktop system. eMachines are cheaper and better.

    Posting from my HP laptop, which I’m fairly happy with, except that they went for the impractical Applesque seamless bottom panel, which makes upgrading guts a pain. It even came with SuSE preinstalled, for a $100 discount over an identical configuration with Windows.

  20. * If HP had a brain they’d have driven a stake through the heart of HP-UX and went with True 64.
    * If HP had a brain Data Protector wouldn’t suck like a syphilitic crack whore.
    * If HP had a brain OpenView wouldn’t be harder to use than Netsaint^w Nagios.
    * If HP had a brain their monitoring agents (SIM, etc.) wouldn’t, well, see the syphilitic crack whore again.

    You can’t even say “HP makes good servers” anymore because MOST People make good servers, for values of “good servers” that include PC (roughly) combatible hardware.

    HP doesn’t have a brain, they have a corporate culture which is much like that of anything you find grown in a petri dish.

    It takes a lot to make me wish I was back in a Dell shop, but HP manages to do it.

    But I digress.

  21. > (The clowns who think I’m an Android fanboy are, of course, too thick to learn anything from this conversation.)

    Chrome OS fanboi, then?

    Chrome OS being architected not unlike you describe Web OS, only: already open source, and much faster. And you can write apps in python. (And with NaCl, you can write apps in C.)

    Maybe the Oracle Java Suit will be a blessing if it forces dalvik off Android, with a diving save by Chrome OS.

  22. I have a fairly recent HP desktop system. eMachines are cheaper and better.

    *shudder*. This is why I build all my own desktops. No, I don’t save much money, but I get what I want.

    * If HP had a brain they’d have driven a stake through the heart of HP-UX and went with True 64.

    Yeah, you’re right. But HP-UX does have some redeeming qualities. Like…um, okay, well, SAM is piece of crap. The LDAP implementation sucks. The NIS impelmentation sucks…Its Samba implementation (CIFS 9000) isn’t half bad. :)

    If HP had a brain Data Protector wouldn’t suck like a syphilitic crack whore.</blockquote.

    Having admin'd Data Protector, Veritas/Symantec NetBackup and Legato Networker, it beats the pants off of Legato. Data Protector wouldn't be half bad if it had decent support for the capabilities of modern tape libraries and much better command line utilities (ala NetBackup)

    f HP had a brain OpenView wouldn’t be harder to use than Netsaint^w Nagios.

    Both are over-complicated pieces of crap based on SNMP.

    You can’t even say “HP makes good servers” anymore because MOST People make good servers, for values of “good servers” that include PC (roughly) combatible hardware.

    Their ProLiant line runs major Linux distros very nicely.. They beat the pants off of IBM’s x-Series. If you think x-Series are decent, try adding a QLogic Fiber Channel card. Good luck if you have don’t have a clear idea of PC architecture. :)

    But since you took us far off-topic, your punishment is to be locked in a room filled with HP-UX Itanic servers for you to admin, a half-supported tape library and HP Data Protector. Did I mention the ServiceGuard NFS cluster?

    Muahahahah….

  23. Russell: You’re actually old enough to have maybe seen “Soft White Underbelly” play at a small venue. (I did, and I saw BOC play twice (more)).

  24. Well I guess I’m going to be adding to the stats next month as I am now the proud owner of my second Android phone. The first is the EVO 4G which I like as a device but not really as a phone (the fact that it doesn’t work in Europe is a problem since i live in France mostly), but it is also too big. The second is the Geeksphone Zero which I bought last week – http://www.geeksphone.com/en/

    Unlocked, designed to be rooted and, for an unlocked, non-carrier branded phone, cheap at EUR 200 ish. Its also half an inch shorter and narrower which means it fits into a packet comfortably whereas the EVO is definitely of the “are you pleased to see me?” type.

    With luck I’ll have my custom iptables based firewall running on it next week…

  25. 62% of all smartphones in use in the US are either Android or iOS. The sum one year ago was 37%.

    1. >62% of all smartphones in use in the US are either Android or iOS. The sum one year ago was 37%.

      What an artful way this is of sidestepping the reality that Android’s userbase growth rate has averaged just about twice Apple’s.

  26. > (The clowns who think I’m an Android fanboy are, of course, too thick to learn anything from this conversation.)

    So now that you are the target, fanboy accusations are worthless?

    1. >So now that you are the target, fanboy accusations are worthless?

      Well, since you mention it, yes. They’re falsified by my actual behavior. I’m not religiously attached to Android; dominance by any other open-source platform would suit me as least as well, and I am capable of recognizing that (for example) webOS from horizon to horizon might actually have been a better outcome.

  27. esr, I think he’s trying to resurrect the old Avis slogan “We’re Number 2 so we try harder!”

  28. @FrancisT:

    I don’t think you’ll be adding to comscore stats, since they count (by survey) people who carry phones. Unless you turn 13 next month (thus finally becoming eligible to be counted), it doesn’t sound like your status has changed in a way they measure.

    As far as Android activations, it sounds like you have activated both devices and not just moved a sim card, so yes, you will show up in that stat the next time Andy Rubin gets up on stage…

  29. Eric’s not an Android fanboy, true. He’s an open source fanboy. He’s been predicting a calamitous drop in Apple’s marketshare for the better part of a year now, and it’s still not happening. Remember when Antennagate was going to permanently ruin the iPhone’s image?

    Many people — myself included — were dead wrong about the effect of the Verizon iPhone vis a vis Android’s growth. But that does not automatically mean that anyone who disagreed with them is right.

    Shit, go back and read Eric’s original iPad posts. It was gonna be the next Newton, remember?

  30. It does indeed look like Apple is at saturation. I wonder how much of the Android numbers is the big phones that compete with the iPhone (Droid, Nexus, Galaxy, et al) and how much is the smaller–and cheap/free esp. with contract–phones (the little LGs and such) that have no competition.

    I think the sub-$50 market segment is probably a big one, though it doesn’t get much attention. Of course, Blackberry also plays there a little bit; it might be their salvation, if they could put out something modern-looking. But if they did that, they probably couldn’t afford to play there anymore.

    I can easily imagine that graph represents two markets: in the high end, which is saturated, iPhone and the big Androids slug it out for some share of the people newly moving into that market; in the low end, which is not saturated because it’s so new, Android vacuums up everyone who wants a smartphone, but not to pay $300. (Or maybe not, but I’d love to see the breakdown, anyway. Dunno that we’ll ever get it, though.)

    I’ll also note that there are other, smaller, market segments where Android exists alone: ruggedized smartphones, semi-custom devices (such as the Army has ordered), e-readers, embedded systems. All rather small compared to consumer market, but still participating in the “ecosystem”.

  31. @melee:

    I can easily imagine that graph represents two markets: in the high end, which is saturated, iPhone and the big Androids slug it out for some share of the people newly moving into that market; in the low end, which is not saturated because it’s so new, Android vacuums up everyone who wants a smartphone, but not to pay $300.

    Don’t forget that Apple has been playing in the low-end market too, with $0 or $20 refurbished iPhone 3s and $50 new iPhone 3s (on contract at AT&T).

    But as several of us here have been saying for over a year, yes, the mass market is price conscious, and that’s where Android can really shine.

  32. webOS may still have some life in it, the new hardware (TouchPad, Pre3) will be shipping soon and that should be an uptick in sales, especially if the TouchPad can achieve what no other non-Apple device has done and not suck. I know I’ll be looking at a Pre3 as the eventual replacement for my Galaxy S, the user experience is imply better on webOS, the only aspects of my Galaxy S I prefer over the Pre are the bigger screen and better mail and browser apps, the Pre3 solves the screen issue and webOS 3 will solve the other two. Frankly it’s surprising to me that my Galaxy S feels slower than my Pre did on a regular basis despite having more RAM and twice as much processor power.

    As to Apple’s share, they’re not growing explosively but they aren’t hurting in the least either. I’d rate the marketshare news as disastrous for RIM, very bad for Microsoft, poor for HP and good for both iOS and Android. Remember Apple’s maintaining if not growing share in a market where another platform is having explosive growth.

  33. Remember Apple’s maintaining if not growing share in a market where another platform is having explosive growth.

    Yes. It will be interesting to see what happens when the market gets saturated.

  34. Arguably, the strategy at Nokia before Elop seemed to be for the company to slowly head for extinction. Elop’s change of course however may only speed up the rate of decline. Investors seem to be voting with their feet.

  35. At this point, I’m going to officially declare the smartphone market “boring”. Apple can’t wind up the monopolist from here, and nobody else is in any position to even challenge Android. Exact Apple-Android relative share (10-90! 20-80! 30-70!) is a detail. The only thing left is to see if Oracle/Apple/whatever lawsuits manage to kill Android, and that isn’t going to happen in the marketplace but the courts.

  36. An analyst claims that a “working class” Apple iPhone would disrupt Android. Whether he’s right or not (which I’m happy to agree to disagree on), I think his concept of $300 being a “working class” phone price point shows that he’s pretty out of touch, or at least extremely US-centric (which, btw, isn’t where either Android or Apple will be seeing huge growth over the next few years).

    http://www.zdnet.com/blog/btl/a-working-class-apple-iphone-would-disrupt-android-rim-nokia/49727

  37. Apple can’t wind up the monopolist from here, and nobody else is in any position to even challenge Android.

    I disagree. It’s still very much anyone’s game. A reversal of fortunes — in anyone’s court, not just Apple’s — is still possible if someone pulls a game changer out of their hat.Admittedly, it looks less and less possible as time goes on, but, like esr, I think we won’t know for absolutely sure until 4Q2011 sometime, when Android surpasses the magical 50% point.

  38. @Steven Ehrbar:

    At this point, I’m going to officially declare the smartphone market “boring”.

    Not me. Maybe the smartphone OS market, but as Morgan points out, the obvious conclusion isn’t yet fully assured.

    The smartphone market itself, however, should only get more competitive. As companies reach out to conquer the dumbphone segment, price, battery life, phone quality and life, and the ability to be a serious primary computing device will all be good product differentiators.

    You’re right in that it may be a foregone conclusion that in a couple of years we’ll be able to put a five dollar bill into a vending machine and pull out a tiny smartphone that unfurls into a full PC with a 30 inch screen and full-sized keyboard, and has massive compute power and only needs to be charged once a month (by putting banana peels in the reactor).

    Nonetheless, I will view each of those developments as exciting on its own merits…

  39. As far as Android activations, it sounds like you have activated both devices and not just moved a sim card, so yes, you will show up in that stat the next time Andy Rubin gets up on stage…

    I thought android activations were based on new Google accounts not sim cards.

  40. @(not)Andy Rubin: “62% of all smartphones in use in the US are either Android or iOS. The sum one year ago was 37%.”

    @esr: “What an artful way this is of sidestepping the reality that Android’s userbase growth rate has averaged just about twice Apple’s.”

    ESR’s comment notwithstanding, I think there’s still a valid point to the “(not)AndyRubin” original quote above. The market is rapidly moving toward a 2-player market, much like the original PC/Apple duopoly from the late 80’s. While Android is likely to be a much bigger player than Apple, both ecosystems appear set to survive and prosper for the foreseeable future.

    All of the minor players are on track to crash and burn, either slowly or rapidly. It’s very difficult for me to see any alternative future where a 3rd player gets and keeps double-digit market share. Does anyone else here see a niche that neither iOS nor Android is positioned to exploit?

  41. Steven Ehrbar Says:
    > At this point, I’m going to officially declare the smartphone market “boring”.

    Any time an open platform wins over a flock of closed ones it is anything but boring. There shall be dancing in the streets.

    Keep the reports coming, Eric, so we can revel in the ongoing victory.

  42. @Patrick Maupin: “An analyst claims that a ‘working class’ Apple iPhone would disrupt Android. Whether he’s right or not (which I’m happy to agree to disagree on), I think his concept of $300 being a ‘working class’ phone price point shows that he’s pretty out of touch, or at least extremely US-centric…”

    An odd argument; he seems to be suggesting that Apple move away from the premium-pricing position that it has historically used as the centerpiece of strategy. Offering cheap refurbished phones from the previous technology generation doesn’t violate this strategy, but cheap, new downmarket phones would.

    I agree that $300 w/out contract is not cheap by today’s standards and will be downright expensive in six months. If you like at the new Android phones that are getting so much attention over the past few months, their retail prices are typically in the $150 – $250 range. I would say that Apple would need a phone that sold at $200 max to play this strategy, and more important they would need to be prepared to drop that price $50 – $100 six months from now. I don’t see it happening, and even if it did, it wouldn’t “disrupt” Android in the slightest.

    Last time I looked, the U.S. was in a “recession” that is probably better described as a depression. A few extra dollars makes a big difference in a price-sensitive economy with unemployment rates north of 9% and short-term interest rates near zero.

