The newest comScore report contains some fascinating statistics on the state of the U.S. smartphone market. And this is where I get to intone “All is proceeding as I have foreseen”, because the prediction I made and was duly calumniated for early in the year has come true right on schedule.
comScore is not just tracking new-unit sales, where Android has been winning for about a year. They’re tracking total subscribers, and Android is now at effective parity with the iPhone and (unlike the iPhone) actually growing share at above statistical noise level. This tracks my predictions from 1Q 2010 when Android was just beginning to slam the hell out of everyone else in new-unit share; I said to expect crossover in 4Q 2010 and it’s happening.
And this is before the marketing around the Nexus S hits, which as I pointed out in my last report represents the first direct push into the iPhone’s turf. It will be very interesting to compare Christmas-season sales figures, if we can get them. The iPhone needs to not just beat the Nexus S during the Christmas rush but make it look like an ignominious failure; otherwise, app developers and others making strategic bets are going to have grounds to suspect that Apple can’t even hold its core demographic in the longer term.
Or, to put it another way: if the Nexus S outsells the iPhone over the next thirty days, it’s game over for Apple – their chances of holding a profitable niche at the high end of the market basically evaporate. If the Nexus S and iPhone sell at rough parity Apple has to do something – another product rollout, a multicarrier announcement – to prove they’ve still got game against the relentless flood of ever-improving Android devices. Only if iPhones outsell the Nexus S by a minimum 5%-6% can Apple afford to continue executing on the same product strategy.
Once again, the fundamental problem Apple has is not the Nexus S itself; rather, it’s that they’re stuck in a comparatively fragile single-vendor/single-product strategy, facing an multi-vendor army that in aggregate has more financial mass, more capacity to innovate, and more freedom to innovate. The Black-Scholes theorem applies; a portfolio of options is more valuable than an option on a portfolio. If the Nexus S fails to take core-demographic share from the iPhone over Christmas this is only a reprieve, because the Android army is going to be driven by competitive dynamics to keep making bets until they get that right; on the other hand, if iPhone doesn’t win decisively, Apple’s only bet fails and Apple’s strategy with it.
What’s even worse for Apple is that with the release of NFC capability in the Nexus S the iPhone is now playing technological catch-up. If they fail, Google will probably lock up the market for phones as portable electronic-payment devices over the next year; if they succeed, they’ll be seen to be chasing Google’s taillights, which would pretty much be death for their brand image. And NFC is not going to be the last bullet in Google’s magazine, so the question becomes: how many bullets can Apple dodge?
Eh, I guess it’s all in how you read things. I don’t see Apple in any weak position or serious danger (though I agree their lock to AT&T is detrimental to their position. Rather, I see that one phone (ok, technically 4 phones) on one carrier holds the same level of command as 50+ phones on 4+ carriers. To be honest, that the iPhone share even continues to grow is just amazing when you consider that the number of people who can afford a smart phone and are on or willing to be on AT&T must be more or less tapped out (as indeed these numbers imply). Think about that, 1 phone on 1 carrier holds 25% of the smart phone market.
While I rightfully criticize Android fans who always say “wait for the next model” in response to criticisms, I do find myself tempted to say, wait until the iPhone is on more than one carrier, especially if the previously cited reports on phone/OS loyalty hold any weight.
Bear in mind too that for both Android and iOS, it’s somewhat wrong to speak only of the smart phone market, as both are platforms for mobile computing, and both have installation bases beyond mere smart phones. Indeed, if iOS were to stay at a mere 25% of the smart phone market (unlikely, unless you seriously expect any move to new carriers to not increase iPhone share), that would hardly mean the end of iOS or its influence on the mobile computing market, not with so many iPod touches and iPads out there.
Still, I have to seriously wonder why you keep insisting that Apple must win this “battle” decisively or they lose, while you implicitly acknowledge that if Android fails to accomplish the same, they don’t necessarily “lose”.
Also, I would like to point out that despite my (and your) interest in NFC technology, we’ve seen this dance before, and we know that geek appeal doesn’t mean anything if the technology doesn’t catch on in the mass market, or at least within a market that can be reliably attached to a smart phone with NFC. See previously such wonders as bluetooth contact sharing, and Cue Cat.