  43. @JonB:

    > I thought android activations were based on new Google accounts not sim cards.

    Could be. I don’t have a clue how it works, other than reports from google correcting Steve Jobs’s assertion that google counts upgrades as new activations.

  44. @Cathy:

    Offering cheap refurbished phones from the previous technology generation doesn’t violate this strategy, but cheap, new downmarket phones would.

    But this might actually be a reasonable strategy for Apple, especially if the cheaper iPhone were easy to distinguish from its more upscale brethren at a glance. The iPhone nano rumors come back to mind. Phones get replaced a lot more often than computers, so if Apple has a starter model to get people hooked on the iPhone way of doing things, that might help with future up-sells. An attempted reprise of the iPod.

    If you like at the new Android phones that are getting so much attention over the past few months, their retail prices are typically in the $150 – $250 range.

    T-Mobile prepaid Comet with free shipping from amazon for $145 and dropping…

    A few extra dollars makes a big difference in a price-sensitive economy with unemployment rates north of 9% and short-term interest rates near zero.

    You’re right on this, although some people might splurge on a phone. But the US smartphone market is saturating much faster than the international market, and the international volume market is going to be price-sensitive even by US standards. If Apple is going to make a Volksfone, more power to ’em, but that ain’t happening at $300.

  45. @Patrick Maupin: “But [offering cheap, new down-market phones] might actually be a reasonable strategy for Apple, especially if the cheaper iPhone were easy to distinguish from its more upscale brethren at a glance. The iPhone nano rumors come back to mind. Phones get replaced a lot more often than computers, so if Apple has a starter model to get people hooked on the iPhone way of doing things, that might help with future up-sells. An attempted reprise of the iPod.”

    That’s an interesting idea. What does the history show us? Do we know if Apple successfully attracted new users with the iPod mini/nano/shuffle that it later succeeded in upgrading to the higher-end models, or if those who came in with the lower-end models tended to stick with them? (I’ve never owned an iPod or even seriously considered purchasing one, so I never really took the time to figure out the differences between the various models.)

    @Cathy: “If you like at the new Android phones that are getting so much attention over the past few months, their retail prices are typically in the $150 – $250 range.”

    @Patrick Maupin: “T-Mobile prepaid Comet with free shipping from Amazon for $145 and dropping…”

    Yep. I just bought an LG Optimus for $200, and looked at the recently-released Samsung Galaxy Prevail for $179. We are seeing a lot of decent-quality phones for $200 or less. I expect we’ll see something similar to the Comet for $100 by year’s end.

    Dumbphones will be dead by the end of 2012 unless peripheral issues like battery life create small niche for them. Nokia waited too long to work out a strategy to transition away from the dumbphone cash-cow business.

    @Patrick Maupin: “The US smartphone market is saturating much faster than the international market, and the international volume market is going to be price-sensitive even by US standards. If Apple is going to make a Volksfone, more power to ‘em, but that ain’t happening at $300.”

    Very true. What does the data show about the fastest growing smartphone markets outside the U.S.? Especially the ones outside of Europe as well? Anyone seen growth rates for Android and iOS in the BRIC nations, or other parts of Asia or South America? I haven’t taken the time to look for such data.

  46. Well, my highly uneducated outlook is this: last time I went to Best Buy, a “free” with 2-year contract Verizon Samsung Fascinate smartphone was available.

    Which means that a consumer on contract with Verizon can now replace a dumbphone with a smartphone for free. Which means that other carriers must surely soon follow or be left behind.

    How exactly is the scenario Eric has been predicting not going to come about within months? Maybe I’m dumb, but I just don’t see it..

    1. >How exactly is the scenario Eric has been predicting not going to come about within months?

      Eric can think of a few ways. Patrick Maupin and I have believed since late January that the numbers indicated that Apple has about saturated its natural market (credit to Patrick for noticing this first). This might explain why Apple market share has been about flat, with a Verizon iPhone bump that’s within statistical noise level if it’s there at all. However, we could be wrong; it might be that this is (for example) an illusion created by supply-chain issues, and Apple just can’t build iPhones fast enough to meet actual demand.

      Another possibility, tied to this one, is that the Apple fanboys are right to believe that Android has been selling mainly as a substitute for iPhones. I think this is wildly unlikely now that Androids have a solid technology lead in 4G support. Furthermore., if it were actually true the Verizon iPhone bump would have been much larger and matched by a dramatic drop in Android userbase growth. But, again, it’s just barely possible if Apple’s problem is really not being able to ship product fast enough.

  47. Oh sorry, it’s only for new subscribers. Which basically means Verizon can now pick off the competition with the offers of a free Android smartphone (Droid Incredible as well.) Still the same scenario – wouldn’t AT&T and Sprint surely have to follow?

  48. A possible cause for Apple’s fixed market share:

    Smartphones sales are limited by production capacity. Production capacity grows exponentially (long term). There are only a limited number of producers and they all grow at the maximum rate (=market rate).

    Apple has only a fixed number of suppliers, so Apple can only grow as fast as these suppliers can expand their production capacity. Which is at the same rate as the market grows.

    Android steadily recruits new producers who outgrow their older markets (RIM Windows, Palm). So Android grows faster than the market.

    Projections are left as an excersise for the reader.

  49. Could be. I don’t have a clue how it works, other than reports from google correcting Steve Jobs’s assertion that google counts upgrades as new activations.

    Maybe i’m making an assumption because the correction from Google also mentioned “only those with Google services”.

  50. “But, again, it’s just barely possible if Apple’s problem is really not being able to ship product fast enough.”

    Hmm…wouldn’t there be widespread shortages of iPhone at the retail level if that were the case? (Are there?) Or can supply-chain problems cool demand of a product to the point where shortages become nonexistant?

  51. “Android steadily recruits new producers who outgrow their older markets (RIM Windows, Palm). So Android grows faster than the market.

    Projections are left as an excersise for the reader.”

    So once it’s a two-man game, Android begins to flatten due to supply constraints?

    I see that smartphones are now a quarter of the cellphone market. It’s safe to say quite a few of the holdouts will never pay iPhone kind of money for a phone now that free w/contract Androids are available — that’s a basic marketing advantage that will be tough to overcome. So you’re killing Apple at the lower end with free phones (call it ‘enticing granny’) and at the higher end with LTE. Brutal.

  52. @JB:

    But, again, it’s just barely possible if Apple’s problem is really not being able to ship product fast enough.”

    Hmm…wouldn’t there be widespread shortages of iPhone at the retail level if that were the case? (Are there?) Or can supply-chain problems cool demand of a product to the point where shortages become nonexistant?

    Supply constraints are just one variable in a very complicated equation that includes the app market potential of near ubiquity, and the coolness factor of non-ubiquity. It’s hard to know exactly the strategy Apple is pursuing, either short-term or long-term.

    Nonetheless, there are “widespread shortages” of the cheap refurb/old iPhones. They appear on AT&T’s website, then disappear a few hours or days later. So when you consider that cost of goods of a new iPhone is under $200 and Apple sells them to the carrier for almost $600 — that’s an amazing margin for this sort of consumer product. It’s entirely possible that Apple is deliberately managing demand on the iPhone 4 by not dropping the price a little bit.

    You also have to consider that they sell iPhone 4s in lots of markets, and can have per-market pricing. This makes the equation even more complicated. They have to come up with the “correct” allocation of iPhones to markets that will give them the desired network effects in each market.

    If I were Machiavellian enough to buy into Steve Jobs’s vision, I would find a job at Apple helping to decide the best strategy for world domination to be very interesting indeed.

  53. @JB:

    I see that smartphones are now a quarter of the cellphone market.

    In the US. And you can’t discount that. Japan used to be where the cool new phones came out that had features that no other country would see for a decade. Now the US market seems to be one of those where a lot of companies ship their product first, to collect higher prices and gauge demand before going global.

    But…

    Smartphone shipments in Q1 2011 were approximately 100 million units (slightly under according to IDC, slightly over according to Canalys).

    I don’t have the exact total installed base numbers at the tip of my fingers, but IDC claims that in Q1 2010, there were only 54 million smartphones shipped globally.

    When you consider that there are over 5 billion cellphones out there in the world, it quickly becomes apparent that (1) the global cellphone market is nowhere near saturated with smartphones, and (2) last quarter’s smartphone run rate was basically high enough to give every cellphone user a smartphone within the next 4 years. So even though a lot of smartphones are upgrades for previous smartphones, replacements for broken smartphones, etc. you can expect that the growth will continue, and within 2 years everybody on the planet who can afford any kind of cellphone will be able to afford one running Android, even if it doesn’t have a big touchscreen.

  54. @esr:

    But, again, it’s just barely possible [that Apple will come from behind and kick some serious Android ass] if Apple’s problem is really not being able to ship product fast enough.

    I think we’re in violent agreement on this point. Barely possible, but in my opinion the most likely scenario has Android-based manufacturers slugging it out over handset features. Of course, not everybody agrees.

    KEYBOARD ALERT!

    .

    .

    .

    .

    Stephen Elop completely disagrees with me, and in his keynote at Qualcomm’s recent developer conference, he gave this assessment of the industry: “the handset battles are over; the ecosystem war is starting.”

    1. >Stephen Elop completely disagrees with me

      Stephen Elop puzzles the crap out of me. The “burning platform” memo was gutsy and brilliant, but every strategic decision he’s made since – beginning with the Microsoft alliance – has been phat stupid. Now he says the handset wars are over, before NoWin has shipped a handset! Makes me wonder what the guy is smoking; he just undermined his own company.

      Article says everybody thinks Nokia’s got three rough quarters coming before they can make a serious play. That’s a quarter too late; Android’s going to cross 50% share in the U.S. before year end and probably worldwide as well, and at that point everybody but Apple can kiss their asses goodbye.

  55. > Makes me wonder what the guy is smoking; he just undermined his own company.

    Actually, it kinda sorta makes sense if you squint and hold your head just right. By the time NoWin has a handset, it will be far, far behind the top-of-the-line Android offerings. At that point, all Elop will have to fall back on is “but the Win 7 Mobile platform is the most wonderful thing in the world, and more handsets will follow.”

    This logic leverages the perception in some quarters that WinMobile is better than Android, and the probability that the change of pace of Android OS development is slowing, but completely misses the probability that Android OS is now good enough, and that, not only the handset capabilities, but also the app capabilities, will keep increasing rapidly. The chicken and egg problems seem insurmountable unless MS and Nokia both make huge, immediate, developments in the platform. Expect Microsoft to start handing out cash to developers like lobbyists hand out cash to Texas legislators…

    1. >This logic leverages the perception in some quarters that WinMobile is better than Android

      Geez. One wonders what kind of mind-share these guys think they own, with WinMobile plus WP7 at 6.7% and falling?

  56. Expect Microsoft to start handing out cash to developers like lobbyists hand out cash to Texas legislators…

    Ding. This is what Microsoft did when it couldn’t win in the high-end CAD workstation market: they started paying CAD vendors to port their stuff to Windows and then started handing out cash kickbacks to large enterprise customers to adopt the Windows ports. Seems like a big cost and hassle for Microsoft for such a small vertical market, but Microsoft is paranoid enough to believe that they must rule every market.

    I expect them to do exactly the same thing for Windows Phone 7, only with a twist. They’ll build or have already built a generic mobile services platform comparable to Google’s, then they’ll start handing out cash to major mobile app producers to port their stuff to WP7 and — more importantly — to add features to their WP7 versions that locks customers in. Whether they can pull this off remains to be seen, but look for Microsoft to attempt to leverage its office suite monopoly to push WP7 phones into corporate environments and its Windows desktop monopoly to push to WP7 into phones at home.

    I doubt that they’ll pull it off, but that’s where they’ll aim.

  57. When my teenage son (who has been an Apple sycophant for years) switched to Android this weekend and calls his new phone “cool” – I know that the Apple iPhone has lost its mojo. And another teenager is now introduced to the benefits of “open” systems. The introduction of Android OS to the mobile phone market was a point of inflection for numerous reasons.

  58. > One wonders what kind of mind-share these guys think they own, with WinMobile plus WP7 at 6.7% and falling?

    Bu bu bu but…..

    Everybody knows and loves Windows! They can’t lose!

  59. “Article says everybody thinks Nokia’s got three rough quarters coming before they can make a serious play. That’s a quarter too late; Android’s going to cross 50% share in the U.S. before year end and probably worldwide as well, and at that point everybody but Apple can kiss their asses goodbye.”

    What is so magic about 50%? Am I reading this too literally?

    I agree they are too late. But I tend to think NoWin in Q4 2011 is too late by about 18 months – 24 months — about when Android took off.

    What do you make of the market analysts that have predicted Window Phone 7 to surpass Android by 2013? Or iPhone by 2015? http://www.pyramidresearch.com/points/item/110509.htm?sc=PP051011_SMRT http://money.cnn.com/2011/03/29/technology/windows_phone_7_forecast/index.htm

    1. >What is so magic about 50%? Am I reading this too literally?

      50% is magic because perceptions of who has won tend to become self-reinforcing above that level.

      >What do you make of the market analysts that have predicted Window Phone 7 to surpass Android by 2013?

      Microsoft has its fanboys too. The idiots.

  60. > Microsoft has its fanboys too. The idiots.

    Some of them are paid to be fanboys. Considering how quickly most of the public and the journalists forget, perhaps not so idiotic after all.

  61. Another undeliberately hilarious article, beamed straight from the Reality Distortion Generator into the brains of the faithful, through the medium of beatweek magazine:

    http://www.beatweek.com/news/8609-iphone-5-2011-precipice-top-five-reasons-to-finally-ditch-the-android/

    Meanwhile, Microsoft’s true strategy is unveiled. First, they’re going to pwn the tablet market, then move down to smartphones. Kinda sorta makes sense, I suppose, since Android doesn’t yet own the space:

    http://www.eweek.com/c/a/Windows/Microsofts-Windows-8-Could-Challenge-Google-Android-in-Tablets-675079/

    1. >Another undeliberately hilarious article, beamed straight from the Reality Distortion Generator into the brains of the faithful, through the medium of beatweek magazine:

      The funniest thing about it is the smell of desperation coming off that prose.

      >Meanwhile, Microsoft’s true strategy is unveiled.

      This article would have been a coffee-through-the-nose experience if I drank coffee. The giveaway that it was either written by an idiot or faithfully reporting idiotic strategic planning inside Microsoft is the burbling about unifying PC with tablet/smrtphone interfaces. Touchscreens != mice, and any attempt to “unify” them would be a dog’s breakfast so unpalatable that even Microsoft wouldn’t ship it.

  62. @phil

    >What is so magic about 50%? Am I reading this too literally?

    50% is magic because perceptions of who has won tend to become self-reinforcing above that level.

    Just to give a single example of what ESR is talking about (more examples exist, this is just one).

    As a developer if there are N players with about the same market share(and assuming all else being equal) then i can pick and choose between who i develop for without much trepidation. There’s potentially equal money there. If there’s a 50% player and several 10-20% players then while i may want to develop for the 25% player, I can’t in good conscience ignore the 50% player.

    All other things being equal some developers will develop solely for the 50% player, but very few will develop solely for the 25% player.

  63. > Another undeliberately hilarious article, beamed straight from the Reality Distortion Generator into the brains of the faithful, through the medium of beatweek magazine:

    Fingers-in-the-ears bleating from fashionistas to fasihonistas. I have a friend in her 50’s who loves her shiny new Android device. “By geeks for geeks” is a load of rubbish from the self-styled fashion elite to make themselves feel important.

    Ok, I’ve calmed down now. Nothing is going to stop the rise of OSS phones OS’s until market saturation is reached. The only way to disrupt Android is with another open system, in which case I’d be looking to Meego before iOS.

  64. esr, I have great respect for you. However, I think that you disappointed me when you wrote: “…any attempt to “unify” them would be a dog’s breakfast so unpalatable that even Microsoft wouldn’t ship it.

    We both know that that’s a pretty high bar. ;-)

  65. @LyleD:

    > “By geeks for geeks” is a load of rubbish from the self-styled fashion elite to make themselves feel important.

    There were lots of funny lines in that article, but that one made me stop and think — what is an iPhone? “By geeks for non-geeks?” When Barbie says “Math is hard” Mattel gets body-slammed, but everybody takes it for granted that Apple acting like their customers are stupid is the best thing. Maybe it is. Maybe they are.

  66. The giveaway that it was either written by an idiot or faithfully reporting idiotic strategic planning inside Microsoft is the burbling about unifying PC with tablet/smrtphone interfaces. Touchscreens != mice, and any attempt to “unify” them would be a dog’s breakfast so unpalatable that even Microsoft wouldn’t ship it.

    Actually thats not entirely true. MS shipped all of XP, Vista and 7 (I think ME as well) onto tablets. And by tablets I don’t just mean tablet convertibles(like what I own) but also “slate” configuration Tablet PCs which are very similar in form factor to iPad. On screen keyboard, gorilla arm issues etc… I think you underestimate Microsoft’s ability to ship unpalatable (they shipped ME afterall).

  67. Russell: Unless I’m mistaken, that’s only for 3G and 4G smartphones (at VW Online.) There’s also a $10/mo. 75MB package for occasional data users.

  68. Patrick: so there’s “peak” demand for a $150 pre-owned iPhone 4 but not for $200 new one? I’m unconvinced that the Apple fanboys are quite correct then.

  69. What do you make of the market analysts that have predicted Window Phone 7 to surpass Android by 2013? Or iPhone by 2015? http://www.pyramidresearch.com/points/item/110509.htm?sc=PP051011_SMRT

    I just love the windows mobile line (I assume it is anyway… could they have picked worse colours?) that just magically explodes skyward faster than android is now (and i’d be suprised to see someone argue that Android is getting pretty awesome growth numbers right now) when both elements of NoWin have sod all market share and decreasing. I know… i know, worth more than the sum of the parts and all that…

  70. @jonb> All other things being equal some developers will develop solely for the 50% player, but very few will develop solely for the 25% player.

    … Because everyone developed for Windows, and very few developed for linux, even though it was below 25%, and Microsoft was > 80%.

    The scales, they fall from my eyes!

    I’m waiting to see the open sores fire up their copy machines after tomorrow. Android fanbois are still lost in the handset mentality, while Microsoft and Apple are moving on to the ecosystem. I don’t expect Eric to understand this, and instead with his usual threats to ban and other bluster worthy of an 8th grade bully.

    But it’s his blog/playground, and he is king of all he surveys.

    1. >Because everyone developed for Windows, and very few developed for linux, even though it was below 25%, and Microsoft was > 80%.

      Comparing developer adhesion to a open-source system with developer adhesion to a closed-source one is likely to mislead you, because those cases aren’t equal – lots of developers will put a bet on an open-source system that they wouldn’t on a closed-source one at the same level of matket share.

  71. About NoWin and Elop’s fairy tales, or speeches.

    Elop is not paid to tell the truth, but to create “shareholder value”. I find it difficult to say which company’s shareholders he is actually creating value for. But he should not tell the truth unless it increases share prices. Especially not when his company is making a nose dive into the ground.

    Investors fear Nokia collapse
    http://www.hs.fi/english/article/Investors+fear+Nokia+collapse+/1135266631344

    The mobile phone manufacturer Nokia’s share price continued to slide in New York on Wednesday after the company had warned on Tuesday that its second-quarter sales and margins would be much lower than expected.
    Never before in the current millennium has Nokia’s future looked as uncertain as today, analysts feel.
    The demand for Nokia handsets is eroding so rapidly that the company is unable to estimate its profit development for the coming autumn.

    (emphasis mine)

  72. @JB:

    > so there’s “peak” demand for a $150 pre-owned iPhone 4 but not for $200 new one?

    I was thinking more of the $50 new iPhone 3s. They sold a lot of those, and could probably sell a lot more. It probably wouldn’t take too long to satisfy the demand for used iPhone 4s at $150 — that’s bound to be a bit thinner.

    > I’m unconvinced that the Apple fanboys are quite correct then.

    Oh, I don’t think they’re correct, and I’m pretty sure esr doesn’t either. We’re just allowing for that tiny possibility. For example, given many Americans’ penchant for ignoring recurring monthly costs, there is no doubt in my mind that Apple could sell beaucoup phones while still making margins that other vendors would kill for, simply by dropping the up-front price of a subsidized iPhone to $0. But they’d have to be able and willing to make a lot more of them.

    The question is: are they not willing to make a lot more of them, or are they merely not able to make a lot more of them?

    I come down mostly on the side of “not willing”, but if you go back and look at the raw data on the shipments when iPhone 4 started (which you can see fairly well on Eric’s graph page) it looks pretty obvious that at that point in time (last summer), they were “not able”. Maybe they still are simply “not able”, but it’s a bit harder to tell that the way the numbers look now.

    In any case, unless and until the supply chain fairy waves her magic wand and makes whatever problems Apple has disappear, there is no practical difference between not willing and not able.

    If Apple releases a differentiated (cheap) product in the fall, showing that it was a case of “not willing” because they didn’t want to cannibalize their high-end market, that will be an interesting data point, but it might be too little too late. It really does seem that Android is “good enough” for a lot of users, and the longer Apple puts off trying to woo Android users and the better Android gets, the harder the job they have in front of them.

    There are four differentiators between this and the iPod market: (1) music is not a big deal — either people also have an iPod and don’t want to waste phone battery on music, or they’ve gotten more comfortable moving music to and from their Android phone. Either way, you don’t see articles where people say “Android is fine, but I need an iPhone because of my music; (2) the UI is familiar enough and there are plenty of people around who can help — every off-brand Android phone works pretty much just like every other off-brand Android phone, not at all like off-brand MP3 players; (3) branding no longer works solely in Apple’s favor. Everybody knows what Android is, google still has plenty of positive name recognition, and there is plenty of advertising supporting the non-Apple alternatives; and (4) people who buy phones to do things other than make calls and surf the web are interested in the actual specifications of the phone such as CPU, memory, screen size, battery life, etc. Apple will seldom if ever be able to compete with the highest-specification Android devices again.

    So, they will lose the low end by not being willing to sell cheaply enough, they will lose the high end by not bringing out high-end replacement products fast enough, and they will eventually be relegated to hard-core Apple fans and people who want or need an iPod/phone combo.

  73. A new entry for bullshit bingo:

    @not(Andy Rubin):
    …while Microsoft and Apple are moving on to the ecosystem

    As required for Bullshit Bingo, I have no idea what this could ever mean. How do you move into the ecosystem?

    Lying on the grass, is lying on an ecosystem. Just as it lying in bullshit (the stuff found on farms) is lying in an ecosystem. Albeit for a micro biologic ecology.

  74. > Investors fear Nokia collapse

    So MSFT can acquire them on the cheap.

    @Winter: you’re soaking in it.

  75. @not(Andy Rubin):

    > Android fanbois are still lost in the handset mentality, while Microsoft and Apple are moving on to the ecosystem.

    Apple’s certainly (so far) delivering shareholder value with their strategy, but despite your denigration of handsets, Microsoft is going to have to get Nokia or HTC or somebody to shift a lot more of those before any “ecosystem” benefits accrue to them or to their conscripted partners.

    1. >Microsoft is going to have to get Nokia or HTC or somebody to shift a lot more of those before any “ecosystem” benefits accrue to them or to their conscripted partners.

      Heh. To rephrase Elop: “Yes, I know Nokia looks doomed, but soon our products will be filled with magic sparkles by the Ecosystem Fairy, and then watch out!”

  76. > Apple’s certainly (so far) delivering shareholder value with their strategy

    Apple makes a ton of money selling more and more Apple stuff to its customers. As you’ve pointed out, they’re wildly successful at it. Apple’s market cap is now higher than MSFT + INTC *combined*.

    Microsoft will acquire NOK (minus the part that makes base stations with Siemens) out of whatever the Finnish equivalent of abject financial failure has happened.

    Google is going to get squashed between the 18 wheeler that is Microsoft, the (tour) bus that is Apple, and the bulldozer that is Oracle.

  77. Microsoft will acquire NOK (minus the part that makes base stations with Siemens) out of whatever the Finnish equivalent of abject financial failure has happened.

    Yeah, that’s not going to cause any antitrust concerns in Finland or with the EU. And while Microsoft has shown themselves capable of making hardware (modulo the occasional blue screen of death), will they really be able to sell enough cellphone games to support their razor/razorblade strategy?

    Google is going to get squashed between the 18 wheeler that is Microsoft, the (tour) bus that is Apple, and the bulldozer that is Oracle.

    Sure, there’s a chance that Google will have to cough up some spare change to Oracle.

    I know that, as a geek, my tastes are not mainstream. Still, I’m not completely alone when I say that Microsoft, Apple, and Oracle have zero that I’m interested in, and I use google every day…

  78. @esr: Stephen Elop puzzles the crap out of me. The “burning platform” memo was gutsy and brilliant, but every strategic decision he’s made since – beginning with the Microsoft alliance – has been phat stupid. Now he says the handset wars are over, before NoWin has shipped a handset! Makes me wonder what the guy is smoking; he just undermined his own company.

    The More paranoid suspicion here in Finland is that Elop is undercover agent whose loyalties remain at Microsoft and whose job is to undermine Nokia. The Conspiracy theory goes along lines that Bill never forgave nor forgot getting snubbed around ’95 and Elop’s job is to get Nokia dismantled so that Microsoft can buy the useful parts and intellectual properties in order to get a kickstart on mobile market…

  79. @Patrick Maupin:
    “The question is: are they not willing to make a lot more of them, or are they merely not able to make a lot more of them?”