Interestingly, the idea behind Cue Cat lives on today in the popularity of smart phone barcode scanners, but this is an example of needing to attach the idea to equipment you can rely on your market to have rather than the technology bringing about a market for itself.
Whether NFC catches on will rely on the banks / credit card companies cooperating, and the merchants, and enough consumers trusting their phone to be their credit card. Those are some significant hurdles, especially with only one model on one carrier currently supporting the tech as, unlike the iPhone, the Nexus S doesn’t command 25% market share on its own.
@tmoney: ” I do find myself tempted to say, wait until the iPhone is on more than one carrier”
I do live in a place where the iPhone is on more than one carrier. It is in fact on all carriers here. In Europe all cellular networks use the same technology, and if you get a iphone without a contract you can use it on any network you fancy. Comscore will certainly bring out a report on the EU Smartphone market too, and that will give some interesting data for comparison… (In the last report the iPhone was still outselling Android phones 3 to 1 in Europe)
Eric should put money on the line. I too think when Verizon gets thr iPhone early next year (which Eric now apparently now acknowledges), android will need to do something drastic, and about the only thing it can do is drop trou. Eric will celebrate the coming of the $99 android candybar phone, without realizing that death of the platform rides in, as these $99 phones will be extremely limited (poor screens, no gps, limited memory, cheap build quality.)
>(which Eric now apparently now acknowledges),
No. I’m not expecting a Verizon iPhone until 2012. If it were going to happen sooner, Apple would have found some way to convincingly leak it to their investor base by now. So far, all the repetitions of “annny month now” from third parties have been unconvincing; Apple needs to back that up with a CDMA SKU shipping to overseas markets before I’m going to believe.
I do see a lot of iPhones around me (Europe). So I would not be surprised if the iPhone still outsells Android. I heard (sample N=1) complaints from teenagers that Android is too power hungry to really use as a music player. I must admit that I have to be very, very careful with G3 and automatic updates to get more than a day out of a charge on my Nexus One. But I have no experience with iPhones.
My thought about the smartphone wars is that the Android producers have no choice. For them, it is android or bankruptcy. They cannot switch to producing iPhones or Blackberries. And their only alternative OS would be Windows, which has none of the pluses of Android and more than enough minuses.
Yeah, whatever… my question is, “What about Oracle?”
Open Office, Google Java lawsuit, Hudson project trademark hassle, SySql fork…
When will the madness stop (or at least, when will it be reflected in Oracle’s share price?)
“Now is the winter of our discontent
No glorious summer for Ellison, poor Larry;
And all the clouds that low’r’d upon his house
In the deep bosom of court filings buried.”
>When will the madness stop (or at least, when will it be reflected in Oracle’s share price?)
The fact that it hasn’t already been means that large investors have about the same evaluation of the lawsuit’s chances that I do. Or to put in another way, they’ve already priced in the expectation that the suit will fail.
Oh yeah, ran across this.
Turns out, its completely contrarian to your POV.
And I gotta say, his analysis seems a bit more based in fact than yours.
And here he weeps openly for Android:
@JTL
“The data shows that Verizon sold, on average, about 3 million smartphones per quarter during the first nine months. That’s about 1 million a month. That was based on three platforms, six vendors and widespread promotion.”
You yourself pointed out before that 9 million Android phones were activated a month (worldwide). This would come down to more than 25 million a quarter.
http://twitter.com/Arubin/status/12727540783251456
AT&T sold 5 million iPhones in Q3. The rest of the world would have to buy 20 million iPhones a quarter to make up for the difference.
Unless of course 300,000 Android activations a day does not translate in 27 million activations a quarter. That is, it is marketing speak.
>Also, I would like to point out that despite my (and your) interest in NFC technology,
I have very little personal interest in it, actually. I keep talking about it because, according to Eric Schmidt, it’s a harbinger of an alliance between Google and the credit card companies that implies both are going to make massive piles of money.
I noted that large investors seem already to have priced into Oracle stock the expectation that its lawsuit vs. Google will fail. Because we are scrupulously fair here at Armed & Dangerous, I will note that they may have made an incorrect inference based on the last large software-IP lawsuit involving Unix technology.
I refer of course to the SCO lawsuit against IBM – massive fail on epic scale. It would be easy, though unwarranted, to infer from that lawsuit that suing big open-source infrastructure projects is necessarily a sign that some bunch of losers is trying a Hail Mary play.
Contingently I think this is highly likely to be true in any individual case, but it isn’t necessarily true.
@esr
>I keep talking about it because, according to Eric Schmidt, it’s a harbinger of an alliance between Google and the
>credit card companies that implies both are going to make massive piles of money.
And Google Wave was going to be the next big thing in social networking and collaboration. Google is one of the few rare companies that likes to take risks on odd projects and see where they go, unfortunately most of them don’t go very far. I’ll believe NFC as the future when I see multiple models using it, using it consistently, and merchants playing along too.
Also interesting to ponder is that the market segment that the credit cards really make their money off of, is also the segment least likely to be able to afford the latest and greatest smart phones (although, equally the segment most likely to spend money on credit that they can’t afford to get one)
Eh. I’m not sure how Eric missed it, but Apple hired NFC expert months ago. They’ve also been playing patent games. Of course, when Google does this, it’s an indication that there’s an alliance between Google and the credit card companies. When Apple does it, it’s an indication that Apple fanboys are blind, or something.
My prediction? When Apple ships an iPhone with NFC next summer, it’s going to have a clear reason for existing. It won’t be a feature checklist with no current use.
Eric’s also being rather stubborn about what constitutes a convincing leak. The WSJ has the leak; the WSJ is one of the conduits Apple’s used for this purpose in the past; Yukari Iawtane Kane, who wrote the WSJ Verizon iPhone story, also wrote the tablet story last January. So it’s the same damn reporter who’s gotten Apple leaks in the past. I’m not sure why this isn’t convincing to Eric. I suppose the question is this: what counts as a convincing leak?
>When Apple does it, it’s an indication that Apple fanboys are blind, or something.
No, this is actually interesting news. But there’s an important difference between “Apple hires an NFC expert” and “Google is shipping an NFC-enabled product now” that you would do well not to forget.
>I suppose the question is this: what counts as a convincing leak?
Well, in this case, it would be a rumor that hasn’t been re-launched so many times over the last year – every time without actual result in shipping product – that I find myself reacting with “yeah, yeah, flog that dead horse one more time, why don’t you?” I might as well ask how many iterations will be required for you to become skeptical.
>When Apple ships an iPhone with NFC next summer, it’s going to have a clear reason for existing.
Actually, I agree with you that Summer 2011 is a plausible timeframe for Apple to do this. But I think you may be missing something.
Google’s NFC has, as JonB has pointed out, a clear reason for existing; the Japanese market (I should have spotted this, as I already knew that phones as payment devices are widely accepted technology there). But let’s suppose it didn’t. I think it would still be to Google’s advantage to crank up the competitive tempo with NFC and whatever other technologies they can load into a ubiquitous-computing device. This attacks Apple in two ways:
1. Jeopardizes their brand image of cutting-edge cool. When that’s your positioning being seen to play catch-up is deadly.
2. Puts competitive stress on Apple’s product-development machine, which fields fewer people than the Android army, can’t try as many things, and has much higher downside costs associated with any failure.
Assuming NFC is a complete and total flop in the states. The descriptions i’ve read of the technology quote it as already seeing heavy use in Japan. So from a phone manufacturer perspective NFC is a necessary function to break into the Japanese market. It’s also rating some air time in Australia (though not under the NFC name of course).
Nexus S will not even register on the radar during this holiday season. Yes, geeks will be all over it, but the masses won’t even notice. The reasons are simple: you can only buy it online or in one store (Best Buy), and either pay full price or get a 2 year contract with T-Mobile – at a price point similar to the iphone. There’s no good story behind it from a general populace point of view.
NFC is interesting, but doesn’t have a supporting infrastructure at this moment so one can’t really use it. My guess is it will enter the play in about a year or two when NFC-enabled points of sale will become more ubiquitous.
I doubt Eric missed it it’s pretty much in the top 10 of NFC searches. The interesting point about this is unless the Verizon leak is accurate AND the verizon IPhone has NFC, Apple is going to be 6 months “behind the 8 ball” on a feature. As you say, this is likely to be only for bragging rights in the US but as I said above… possible killer feature for the phone in Japan.