    Apple already commands 25% of the production capacity of Smartphones in the world. That is with a fixed set of companies. If Apple want to produce more Smartphones they will have to grow production capacity. If they want to grow faster than the market, they need to grow their production capacity faster than the other producers. Or recruit more producers (read factories).

    Can they do that?

    @not(Andy Rubin):
    “Google is going to get squashed between the 18 wheeler that is Microsoft, the (tour) bus that is Apple, and the bulldozer that is Oracle.”

    I get the impressions that you do not like Google. This seems to cloud your judgment. That might be helped by trying a different search engine. Bing often has outdated information.

  80. I’ll bite:

    @not(Andy Rubin): … Because everyone developed for Windows, and very few developed for linux, even though it was below 25%, and Microsoft was > 80%.

    Yes, exactly. If you compare number of Linux-only products with number Windows-only products then you’ll see that your example is right to the point. Even things like GNOME and/or KDE are ported to Windows!

    Yes, there are some Linux-only packages – but they all target niches where Linux is big for some reason: servers, render pharms, scientic, etc.

    @not(Andy Rubin): Apple makes a ton of money selling more and more Apple stuff to its customers. As you’ve pointed out, they’re wildly successful at it. Apple’s market cap is now higher than MSFT + INTC *combined*.

    Yeah, but Apple earnings are smaller then Microsoft’s :-) So Apple must continue to perform much better only to keep it’s market cap. If iPhone5 will not be run-away success everyone expects market cap will tank (it does not mean Apple will tank too: Microsoft’s price is less then half of it’s peak even as Microsoft brings records profits). But this has little relation to our discussion so I’m not sure why you even bring this up.

    @not(Andy Rubin): Microsoft will acquire NOK (minus the part that makes base stations with Siemens) out of whatever the Finnish equivalent of abject financial failure has happened.

    And it will help them… how exactly? What Microsoft’s hardware endeavor you can name? Sure it produces good keyboards and game consoles, but it’s not a leader there. And if Microsoft will start to produce hardware then it’ll push other producers ever further in the Android camp.

    @not(Andy Rubin): Google is going to get squashed between the 18 wheeler that is Microsoft, the (tour) bus that is Apple, and the bulldozer that is Oracle.

    Everything is possible in the vastness of the multiverse, but probability of that happening in our universe is quite low.

  81. @khllm> Yeah, but Apple earnings are smaller then Microsoft’s :-)

    Barely.

    CUPERTINO, California—April 20, 2011—Apple® today announced financial results for its fiscal 2011 second quarter ended March 26, 2011. The Company posted record second quarter revenue of $24.67 billion and record second quarter net profit of $5.99 billion, or $6.40 per diluted share. These results compare to revenue of $13.50 billion and net quarterly profit of $3.07 billion, or $3.33 per diluted share, in the year-ago quarter. Gross margin was 41.4 percent compared to 41.7 percent in the year-ago quarter. International sales accounted for 59 percent of the quarter’s revenue.

    Redmond, Wash. – Apr. 28, 2011 – Microsoft Corp. today announced third-quarter revenue of $16.43 billion for the quarter ended Mar. 31, 2011, a 13% increase from the same period of the prior year. Operating income, net income, and diluted earnings per share for the quarter were $5.71 billion, $5.23 billion, and $0.61 per share, which represented increases of 10%, 31%, and 36%, respectively, when compared with the prior year period. Diluted earnings per share included a $0.05 tax benefit primarily related to an agreement with the U.S. Internal Revenue Service to settle a portion of their audit of tax years 2004 to 2006.

    > So Apple must continue to perform much better only to keep it’s market cap.

    Apple’s share price has been de-coupled from its earnings for many years. For reasons nobody can understand, Apple’s consistent earnings growth above 70% is not seen as valuable. Apple appears to be valued at a multiple of it’s cash. In fact, at the current stock price, Apple’s cash accounts for around 20% of the price.

    Find that situation with any other stock.

  82. > lots of developers will put a bet on an open-source system that they wouldn’t on a closed-source one at the same level of matket share.

    Nice cover attempt, but you still lost the point.

  83. @Winter> I get the impressions that you do not like Google.

    Only because it’s all smoke and mirrors.

    Long-term, my anger with Google is over their betrayal of the principles of an open Internet and net neutrality. This will harm innovation, curtail access, stifle competition and, drop America’s standing in the global web.

    But right now, at this moment, what is mostly pissing me off, is that they betrayed the open Internet while making a public fucking statement that used a pile of bullshit PR speak and legalese because they figured that was all it would take to dupe peons like you and me.

    There must be some universal law for public companies, at least those in the tech sector:
    the more you steal, er, copy, from others, the more your share price refuses to budge

    Google is an advertising company. That is where all their money comes from. And over the past ten years, everything, yes, everything they have developed, has, in fact, been a copy, usually a poor one, of what others are doing.

    Except for the stuff they just buy outright.

    All of which leads to doctored, artificial search results. Always.

    How does that make them feel, I wonder?

    Is Larry Page too busy running that giant company to care?

    Does Andy Rubin have no shame?

    Should we assume Marissa is smart enough to know that Google has not innovated — at anything — in a decade?

    And is Sergey doing fuck all?

    Either way, Google has discovered Groupon. A small section of the new web they cannot control. Which, clearly, on the ‘open’ web, is not acceptable.

    http://www.youtube.com/watch?v=CjDA_mbkZjE

    1. >Only because it’s all smoke and mirrors.

      There’s a smartphone sitting next to my left hand for which the entire OS is open source, and which is neither locked to a cell carrier nor boot locked; I could install a complete third-party stack from CyanogenMod on it at any time. I can tether with it, and frequently do.

      That’s not smoke and mirrors.

  84. > That’s not smoke and mirrors.

    You keep ranting about your Nexus One, but the N1 was a failed experiment by Google, one the carriers forced them to abandon. (Hey, I own *two*, a T-Mobile and ATT models, both courtesy of Google.)

    Which of these is present in a current version of Android (from Google):

    Rule of Modularity: Write simple parts connected by clean interfaces.

    Nope

    Rule of Clarity: Clarity is better than cleverness.

    Hmm

    Rule of Composition: Design programs to be connected to other programs.

    Nope

    Rule of Separation: Separate policy from mechanism; separate interfaces from engines.

    maybe

    Rule of Simplicity: Design for simplicity; add complexity only where you must.

    Hahahahaha! (No)

    Rule of Parsimony: Write a big program only when it is clear by demonstration that nothing else will do.

    not even worth making the obvious joke.

    Rule of Transparency: Design for visibility to make inspection and debugging easier.

    tee hee (nope)

    Rule of Robustness: Robustness is the child of transparency and simplicity.

    True. Android fails here (too)

    Rule of Representation: Fold knowledge into data so program logic can be stupid and robust.

    I think ‘data’ is what Android is all about. Google’s acquisition of your data.

    Rule of Least Surprise: In interface design, always do the least surprising thing.

    tell me with a clear face that Android follows the law of least astonishment.

    Rule of Silence: When a program has nothing surprising to say, it should say nothing.

    Point for Google!

    Rule of Repair: When you must fail, fail noisily and as soon as possible.

    I’m not sure I even agree with this.

    Rule of Economy: Programmer time is expensive; conserve it in preference to machine time.

    Hey everybody, let’s write code in Java!

    Rule of Generation: Avoid hand-hacking; write programs to write programs when you can.

    No comment

    Rule of Optimization: Prototype before polishing. Get it working before you optimize it.

    point 2 for Google!

    Rule of Diversity: Distrust all claims for “one true way”.

    Unless they come from Google. “Dear handset manufacturers, all your codebase are belong to us!” — Google

    Rule of Extensibility: Design for the future, because it will be here sooner than you think.

    Chrome OS: check
    Android: not so much

    1. >You keep ranting about your Nexus One, but the N1 was a failed experiment by Google, one the carriers forced them to abandon.

      You’re funny. I was actually referring to my T-mobile G-2, which was a pretty big success. But you’re also funny because the “failed” N1, even more than the iPhone, set the design pattern that later handsets imitated.

  85. @esr
    “There’s a smartphone sitting next to my left hand for which the entire OS is open source, and which is neither locked to a cell carrier nor boot locked; I could install a complete third-party stack from CyanogenMod on it at any time. I can tether with it, and frequently do. ”

    Same over here.

    @not (Andy Rubin)
    “Long-term, my anger with Google is over their betrayal of the principles of an open Internet and net neutrality. This will harm innovation, curtail access, stifle competition and, drop America’s standing in the global web.”

    Eh. And so you support MS and Apple as an alternative to Google? Sound less than brilliant to me.

  86. What I declared boring is the market, not the technology. Lots of interesting evolution happened in PC tech after the mid-eighties . . . but the market structure of clonemakers selling x86 systems running a backwards-compatible Microsoft OS never changed. The fact that the Mac, the Amiga, and the Atari ST did all sorts of “disruptive” things DOS couldn’t match in 1985 didn’t stop Microsoft from having until 1990 to get things just halfway right with Windows 3.0.

    I’m declaring that we’ve reached the same point with the smartphone. No single player can seize the market from Android phonemakers any more than any single player could seize the market from DOS clonemakers. A new open platform might be able to displace Android as a technology . . . but the replacement would be open source and able to run .dex files, because that’s what it would need to do to convince the phonemakers to drop Android. Which would leave the market structure unchanged.

    (Events, of course, could prove me wrong, but I don’t give them any strong possibility of doing so.)

  87. > but the replacement would be open source and able to run .dex files

    when all you have is a hammer…

  88. > No single player can seize the market from Android phonemakers any more than any single player could seize the market from DOS clonemakers.

    Fundamental difference: Microsoft got paid for DOS.

  89. Fundamental difference: Microsoft got paid for DOS.

    And Google’s donating Android to the world out of the goodness of their heart?

    Google has profit motive, and they must either derive some profits from Android or expect to do so in the future.

    (For those non-libertarians in the crowd, profit motive is a good thing.)

  90. @not(Andy Rubin):
    “when all you have is a hammer…”

    Actually, Android is the market leader in mobile phones. More Android phones are sold than iPhones. Sales of windows phones are rounding errors in the market. So, he has a better toolkit than “you”.

    Until something changes quite drastically, this is all just whining from people in denial.

    Android is well under way to capture 75% of the market, 95% is still possible. MS is on their way to to be dismembered, when the shareholder realize MS’ management is wasting good money on a lost case. Just live with it.

  91. @Winter:

    > Apple already commands 25% of the production capacity of Smartphones in the world

    I think you’re confusing Apple’s share of the US market (discussed quite heavily on this board) with Apple’s share of the global smartphone market, which is more like 17%.

    I also think your use of the verb “command” might not reflect the reality. Apple relies on the same ODMs as anybody else. Foxconn, for example, who makes most of Apple’s stuff, still probably sells more stuff to Nokia, although obviously that’s dropping daily.

    Honestly, there’s not that much difference between building a dumbphone and a smartphone, and Apple only “commands” 3.9% of the total global handset market.

    Finally, Gartner thinks dumbphone demand is softening considerably. That must mean that smartphone prices are getting low enough to be competitive for a lot more customers. Will Apple follow or not?

    http://www.rethink-wireless.com/2011/05/19/gartner-reduces-handset-forecasts-2011.htm

  92. BTW, according to that article, Nokia still ships 25% of the phones in the world.

    So if Nokia and MS could wave their magic wands and magically convert all their shipments into NoWin phones without incurring a lot of extra cost, then yes, all of a sudden MS is a contender.

    Last quarter, Nokia shipped 60% more phones than Apple. Even if their smartphone shipments drop by 80%, they still have channel infrastructure and a presumably loyal customer base to sell into, so if they don’t waste too many quarters, they have enough goodwill they can probably claw their way back up to an Apple-sized percentage of the global smartphone market.

    This possibility must be the thing that some of the pro-MS analysts are counting on. It will be interesting to see how it plays out. The analysts are probably not reckoning on the power of open source to keep driving a juggernaut, but some of us on this blog may not be reckoning on the strength of Nokia’s brand name because, well, you just don’t see those in the states any more.

    1. >The analysts are probably not reckoning on the power of open source to keep driving a juggernaut, but some of us on this blog may not be reckoning on the strength of Nokia’s brand name because, well, you just don’t see those in the states any more.

      That is a reasonable point. From here in the U.S. it’s probably difficult to properly evaluate either the overseas strength of Nokia brand loyalty nor the speed with which cheap Android smartphones are eroding it. I’m not sure it’s a less fraught project from Finland, though. Nokia being Finland’s flagship brand, I suspect that it’s very easy for planners living there to unconsciously overestimate the amount of goodwill it has anywhere else.

      So while you’re right to point out that we don’t know, it’s likely they don’t actually know either.