Actually no. the “Indication” of an alliance between google and the credit card companies was basically from the mouth of Eric Schmidt (search for Eric Schmidt and NFC). ESR asserted in the last smartphone wars that the credit card companies are ok with this (I haven’t personally seen quotes to that effect but haven’t seen denials either).
ESR says: I’ve incorporated your corrections.
>ESR asserted in the last smartphone wars that the credit card companies are ok with this (I haven’t personally seen quotes to that effect but haven’t seen denials either).
I infer this from two things:
(1) It’s not Schmidt’s style, or hasn’t been in the past, to make a claim like that without having solid reasons to believe it.
(2) I think I understand the credit-card companies’ business model. NFC deployment would have, for them, the dual good effect of (a) reducing POS hardware costs, and (b) ereducing the consumer’s friction cost to do payments. They’ll like the first effect; they’ll love the second.
For what it’s worth, and I recognize I have no evidence, I was skeptical of every rumor up until the one that appeared in the WSJ. I am somewhat skeptical of this one because the timing is strange, for that matter. The WSJ usually leaks much closer to the Apple event. The timing of the WSJ article would normally point to a Q1 2011 event, but that’s bad for Apple’s normal iPhone development cycle: either it’s an iPhone 4 on Verizon, which is a harder sell if you know there’s an iPhone 5 coming out in the summer, or it’s an early iPhone 5. I guess the latter is more plausible. Still, the timing is odd.
I don’t even remotely believe the latest rumors about an LTE release right after the holidays.
Anyhow, OK. I think you’re saying that you won’t believe any rumors until the actual announcement. That’s fine and understandable.
Regarding NFC and Japan: yeah; when and if the Nexus S is released in Japan, that’ll matter. Actually, the interesting thing there is that Apple has something like 60% of the smartphone market in Japan. It’s already competing with NFC and doing OK. Obviously they want to get NFC in their phones, but it’s not like Google’s creating pressure that’s not already there. Gotta be careful about wearing those fanboy goggles: there are other smartphone providers.
>Anyhow, OK. I think you’re saying that you won’t believe any rumors until the actual announcement.
To be very clear: the event I’m watching for is an announcement that Apple’s going to ship a CDMA iPhone for overseas markets where they’re in demand and no carrier has an exclusive. If that happens, and then there’s a credible rumor that the AT&T exclusive is about to pop like a balloon (one coming through known Apple leak channels) then I’ll believe a Verizon iPhone is imminent.
I’ll say, further, that I think the window of highest probability for such an announcement is 1Q 2011. If I were Jobs, I’d want to announce it at MacWorld, giving them 120 days or so to prep for a mid-year Verizon launch. This is the optimistic scenario for Apple fans.
Past performance says Apple has roughly year-long development cycle on new iPhone models. If we don’t hear anything official about CDMA in 1Q 2011, on the other hand, I’m going to take it as confirmation of my belief that AT&T has Apple on the hook until 2012 and Jobs’s detestation of multiple SKUs has won out over the desire to have a CDMA option in place early.
@ JonB
>Assuming NFC is a complete and total flop in the states. The descriptions i’ve read of the technology quote it
>as already seeing heavy use in Japan. So from a phone manufacturer perspective NFC is a necessary function
>to break into the Japanese market.
I should have clarified that I am mostly speaking of and to the US market here. Like most of these sorts of tech, I expect it to catch on elsewhere first. Also note that despite my pessimism, I do expect to see NFC or some similar possibility appear here, but I am suspicious as to its chances of wide spread use. As much as I would love to be proven wrong, I think that we the American people are far too distrustful of automation technology and that it will probably take until the current generation of middle and high school students who have grown up with a cell phone in their pocket before cell phone as a credit card really catches on.
Also, while I am aware that cell phone as credit card has been around in Japan for a while, is it actually NFC as the Nexus S has, or is it some other tech? Also, it is clearly not a necessary function for entry into the Japanese cell market, for reference see the success of the iPhone in Japan.