  93. Well, we’ll see what Jobs pulls out of his hat today. iCloud might have some interesting possibilities- some kind of streaming deal with the big media conglomerates might give another chance to pitch DRM as a user-benefit.

  94. @Patrick Maupin:
    “So if Nokia and MS could wave their magic wands and magically convert all their shipments into NoWin phones without incurring a lot of extra cost, then yes, all of a sudden MS is a contender.”

    Interesting possibility.

    That would mean that NoWin comes up with a design somewhere this summer and orders 25 million handsets for Q4/Q1. At ~$200 a piece, this would mean an investment of $5B. Given WP7’s track record and Nokia’s current state, MS will have to come up with some securities up front to get the phones in the shops.

    In other words, MS will have to buy Nokia.

  95. @not(Andy Rubin):

    Rule of Repair: When you must fail, fail noisily and as soon as possible.I’m not sure I even agree with this.

    Of all your statements, I either agreed or disagreed immediately, but this caught my attention as an exception to your commentary. not(Andy), you did a really good job of avoiding personal comments, or at least separating them well … until here. Now you have my curiosity piqued: with your personal feelings about Rule of Repair clearly stated, do you think that Android does a good job of following it’s dictates? [I read your personal comment as meaning to imply you feel that this is not an important UX/UI rule; however, that also would require clarification.]

  96. @Patrick Maupin ‘The question is: are they not willing to make a lot more of them, or are they merely not able to make a lot more of them?
    I come down mostly on the side of “not willing”’

    I would guess right now Apple can meet demand or at least it’s close, I have seen no recent signs of shortages/lines/etc for the iPhone 4. But what possible reason could there be for “not willing?” The only one I can think of is they don’t want to risk over-production, so they intentionally under produce a bit.

    And whoever said MSFT has higher profit than AAPL has out-of-date info. Google is your friend.

  97. @phil:

    I would guess right now Apple can meet demand or at least it’s close, I have seen no recent signs of shortages/lines/etc for the iPhone 4.

    My comment was in the context of a discussion where I was pointing out that perhaps Apple was regulating demand by not dropping price. They have a jaw-dropping margin on the iPhone and a stagnant market share. Perhaps that is what they want, or perhaps they would prefer to reduce their margin to grab more market share, but realize they don’t currently have the build capacity to pull that off.

    For example, if you look at the iPad 2 right now, it is arguably underpriced, in that there have been shortages and more shortages are predicted, so Apple probably could have sold just as many at a somewhat higher price.

    In electronics, it’s difficult to raise prices once you realize that you are leaving money on the table, but it’s quite easy to delay price reductions to keep from leaving additional money on the table, so if a price reduction on the iPhone would cause iPad-style shortages, keeping the price high is the smart thing for Apple to do for the moment.

    One can imagine, however, that they are, through force of will and lots of cash, creating huge build capacity for the upcoming iPhone 5. Since Apple tries to capture a huge amount of value right at product launch, I expect that the iPhone 5 will launch once they have both huge stockpiles and huge capacity in all corners of the supply chain.

    So I think the initial pricing on the iPhone 5 will give us a good clue as to whether the iPhone 4 had been supply constrained. If it remains high, the answer is no, but if they drop it considerably, the answer is “yes, but we’re ready for the world now.”

  98. I think that Apple’s flat share is two things, they’ve saturated the market for people willing to pay $200+ for a phone plus have to have some of the most expensive plans offered in the cell market. I wonder if there is a breakdown anywhere of what sort of plans most new android growth is coming from. The other part is that Android offers nothing to those owners to make it worth switching.

    Full disclosure, one week ago, I walked into a store and walked out with a $200 LG Optimus V and a new prepaid virgin mobile account. Despite the fact that I am and remain an iPhone fan, and the fact that I was willing to spend $200 on a phone, I am not willing to spend $70 / month on a single line of cell service, and while using my unlocked 1st Gen iPhone on a voice only T-Mobile plan was fine, I finally had a need for mobile data, and $25 / month for that is too good to pass up. This whole every smart phone must have a $30 data plan and massive amounts of minutes thing that VZW and AT&T are pushing (and to a lesser extent Sprint and T-Mobile) is a royal pain, and I think Android and Apple have tapped that market and android’s gains are entirely in the space of flexible plan options.

    Now that I have one and have it side by side, I can finally put my finger on the “lag” issue I’ve complained about before (pun not intended). There is a lag, and it appears to be pervasive throughout the android system. Place your finger on any scrollable pane in Android (i.e. the web browser or the apps menu), now just scroll back and forth, don’t flick, just move your finger up and down. There is a slight, maybe 12-15 pixel lag between your finger movement and the display moving. This just doesn’t happen on an iPhone (barring the phone being busy processing something, but Android would do the same, and I’m not counting that against either). When you move a screen element on an iPhone, it is plastered to your finger and it creates a very believable illusion that you are actually moving the on screen element. In android on the other hand, you have the feeling that you are telling the device to move the screen element. It breaks the touch illusion, and is sufficiently annoying that I would gladly switch back to an iPhone if I could get the right plan for it. I am willing to accept that it is possible that this is an issue of hardware, but at the same time, bear in mind that my reference device is a 1st gen iPhone from 2007. The optimus is a new phone released this year. I would expect it to perform at least on par with my iPhone.

    A couple other minor annoyances that are either android problems, or something that I would love a suggestion on an alternative to fix it. The default browser when zooming on a text element rearranges the text to flow in the new visible area, rather than having a minimum and forcing a left right scroll if I zoom any further. This leads to some ridiculous moments such as a wikipedia article having 2 words per line when zoomed in until you get past the side image at the top of the article.

    The browser also stubbornly persists in maintaining the last zoom and page position. If I zoom in on a page, scroll down half way and click a link, it will take me to the new page and keep the zoom and try to scroll me to the same position.

    No quick scroll. Granted the iPhone only has this for scrolling up, but almost every page and list in an iPhone can be scrolled to the top by tapping the status bar. This is especially grating considering that every once in a while for no apparent reason, when scrolling either up or down, the phone will randomly scroll to the top or bottom.

    Inconsistent back button behavior. Most of the time, it will take you back to the absolute last item you were in, even if it was a different app. This is good. OTOH, I have on occasion pressed the wrong button or element which has taken me out of one app and into another. Pressing the back button here will take me back a level within that same app. This is bad.

    Silent mode isn’t completely silent. Programs still start with last volume set unless you manually change it before hand.

    The apps menu is a crappy place for being the only place a newly installed app goes. Unless you remember the exact name of the app you downloaded, you’ll be scrolling up and down that list looking for it. New apps should either automatically appear on a screen page, or there should be an option to sort the app list by recently added.

    Speaking of apps, deleting them sucks a lot. I should not have to install a programs specifically to make deleting apps suck less. Press and hold and delete is a lesson Android needs to take badly from the iPhone.

    Also on the topic of apps I shouldn’t have to download, no timer app built in? Really?

    No android 2.3. Yes I know, I can root the phone and install it, wasn’t the whole big deal that I didn’t have to “jailbreak” anymore to do what I want with my phone?

    Google account syncing. Believe it or not, I don’t like having my entire life on Google’s servers, and enjoy the fact that my iPhone syncs to MY computer and not Google’s.

    Now, I was always told to end reviews on a positive note, so here’s some things that are definitely awesome:

    Notification bar, iPhone needs this bad.

    Navigation, granted I know that newer iPhones have GPS, but I always wanted to have a Google Maps powered GPS, and now I do.

    Google account syncing. Despite my objections above, it is nice to not need $99 / year to push contact and calendar updates. Too bad I can’t sync it with my mac.

    All in all, I have to agree with the assessment that Android is indeed “good enough”, however this is not in my mind a good thing. Being better is how it will best drive competition and the things that are wrong really are an annoying, death by papercuts type of problem which plagues linux too. All in all, I can live with my phone, but I can not wait until the day I can either have an iPhone on virgin mobile, or the other carriers realize that some of us really just want data, and don’t need that much voice time.

    So now that I’m part of the dark side, what stuff should I be getting and doing? Is the rooting of the phone really worth the upgrade to 2.3? And what sort of stuff can be done to backup and preserve the phone before rooting? I know where to get a new image for my iPhone, where do I get a stock image for a virgin mobile phone? Or can I just back the whole phone up as is somehow?

  99. @Winter:

    The delta between the build cost of Nokia’s current average phone and an iPhone is probably more on the order of $100, not $200 (and dropping rapidly). Given that they wouldn’t have all those phones in the channel at one time, and that their ASP should go up by at least $50/phone to start (but also dropping rapidly), and given that they probably don’t (or shouldn’t) mind selling somewhat fewer phones, I think your estimate of the investment required to do this is off by a factor of at least 5, maybe double that.

  100. Sorry for the massive blog entry type reply, but I figured for all the crap I give Android here, I might as well give it a fair overview of my initial impressions having the phone for real too.

  101. @tmoney

    > The default browser when zooming on a text element rearranges the text to flow in the new visible area, rather than
    > having a minimum and forcing a left right scroll if I zoom any further.

    I have noticed differences in different roms’ browsers. In mine, that behavior only happens with a double-tap zoom on a paragraph, not on a pinch-zoom. There may be an option in yours. All the various roms’ stock browsers badly need hardware acceleration, though.

    > Inconsistent back button behavior.

    This is on the app developer’s head. The ‘back’ button is really just a stack pop that removes the top Activity. If App A, when trying to push Activity 2 on App B, instead pushes the entire App B stack, then you will get the behavior you have seen, and you have to pop back down through App B to get back to where you were.

    I have made notes about my opinions of Android’s UI in the past, but after spending several months with it now, I can say fairly strongly that it is the Activities stack paradigm that is the serious winner, once you grok it.

    > Inconsistent back button behavior.

    Blame the carrier and the manufacturer. Or better, complain. : )

    I agree the Apps drawer is… broken might be an apt word. I am not a fan; putting it in page mode rather than scroll mode makes it a little easier to deal with, and my ROM (CM 7.0.3.1) supports customized apps sets, though I usually just use folders on the home screen.

  102. Oops!

    On the third quote, s/Inconsistent back button behavior./No android 2.3./

    Curse Windows requiring ^c for copy…

  103. Nokia went with Windows Phone because Meego sucked, and Google was too stringent in its requirements for Android devices, not allowing Nokia to differentiate as they’d like.

    Funny how from the standpoint of the world’s largest handset manufacturer, Microsoft provided the most open option…

  104. Jeff Read Says:
    > Funny how from the standpoint of the world’s largest handset manufacturer, Microsoft provided the most open option…

    The article you linked all reads like after-the-fact rationalization. Unsurprising considering that Nokia’s stock is getting trounced. If the article contains a germ of truth about Microsoft being very accommodating then its either a) an act of desperation, b) a play to get Nokia on their knees. In either case Nokia has made a deal with the devil.

    And the part about Google being “snobbish” is hardly worth comment. You don’t make bet-the-company decisions on such things. If Google was unwilling to allow Nokia to gut Android, then good for ’em. Otherwise, Nokia made a bad decision for a bad reason.

  105. >Fundamental difference: Microsoft got paid for DOS.

    And the difference it makes is that in that case, the clonemakers were, in fact, vulnerable to a single player—Microsoft. Microsoft arrogantly cutting off sales of MS-DOS to try to force a transition to something like OS/2 1.0 could actually have ended the clonemaker status quo and made space for a new platform. Google can, at worst, make itself as irrelevant to the future of Android as the Xfree86 people made themselves to X.

  106. @Jeff Read:

    > Funny how from the standpoint of the world’s largest handset manufacturer, Microsoft provided the most open option…

    That’s not “funny.” To the extent it’s true, it’s merely a good example of the free market at work. But I’m having a hard time deciding how true it is. It’s my understanding that google’s leverage primarily extends to use of the app store. To the extent this is true, Nokia could have opened their own app store with minimal investment. What Android developer in their right mind wouldn’t also submit an app to Nokia?

    @Steven Ehrbar:

    Google can, at worst, make itself as irrelevant to the future of Android as the Xfree86 people made themselves to X.

    Google could do much worse than this. Although it seems quite unlikely, Google could decide not to open up future versions of Android. Unlike the Xfree86 team, google could throw some pretty massive resources at its version of reality.

  107. At the WWDC, Apple’s just claimed over 200m iOS devices sold and claims it has almost half “the market.” Whichever market that is. Anyway, they’re saying 44% iOS, 28% Android. Sounds… fishy, but I guess they’re including the tablet market too.

  108. Oh hell, and now they have had to pretty much admit they’re in a genuine battle with Android. Their new iOS notification? A “Notification Center” you access by… dragging down from the top of the screen. *munches popcorn*

    1. >Their new iOS notification? A “Notification Center” you access by… dragging down from the top of the screen. *munches popcorn*

      *snrk*

      So it’s come to this, then. Apple is now chasing Android’s taillights even when it comes to UI. Savor the multi-leveled irony.

  109. I have got to start wishing for things more often. That’s the 3rd time this year I’ve wished for something and the same day it is announced that thing is happening.

    >This is on the app developer’s head. The ‘back’ button is really just a stack pop that removes the top
    >Activity. If App A, when trying to push Activity 2 on App B, instead pushes the entire App B stack, then
    >you will get the behavior you have seen, and you have to pop back down through App B to get back to
    >where you were.
    >I have made notes about my opinions of Android’s UI in the past, but after spending several months
    >with it now, I can say fairly strongly that it is the Activities stack paradigm that is the serious winner,
    >once you grok it.

    I definitely get it, and I do like it, especially with the number of apps with ads that get in your way, it’s useful to be able to say “go back to where I was”. Unfortunately, as I said, it’s inconsistent and doesn’t always do this. Also, too many developers lean on it as a crutch for app navigation too (i.e., the browser, where half the time it means go back one page and half the time it means go back to whatever app brought me here).

    I don’t have any good suggestions for how you unbreak this other than trashing the button completely, and I’m not so sure that’s a good idea.

  110. @tmoney
    > Also, too many developers lean on it as a crutch for app navigation too (i.e., the browser

    UGH, total agreement here. I cannot get used to that hard functionality shift at all.

    The problem is it makes sense in both contexts, but not when they have to operate together. But, really, the stocker Android browser (under whatever guise) is pretty sub-par and needs a lot of work.

  111. @jsk:

    At the WWDC, Apple’s just claimed over 200m iOS devices sold and claims it has almost half “the market.” Whichever market that is. Anyway, they’re saying 44% iOS, 28% Android. Sounds… fishy, but I guess they’re including the tablet market too.

    Actually, that number also includes over 60 million iPods. See, e.g.:

    http://www.asymco.com/2011/04/19/review-of-apples-unit-numbers-released-in-legal-filing-prior-to-earnings/

    To be fair, by now “over 200 million” is probably closer to 210-220 million, but still, I stand by my point from a few articles back that it is somewhat disingenuous for Jobs to simultaneously claim that 7″ Android tablets are too small to be useful, yet 4″ iPods are highly relevant.

    BTW, last quarter (Jan-March) the total number of Android handsets shipped was, IIRC, around 2 million more than the total number of iPhones and iPads shipped. When Android pads start to take off seriously, Jobs will need to come up with a new way to spin the numbers, because iPod sales aren’t going anywhere — total iPod sales to date are less than 2 quarters of Android handsets at the current run rate, so that extra 60 million units of iPods is going to rapidly diminish in spin value.

  112. @Patrick

    > Actually, that number also includes over 60 million iPods.

    Ack, yeah, ok. I knew I was missing something. I even own an iPod Touch, heh. Don’t use it much anymore, but it does have darn good battery life even after 2+ years.

  113. > So it’s come to this, then. Apple is now chasing Android’s taillights even when it comes to UI. Savor the multi-leveled irony.

    What’s worse for Apple is they didn’t even have much of a strong showing outside of that much-desired update to notifications. The other highlights I saw were support for OTA updates and Wifi iTunes syncing, so no more cables needed. The rest of the new features are just Apple versions of existing 3rd party solutions. All good-looking, but not much _NEW_.

    If iCloud is done right, with it coupled to all iOS devices as well as OSX and Windows, it could end up being the real boon to the Apple world, but I’m not impressed.

    Makes me really think the next phone will be an iPhone 4S or 4G or some nonsense. iOS 5 isn’t expected until the fall, when the next phone is due anyway. There isn’t enough feature jump to justify a whole new form factor, so doubtful it will be iPhone 5.

  114. >So it’s come to this, then. Apple is now chasing Android’s taillights even when it comes to UI.
    >Savor the multi-leveled irony.

    Isn’t this the whole point of competition, that good ideas propagate? Again, see above, it’s not that Android doesn’t have its good ideas, it does and the notification bar was one of them, its that the experience as a whole is a death by a thousand paper cuts. It’s “good enough” but the market would be a whole lot more interesting if it was “better”. The biggest win for consumers would be if Google and Apple were constantly leap frogging each other, certainly the device manufacturers aren’t going to spend money on making Android better.

    1. >Isn’t this the whole point of competition, that good ideas propagate?

      Well, yes. But it’s entertaining in view of the cult dogma that all good things in UI issue from Apple and others merely imitate their surpassing gloriousness. This is ahistorical nonsense, but Apple marketing has succeeded in making it widely believed even by people who ought to know better.

      I said “even when it comes to UI” because Apple had already lost the leads in marketshare, mindshare, handset technology, and networking. UI was the last redoubt.

  115. @jsk:

    > All good-looking, but not much _NEW_.

    The other new elephant in the room might be legal, rather than technical. The agreement with the labels on the ability to combine the selling of music with the music locker. Of course Amazon already started down this path, practically daring the labels to sue them, and google has had its checkbook out for months, just waiting for the labels to offer something worth buying. So, I assume the industry will lurch forward and Amazon’s and google’s offerings will rapidly come up to speed, but it’s certainly possible that Apple applied enough cash to throw a monkey wrench into that assumption.

  116. @tmoney:

    > Isn’t this the whole point of competition, that good ideas propagate?

    Sure, but for every tmoney who just wants google to apply a bit more spit-shine on the project and make everything work together nicely, and then for Apple and Google to spur each other on to ever greater heights, esr is probably hounded by at least half a dozen zealots who insist that Android couldn’t possibly have anything new or useful to contribute to the universe. So allow him his gloat.

  117. Because everyone developed for Windows, and very few developed for linux, even though it was below 25%, and Microsoft was > 80%.

    Thank you for mentioning the best example of this effect for me.

    Leaving aside Eric’s point, there are few true “linux only” products, none of them(that i know of) commercial (while i don’t think i don’t think i need to rely on that crutch, i did imply it in my original statement). Name a major linux project and I can probably find someone who has compiled it for windows.

    Now try the same thing the other way. If it’s commercial software there might be a linux version, Maya 3d comes to mind, more likely you’ll be able to emulate it in wine(doesn’t work so well for most games) otherwise your SOL. If you’re talking about open source then, as eric mentions, all is not equal. it’s difficult to find examples of windows only projects but not impossible. Notepad++ comes to mind.

  118. Well, to be fair some things Apple does will be hard to replicate on Android. The well-known jerkinness of Android comes directly from the decision to use Java. Java’s standard library is insane: simple thing like single call to Printf can create literally hundreds of objects – which then must be garbage-collected. And it’s very hard to do that in 10ms (required to have smooth interface).

    I’ve seen quite smooth interfaces on Android (a lot of games have them) but it’s only possible if you stop using most standard Java classes (Android-supplied classes are usually Ok) and start counting each and every allocated and deallocated object (typical approach is to put all objects in some kind of array to protect them from Java’s GC, then you should “release” said array when time is not critical and ask Android to run GC). Not something you can expect from typical Java Mouseclicker…

    Note: MacOS offers garbage collector, iOS does not. I think eventually, when phones will be significantly more powerful, iOS will finally offer GC and people will explain that it, too, was “stealed” from Android.

    It’s just difference in approach: Apple only includes features if they know they’ll work well while Android offers all kind of features and then makes them werk well few years down the road. One example: it was impossible to create good multitasking experience till last year (hardware was not powerful enough) – thus Android offered bad multitasking and Apple nothing at all… but do you really think Apple had no way to support multitasking in *nix-derived OS?

    P.S. This being said “Notification Center” is clearly was just an omission: you don’t need anything special in hardware to implement it.

  119. > The well-known jerkinness of Android comes directly from the decision to use Java.

    I consider this to a be a red-herring. Java doesn’t help, but the real problem is simple lack of hardware accel in the basic UI. I will be more than surprised if Ice Cream Sandwich doesn’t have ubiquitous HW graphics support ala iOS.

  120. Eric,

    You’re right, Android wasn’t a knock-off of iOS at all.

    Oh, look, original version of Android! bit.ly/jx0exM

  121. > Actually, that number also includes over 60 million iPods.

    We’ve been over this before. iPod *Touch*, not iPod. Even Eric was confused, not understanding that the iPod Touch runs iOS, and therefore counts in the total number. (Ordinary) iPods (iPod nano, etc) don’t ru. iOS.

    200+ million my harpies. Suck on it.

    1. >Even Eric was confused, not understanding that the iPod Touch runs iOS, and therefore counts in the total number.

      That’s not what I recall. It makes no real difference to me what runs iOS and what doesn’t. The correct question is which devices create positive network externalities for which others.

  122. @Non-carbon life form on some strange planet in an alternate multiverse (who occasionally engages in halting, cryptic, tenuous and tendentious communication with this planet):

    > We’ve been over this before. iPod *Touch*, not iPod.

    Which is relevant to the point, how?

    > Even Eric was confused, not understanding that the iPod Touch runs iOS, and therefore counts in the total number.

    But I wasn’t and I didn’t. And I specifically mentioned a screen…

    > 200+ million my harpies. Suck on it.

    yeah, 110 million phones (will be surpassed by google this quarter), 30 million tablets (will be surpassed by google sometime next year), and 60 million iPods (will be gathering dust in dresser drawers by year after next). What’s your point?

  123. > >Even Eric was confused, not understanding that the iPod Touch runs iOS, and therefore counts in the total number.

    > That’s not what I recall.

    My memory says: Eric dismissed iTouches due to not realizing they had App Store access, though he did know they ran iOS. Not having App Store would have made them extremely irrelevant, thus the initial dismissal.

    Now, here’s me telling why the iPod Touch will not matter in the long run:

    Cost.

    Right now I have a 2nd Gen iPod Touch, from roughly 2 1/2 years ago. A nice little device, esp. once jailbroken, with good battery life. I spent some time in the App Store, did the iOS thing. But guess what I can do with it now? I can still run any App I had downloaded, and can listen to music and such. But my model was obsoleted with a certain iOS version, and the App Store is outright draconian about not providing old versions of software for previous iOS releases, and locking those old versions out.

    Since I can’t upgrade to the latest iOS, that means it has become a future-worthless dinosaur. There is no upgrade path, no carrier to provide a deal. If I want to keep running the latest iOS apps with it, I have to go buy an entirely new device, and they are not cheap.

    As I see it, the iTouch is doomed to irrelevance for simply not being compelling enough vs. the iPhone, and for many is actually an overall more expensive option considering an iTouch owner will still need a phone solution on top of it. Strategically it’s good for fluffing numbers, and filling a relatively small (and shrinking) niche, and that’s about it.

  124. The funniest thing is the following fact: if you add together number of iPod touches and iPhones you’ll get larger sum (it’ll be more then number of Androids currently sold), but market share story will be different:
    http://www.businessinsider.com/chart-of-the-day-ipod-touch-apple-android-share-2011-4
    If add iPad it’ll be more-or-less flat again :-)

    It’ll be interesting to see how long Apple will be able to add new kinds of devices to keep the market share intact…

  125. So I don’t understand the comment about HP being closed source, when WebOS is the ONE platform that doesn’t need rooting, provides you a full bash prompt out of the box, looks more like a linux distribution than any other phone OS, and not only is the kernel interchangable, many good builds from good (former OpenMoko) hackers are available which stably overclock the Pre’s ARM to over 1GHz.

    I think the biggest hinderance to WebOS is *that* it has only been available on one hardware platform, and licensing the OS is the smartest thing HP’s ever considered about/for it. Nevertheless, Android does have an enormous amount of momentum, and I have no idea if it’ll make a difference in the long run having WebOS available on more devices, even if it *is* more open.

    I just think it’s odd that even today people think WebOS is somehow closed..

    US$0.02++
    -m

  126. >Since I can’t upgrade to the latest iOS, that means it has become a future-worthless dinosaur. There is no upgrade path,
    >no carrier to provide a deal. If I want to keep running the latest iOS apps with it, I have to go buy an entirely new device, and they are not cheap.

    They’re no worse than the carrier upgrade path is. The cheapest (new) iPhone is $200 with contract, the cheapest iPod touch, $229. Refurbs are available as low as $150 from Apple directly, and I’m sure a little amazon-fu could turn up some cheaper than that. If you’re the person who will buy an iPod touch, but not an iPhone, I don’t really see the lack of carrier subsidy stopping you. And FWIW, you shouldn’t be having that much trouble finding apps to run on your touch. As I mentioned, I have a first gen iPhone, and it’s only just recently I started running into Apps I couldn’t use.

  127. Java’s standard library is insane: simple thing like single call to Printf can create literally hundreds of objects – which then must be garbage-collected.

    Depends entirely on the implementation. I’m not familiar with the internals of Dalvik, but I know that the Hotspot VM has been able to eliminate most short-lived heap allocations via escape analysis since somewhere before 1.5. Plus, Android is perfectly happy to send a process a SIGKILL if it doesn’t think it’s important at the moment.

  128. khim Says:
    > The well-known jerkinness of Android

    Myth. Pure myth. I have a shiny new Samsung Infuse 4G that’s smooth as glass.

  129. > Myth. Pure myth.

    Not myth. A very real, though sometimes more or less noticeable, problem. It has gotten better, but let’s not handwave it away. Right now manufacturers are fighting it by using more powerful CPUs, while the software side is getting more and more tuned. But they’re only hiding the real problems…

  130. >Myth. Pure myth. I have a shiny new Samsung Infuse 4G that’s smooth as glass.

    And I have a shiny new LG Optimus and it’s not.

  131. @esr> That’s not what I recall. It makes no real difference to me what runs iOS and what doesn’t. The correct question is which devices create positive network externalities for which others.

    Quoting:

    I did some resarch and found I was operating under a misapprehension. Never having used one myself, I didn’t know iPad touches had access to the app store. That changes the category they belong in for thinking about network effects.

    Presumably you mean iPod Touch, not iPad Touch.

  132. > But they’re only hiding the real problems…

    Survey says! Garbage collection.

    > Depends entirely on the implementation. I’m not familiar with the internals of Dalvik, but I know that the Hotspot VM has been able to eliminate most short-lived heap allocations via escape analysis since somewhere before 1.5. Plus

    Escape analysis can make your code run 110 times faster – if you are a really really bad programmer to begin with :-)

    Here is all you get with dalvik: http://www.java2s.com/Open-Source/Android/android-core/platform-dalvik/com/android/dx/ssa/EscapeAnalysis.java.java-doc.htm

  133. > Survey says! Garbage collection.

    Another red herring. It may have an impact, sure, but what I meant was all the software rendering being done where it darn well shouldn’t be, and where in iOS it’s hardware. If all the boring UI stuff could be offloaded to a GPU, Dalvik could garbage-collect to its heart’s content and no one would notice.

  134. > As I see it, the iTouch is doomed to irrelevance for simply not being compelling enough vs. the iPhone,

    Don’t be silly! It’s Apple’s gateway drug.

  135. The iPod Touch will support iMessage, and if you don’t know why that’s relevant (including to the externality discussion) you really aren’t qualified to be talking about the mobile device market.

  136. “The great task in front of us over the next two years is to lift the experience of the Linux desktop from something that is stable and robust and not so pretty, into something that is art,” Shuttleworth said to applause from the audience. “Can we not only emulate, but can we blow right past Apple?”

    – Mark Shuttleworth, founder of Canonical, maker of Ubuntu, 7/22/2008

  137. > The iPod Touch will support iMessage, and if you don’t know why that’s relevant (including to the externality discussion) you really aren’t qualified to be talking about the mobile device market.

    Indeed. It will support iCloud, as well.

    Apple just changed the game. Again.

  138. IF iCloud is handled correctly, it should be pretty neat. Nothing mind-blowing, but certainly useful and better than anything that came before it. Game-changer? Ehhhh, maybe. We’ll see. Raises the bar for others to build something to compete. That’s fine. By coincidence I was planning on building an almost-identical solution for my home systems this week.

    As for iMessage… eh? Removes some of the friction of moving from device to device, consolidates separate functionality into a single application. Alright. I’ll allow that it gives poor little orphan iTouch a bit more life, but really, if you’re trying to convince that it makes the thing relevant, I’m not buying it.

    As for ‘gateway drug,’ it must not be very potent. I bought one, thought it was pretty neat, then felt so damned alienated by its iTunes umbilical and the overall device’s stupid inflexibility that it pretty well pushed me away from Apple mobile products altogether. I had to borrow a friend’s computer running Windows just to turn the fucker on the first time. My opinion was not changed when work handed me an iPhone 4, either.

  139. >The iPod Touch will support iMessage, and if you don’t know why that’s relevant (including to the externality discussion) you really aren’t
    >qualified to be talking about the mobile device market.

    Actually, this is somewhat important. Let’s go back to what I said about a primary reason in the LG purchase. Virgin mobile offers me unlimited data / text and 300 minutes for $25 / month, no contract. With something like iMessage (or really any of the billions of free SMS services) and a cell modem for data in an iPod touch, suddenly, you have a very viable option for people who just don’t need a lot of voice. Heck, I can already use skype on my LG and make pretty much all the voice calls I want, 300 minutes be damned. If they can get a 3G data modem into an iPod touch, combined with iMessage and googlevoice/skype etc, you now have the “iPhone nano” without having to dilute the price points of real iPhones. The touch literally becomes an iPhone without the phone part, and that might just be OK for a lot of people.

  140. > If they can get a 3G data modem into an iPod touch, combined with iMessage and googlevoice/skype etc, you now have
    > the “iPhone nano” without having to dilute the price points of real iPhones. The touch literally becomes an iPhone without
    > the phone part, and that might just be OK for a lot of people.

    Good point. Big if, but let’s not be surprised if, along with iOS 5 and the next iPhone, Apple unveils the iPod Touch 3G or some such. In which case, yes, iMessage is very well positioned in that device space.

    Pretty trivial to replicate, of course. We’ll see how it plays out.

  141. Yeah, it’s certainly an if that plays on I think too things, whether Apple is afraid such a product would dilute the iPhone too much and whether Apple’s relationship with the carriers would make them pass on such a deal. As to diluting the iPhone, this is Apple we’re talking about, not afraid to kill a popular product in pursuit of something better. And really I don’t think the profiles of the people that would buy an iPod touch even with a cell modem would overlap too much with those that would buy an iPhone.

    As to the carriers, well that’s certainly a challenge. On the one hand, Gruber points out that apparently the cell companies found out about iMessage at the same time we did, so they aren’t above antagonizing the carriers. On the other hand, the US (and a few other markets) are not entirely GSM / SIM card based markets, which means even if they made a carrier agnostic device, they still need to convince the non-GSM carriers to activate them on the network. They might be able to bully Verizon with the 4g agreements, but beyond that? Then again, anything that pushes the carriers closer to being simple bit pushers is probably a net positive in Apple’s eyes, as long as they don’t push so hard the carriers lock them out all together.

  142. The iPod Touch is a very, very nice device. Essentially a Smartphone without the phone. And it is totally beside the point.

    There are 5 billion people that use cell phones. They desperately want the communication functionality. See, eg,
    http://www.rnw.nl/english/article/microcredit-your-mobile

    All people want better communication (ie, the Internet). Apple will not deliver them the connectivity, with or without iPods. Neither will RIM. Nor will any single producer. Android will deliver the required handsets, or another “open” solution. Even MS is more likely to connect the world than Apple. In this game, iPod touch, nor iPad, nor any other tablet matter much. (in the end we will use docked phones)

    And the platform that delivers those extra 4 billion handsets will “win” in the only sense that matters.

  143. About iMessage

    Why Apple’s ‘iMessage’ Repeats BlackBerry Messenger’s Mistake
    http://techland.time.com/2011/06/06/why-apples-imessage-repeats-blackberry-messengers-mistake/

    But iMessage is slated to work only across iOS devices (iPhones, iPads and the iPod Touch) which, while giving Apple users the proprietary messaging service they’ve always wanted, may very well turn out to be a big mistake. In choosing to keep iMessage iOS-only, Apple is allowing third-party companies – like the wildly popular WhatsApp – to stay one step ahead by working across different platforms.

    If this has to save iOS? I already have WhatsApp.

  144. Florian M is at it again. Android doomed!

    Instead of linking to his spin, I link to Groklaw’s PJ who gives some more information. Basically, these are some rotten patents given to MS, much like in the Nook case, but now asserted against Motorola. The USA suing itself back into the 19th century, it seems.

    On Premature Predictions Based on Claim Construction Orders, by pj
    http://www.groklaw.net/article.php?story=20110605163439627

    The last time there were some scary articles about a claim construction order, with some concluding that Oracle’s scoring points against Google in the number of claim construction terms the judge agreed with was a leg up for Oracle, I suggested that it was too early to tell what it meant and that journalists could save themselves embarrassment by just reading the filings themselves and not relying on Mueller or on any lobbyist for what they allegedly mean. Sure enough, afterward Oracle’s claims were so severely cut back by the judge, from 132 to only 3 with a hope of possible adjustment in the future, it was clear that any claimed “leg up” from the claim construction order was illusory, because it will only matter if on the last lap one of the claims that survived is about a patent ruled valid and which turns on a particular construction that Oracle prevailed on. See what I mean? And at this point, no one knows what will happen in that litigation’s last lap or in this one. So I reiterate my advice. Wait and see.

  145. Oh Yeah Christopher, escape analysis wasn’t available prior to the 1.6 release.

    reference: http://java.dzone.com/articles/escape-analysis-java-6-update

    So quit lying in support of your position. The simple fact was easy enough to Google.

    You’re correct on your fact, but you might want to be a little less quick to accuse people of lying when all you know about is a single factual error. I apparently conflated two sections from an IBM paper addressing escape analysis and optimizations in the allocation code.

  146. @Christopher Smith: I’m not familiar with the internals of Dalvik, but I know that the Hotspot VM has been able to eliminate most short-lived heap allocations via escape analysis since somewhere before 1.5.

    Just a simple fact: one of the biggest selling points of Android 2.2 was JIT. Before that it was pure interpreter. Even with JIT you still need to analyze new codepath when it’s triggered and…

    @Christopher Smith: Plus, Android is perfectly happy to send a process a SIGKILL if it doesn’t think it’s important at the moment.

    and can freeze said process before doing that – keeping data around and throwing away all these JIT goodies.

    Java is about the worst platform for the mobile, but if you goal is lots and lots of crappy apps then it’s good choice. iPhone showed that people will be happy to write crap in any language so I’m not sure what exactly Java enabled for Android…

  147. jsk: If all the boring UI stuff could be offloaded to a GPU, Dalvik could garbage-collect to its heart’s content and no one would notice.

    Sadly that’s not true. You still need to finish your GC in 10ms or so. It’s possible to write GC which guarantees that – or so I’ve heard. Anroid JIT is not so sophisticated.

    In fact GC is biggest unfulfulled promise of computer world: I remember about quarter-century ago I’ve seen “cool new Lisp-based development platform” (I don’t remember which one, sadly) which stopped every couple of minutes to GC the world and I’ve asked “well, that’s pretty, but sometimes it stops responding – am I doing something wrong”? The answer was “oh, it’s GC – it’s primitive here because it’s pre-production, but we’ll have cool GC soon”. Since then all general-purpose systems with GC I’ve seen exhibit this inherent problem. One exception: reference-count GC-schemes don’t usually introduce additional latency (but then purist spurn them as “not GC at all”).

    I’m cautiously optimistic, though, because slightly later but similar promise was indeed fullfilled (even if it took years and years): today complex multiple layers of indirection in “C++ style” code are indeed correctly optimized away by the compiler. But GC problem is harder: with C++ you can spend as much as you want to do your analysis (because the C++ compilation is done “offline”), but with JIT and GC situation is much harder.

  148. Android will not die…..

    It will remain a popular smartphone OS and if the status quo remains…
    It will still exceed the Iphone by 2X the volume.

    but that doesn’t mean android as of now is not in a big mess…
    It is a HUGE MESS! And Google knows it!

    First is android is heading to Fragmentation Disaster. When a Os is fragmented it just mean that there are many variation to a certain Os which make them all a bit different.

    but android variation is bigger than Window home vs pro variation.
    That means the android OS..

    There are just many OS! So no real network effect taken place.

    Ms ..apple…HP…Nokia…Samsung knows this very well! They all make their own OS and are very success with them for their given time.

    Any Os can naturally coexist in a market with no unified dominate OS. There is only a small network effect to it,

    It is much harder when there is one Monopoly Os like in the Pc with Windows in 90%+ win the Mac taking the remaining 5%+. A new Os is hard to co exist with that environment because of scaling issue and network effect.

    The other issue with android is Google is not making any money out of it. It was originally deem as a defense against Win Mobile.

    Google still make it’s most of it’s mobile money from Apple’s IOS device from search,browse and ads than from it’s android OS revenue.

    But is spends most of its Mobile Money on the Android and pay Apple a small token.

    But is even more ironic that google still make most of it’s money in the desktop space while Apple now makes most of its money in the Mobile space.

    With that in mind, see how Google is still pushing Chrome OS so hard in the desktop space? when we all know it is a flop?

    why So? Isn’t android successful enough?

    The Money google make in the desktop space is an easy target and cannot be offset with the Money in Mobile space. And the very fact the mobile space is still inherently unstable.

    So what can Google do to improve Android?

    first they must retake android and start to stop/limiting the Open source part of it. You can only control what you own.

    Notice there is no Honeycomb Opensource and how old gingerbread is today and how little an improvement gingerbread compare to froyo?

    By controlling and closing Android, Google will be able to make real money by stopping the fragmentation issue of android.

    but Googles seams to be still confuse and hesitant so much they sack their CEO eric. And afraid of pissing their ally and partner and still play the openness issue.

    Even Apple seams not afraid of fragmented Android. So much that they recently continue to renew the map and search services from Google.

    Apple doesn’t need to do anything drastic against Googles just small moves but with big bait.

    IOS 5 hugh improvement over IOS 4. but rather show Google any improvement in speach, map or UI… apple just solve the notification and wireless sync issue in IOS.

    so that is nothing that google can really copycat apple from IOS5 that it can’t try to copy from IOS 4.Today Google fail to innovate from froyo!

    Cunningly Jobs just sucker punch Google by allowing Google to swing aimlessly at the rumour NFC and payment thing.

    Google became a headless chicken and rush everything out on NFC meanwhile Job smack in Google head and Ko them in the clouds space.

    The Cloud is suppose to be google forte.

    But apple made the icloud both free and AD FREE……WOW
    Thus denying Google any future income from the cloud and making the mobile strategy even worst.

    so status Quo still remain for google.
    Little money on mobile…no money for cloud….and most money from desktop.

    Apple if they wanted they could Nuke Google out of Apple desktop e space. and by no means is the Mac space not a very lucrative space.
    And Safari and Itune in PC space.

    But….

    What really’s apple is afraid is not Android os or Google per se…
    What Jobs is really afraid is the red Machine AKA Microsoft..

    Laugh all you want about Microsoft but in no way are they dead!

    Imagine a unify unfragmented window mobile system twice the size of IOS!

    Would Apple make money from Apps? Itune?? Accessory?? Exclusivity? developer support? Search? Ads? Volume? with this environment?

    So you android fan just remember that Microsoft and windows is not dead either!

  149. First is android is heading to Fragmentation Disaster. When a Os is fragmented it just mean that there are many variation to a certain Os which make them all a bit different.

    Except that everyone who has a reason to know has said that fragmentation at the OS level is non-existent. If there is fragmentation it is at the handset level(Different CPU speeds, different memory models, different Feature-sets). Which i’d like to say kills your point but I couldn’t pick much out of the random spacing in your post.

    Paragraphs. They are your friend.

  150. Not myth. A very real, though sometimes more or less noticeable, problem. It has gotten better, but let’s not handwave it away. Right now manufacturers are fighting it by using more powerful CPUs, while the software side is getting more and more tuned. But they’re only hiding the real problems…

    And yet last time we went down this path, I got a repeatable test from JSK and reported that my milestone at the time (was 2.1) had no issues. It was the first test I did when i went to 2.2 again no issues. It’s now the first test I do on any Android i get my hands on. I’ve yet to see issues. (Galaxy S and Galaxy S II are the phones that come to mind)

    Perhaps the point to be made here is that some phones are just poorly made. So avoid those manufacturers if it bugs you.

    Some points that a small google search gave me:
    * CNet stated that the LG Optimus uses a resistive screen and a quick search around for Capacitative vs Resistive touchscreen didn’t list LG as one of the companies that were improving the gap between the technology responsiveness wise. However Wikipedia says Capacitative so grains of salt apply. Another review also pointed at the touchscreen as being the issue however so potentially cheap parts is cheap parts.

    * It did get roundly panned in the few reviews that showed up early as having responsiveness issues. None of the reviews attributed it to android, they all attributed it to LG cutting corners in the wrong places.

  151. Except that everyone who has a reason to know has said that fragmentation at the OS level is non-existent. If there is fragmentation it is at the handset level(Different CPU speeds, different memory models, different Feature-sets). Which i’d like to say kills your point but I couldn’t pick much out of the random spacing in your post.

    Since no-one else has called me on this, let me call myself just briefly.

    A specific datapoint i was thinking of(the rovio interview) was actually saying that fragmentation at the device level wasn’t a problem, not at the OS level as i’d remembered. But having said that,every complaint thread i’ve seen was talking about issues at the device level with lots of “not the problem” style comments when anyone talks about OS fragmentation.

  152. In regard to the jerkiness problem, I have found it can be a real problem on my Evo if there are too many background sync processes tying up the radios. Tuning them so so the radios aren’t kept constantly going fixes the problem altogether. So maybe some reports are application related rather than OS related.

  153. @khim:

    > (but then purist spurn [reference counting] as “not GC at all”).

    Perhaps because doing reference counting properly doesn’t involve hard questions of computer science, but merely the tedium of putting your increments and decrements at the right places…

  154. Its difficult to believe that Florian is paid for his writing, since it consists of cutting and pasting Microsoft’s press releases without even bothering to improve the grammar.

  155. Perhaps because doing reference counting properly doesn’t involve hard questions of computer science, but merely the tedium of putting your increments and decrements at the right places…

    The difficult bit is trying to make a pure reference-counter deal with circular references.

  156. > The difficult bit is trying to make a pure reference-counter deal with circular references.

    True, but even that seems to have been solved reasonably cleanly these days, although I think some of the solutions to that are CPU time intensive and could cause lag just like mark/sweep.

    It’s interesting to note that, for the longest time, Python didn’t garbage collect circular references, but that didn’t impact most Python programs, and the solution for impacted programs was fairly simple (the occasional explicit object unlinkage).

  157. Patrick,

    > True, but even that seems to have been solved reasonably cleanly these days, although I think some of the solutions to that are CPU time intensive and could cause lag just like mark/sweep.

    > It’s interesting to note that, for the longest time, Python didn’t garbage collect circular references, but that didn’t impact most Python programs, and the solution for impacted programs was fairly simple (the occasional explicit object unlinkage).

    I haven’t investigated those reasonable solutions, but here is one that occurs to me. What if each variable kept an out-of-scope bit, set to true if it became out-of-scope but it’s reference count was not zero. When a variable went out of scope, and it’s reference count was not zero the GC would set it’s out-of-scope bit to true and then check all the refererences to see if it was part of a cycle (using something like Devel::Cycle in Perl for example) and if all the other variables were out-of-scope. If so they would all be garbage collected.

    Yours,
    Tom

  158. > What if each variable kept an out-of-scope bit, set to true if it became out-of-scope but it’s reference count was not zero.

    I could be wrong, but that seems like a potentially huge burden on the normal case to deal with the edge case.

    Typically, if I care (if it’s a long running process where garbage collection might be an issue) I use “weak references” on back references, to make circular references not quite so circular.

  159. Patrick,

    > I could be wrong, but that seems like a potentially huge burden on the normal case to deal with the edge case.

    > Typically, if I care (if it’s a long running process where garbage collection might be an issue) I use “weak references” on back references, to make circular references not quite so circular.

    Yes, if you have a large data structure with frequent addtions, the burden would be huge. For objects it has the advantage of calling the destructor as soon as possible. To decrease this, you could merely add the out-of-scope bit and periodically look for cyclical references which have gone out of scope. This is more like a mark and sweep, but with less work to do.

    Yours,
    Tom

  160. > To decrease this, you could merely add the out-of-scope bit and periodically look for cyclical references which have gone out of scope.

    OK, you confused me by your terminology. If I understand correctly, you’re saying that whenever you decrement an object’s reference count you also set a bit in the object that indicates that it is a possibility that the object is part of a now-unused circular reference. (The bit only needs to be set if you didn’t decrement the reference count all the way to zero, but that’s merely a possible architecture specific optimization.)

    That might be useful, but since the garbage collector needs to then find those objects with your new bit set, and find other objects that don’t have the bit set that might be in a cycle with it (and it could be a huge cycle with multiple sub-loops), basically the garbage collector has a lot of searching to do in any case.

    In Python, most objects (ints, floats, strings) cannot possibly appear in circular references, so this additional code to set and check this bit on these kinds of objects would be wasted overhead.

    Python has several optimizations in this area. The first one (last time I looked) was to keep all container objects in a doubly-linked list so that you can easily scan for and delete these circular references. I think there’s also some generational stuff going on, but it’s been awhile since I’ve looked at any of that.

  161. Patrick,

    No, I meant an out-of-scope bit, set when a variable goes out-of-scope.

    > In Python, most objects (ints, floats, strings) cannot possibly appear in circular references, so this additional code to set and check this bit on these kinds of objects would be wasted overhead.

    Well, in Perl you would really only need to check this bit on variables which contain references, so you’ve optimizes my idea some already. Thanks!

    > The first one (last time I looked) was to keep all container objects in a doubly-linked list so that you can easily scan for and delete these circular references.

    Another optimization which would help.

    Yours,
    Tom

  162. > No, I meant an out-of-scope bit, set when a variable goes out-of-scope.

    But how do you define “going” out-of-scope?

    If I have code that does:

    for i in range(100):
    print i

    what do you consider to be “in scope?” The implicit list returned by range, each i, what?

    If I have code that goes:

    i = 2
    i = 3

    Surely the object first bound to i is now out of scope.

    If I say:

    del foo.bar

    Then surely, you would need to see if foo.bar was part of a circular reference, no?

    Let me phrase my question differently:

    Under what circumstances (other than reference count going to zero when the deletion is immediate) would a decref on a container object not need to have your special new bit set, and why?

  163. Patrick,

    > Under what circumstances (other than reference count going to zero when the deletion is immediate) would a decref on a container object not need to have your special new bit set, and why?

    No, I think you are right. Set on decrement.

    Yours,
    Tom

  164. When programming apps for J2ME, I never noticed slowdowns caused by the GC. I tried to optimise my network code, but some of the rubbish I wrote for GUIs should really have caused the GC to go nuts in an obvious manner.

  165. Nice catch, Michael. The really interesting thing there is that it expands the network: if developing for the iOS is also developing for the Mac, the network is that much bigger. And of course the iCloud API — which works on the Mac and the PC alike — is also an effort in that direction.

  166. > which works on the Mac and the PC alike

    Sorry, I’m bored and feel very snarky tonight. Stop calling systems with Windows installed on them a “PC.” Just stop. It doesn’t mean anything. Say Windows. I was sick of that Apple marketing ploy (can’t actually SAY “Windows” in the ads, can we?) in the late 90s, and I’m still sick of it today.

  167. >Sorry, I’m bored and feel very snarky tonight. Stop calling systems with Windows installed on them a “PC.” Just stop. It doesn’t mean anything.
    >Say Windows. I was sick of that Apple marketing ploy (can’t actually SAY “Windows” in the ads, can we?) in the late 90s, and I’m still sick of it today.

    I could be wrong, but I was always under the impression that the reason we use “PC” to refer to windows PCs comes from the old days of system requirements when software for a windows computer required an “IBM Compatible PC”. Hardly a marketing ploy.

  168. > But, um, why are you blaming this on Apple?

    Because Apple took the term “Personal Computer,” something they used to call themselves, and actively redefined it in marketing to obliquely refer to Windows rather than any specific hardware. It started small, with the early iMac commercials referring to “PCs” having too many cables. I thought it was annoying at the time, because to my mind, ‘wait, how is the iMac not a PC too?’ Over time, especially when the whole “I’m a Mac” “I’m a PC” thing got going, it started to amplify. It’s a very clever strategy that has worked, but it still makes my eye twitch due to the inherent inaccuracies. But really, it’s just snarky nitpicking, so feel free to ignore me.

  169. Well, I do fundamentally agree with you about appropriate terminology, so in my book it’s a good snark.

  170. jsk,

    The term “personal computer” when spelled out can refer to any number of microcomputers intended for a single user: IBM, Apple, Commodore, Amstrad, etc. products were all marketed as “personal computers”.

    However the initialism “PC” has, since at least the mid-1980s, almost exclusively referred to IBM’s line of 8086-family machines and, later, those machines from other vendors designed to be compatible with IBM’s. This is reflected in the names of publications at the time — PC Magazine, PC World, etc. which dealt with IBM compatible products primarily and only briefly mentioned Macintosh, Amiga, or other personal computers.

    These days a “PC” generally means an x86-based computer running some variant of the Windows operating system. What the Apple marketing hides (understandably) is that modern Macintosh computers fall under this definition as well, as they have Intel x86 CPUs and can be configured to run Windows. Hence, Macs now count as “PCs” in addition to having always been personal computers.

  171. @esr “Ah, the whining of an embittered loser. It is sweet unto mine ears.”

    Hoo-boy, you have really lost the thread. Calling me an “embittered loser” really shows how far you’ve sunk lately. In what way am I a loser? I don’t win or lose anything based on the market share of any phone OS. Wait, I recently switched to Android, so does that make me a winner now? No, I’m not a member of your cult, so I guess I’m still a loser.

    You have completely lost objectivity and your screeds have become less rational as time goes on. It’s a bit sad, really.

Leave a Reply to LyleD Cancel reply

Your email address will not be published. Required fields are marked *