@esr
Apple’s brand image has never been cutting-edge cool in terms of feature lists. This is why the first iPhone, despite its lack of features (need I dig up old reviews?) was such a smashing success. Apple’s cutting edge is on the desirability front, not the feature front. In general, you could sum up Apple’s approach to features as perfecting existing technologies. In general, Apple picks from existing technologies that have some momentum and gives them that push and polish they need to become consumer technologies (Wi-Fi networking, iPhone, iPad etc) or make it the only option for their customers almost forcing a new market into existence (USB). Unless Google can make NFC a consumer tech, Apple will hardly be playing catch up or having their “image of cutting-edge cool” placed in any danger. I should note that Apple has not always succeeded at this. .Mac/Mobile Me has been Apple’s attempt at pushing computing into the “cloud” but they’ve pretty much failed and continue to fail at it (cost is a, but not the only, reason), and here Google has certainly eaten Apple’s lunch.
As to consumer friction to do payments, I wonder if the CC companies really don’t see the danger in smart phones being their payment method of choice. I certainly do. CC money making relies on carrying a balance, which in part relies on over spending. If NFC on smart phones catches on, I immediately envision a multitude of apps that won’t let you make an NFC swipe after exceeding some user set spending limit without explicit actions designed to slow the transaction down and add more friction. Who do you suppose has less friction, they guy that has to reach for the card he keeps in the wallet he has on him anyway, or the guy that has to type in his password and confirm his current bank account balance before his phone will let him pay? Of course, such apps are voluntary, but again, see what I said previously about American distrust of automation tech.
@Bryant,
Apple does have a tendency to announce or release new products 15 days after xmas (full disclosure if I haven’t mentioned it before, I used to work Apple retail, I dreaded the days xmas+15 through xmas + 30 or so because almost inevitably, Apple would release or announce something new, and the holiday return period ended 14 days after xmas). Further the original iPhone was announced some (6?) months before it’s debut, so an announcement soon for the next iPhone in July would not be unprecedented.
@esr
I don’t honestly follow cell tech enough to know this, what markets outside the US does CDMA have a large foothold in that are also smartphone markets? It was my understanding that GSM is pretty much the world tech and that CDMA was almost exclusively a US problem.
>It was my understanding that GSM is pretty much the world tech and that CDMA was almost exclusively a US problem.
See this list.
I think there’s almost no chance Apple will make that announcement before they announce a Verizon iPhone. Apple’s instincts always lean towards surprising people. Announcing a CDMA iPhone for non-US markets removes the surprise of a Verizon iPhone in exchange for very little. But we’ll see!
@tmoney — you make a good point! So maybe a January announcement for the usual June release. Also, yeah, all of that about the Apple coolness factor.
Apple’s cutting edge is on the desirability front, not the feature front.
It doesn’t matter how cutting edge it is, if it doesn’t work like a good smartphone. Apple gets props for making a solid device and raising the bar.
That list doesn’t have numbers associated with it, making it less than useful for determining how widely CDMA is used outside the US. A quick look says that there’s no relevant CDMA penetration in Western Europe or the Middle East. South Africa has no CDMA — I assume, perhaps incorrectly, that South Africa is the major smartphone market in Africa. I think there’s still a lot of CDMA in Asia. In particular, Japan is a CDMA territory. (The Nexus S is not a CDMA phone.)
I am pretty reluctant to say that it’s either important or unimportant for Apple to ship a CDMA iPhone without actual numbers, but that’s me.
@TJL:
Cherry-picked facts. A lot more of those iPhones are replacements than the Android phones. Also, that report is based on data from “independent stores”, not company-owned stores.
http://blogs.wsj.com/digits/2010/11/01/android-overtakes-iphone/ says 5.5 million iPhones sold in the quarter.
The comscore report says that iPhone users in the quarter were 24.6% of 60.7 million smartphone owners, and that the previous quarter, iPhone users were 23.8% of 53.2 million smartphones. This means that the actual increase in iPhone users for the quarter was only 2.27 million iPhones. This discrepancy (the “missing” 3.2 million iPhones) is easily explained by AT&Ts upgrade policy.
Speculation about whether AT&Ts easy upgrade is due to fear of Verizon or fear of Android is best left to others.
Of course, if you combine that data with the comscore report, you will find that 5.5 million is 243% of the total increase in U.S. iPhone users over the last quarter, so it’s likely that 3.2 million of those 5.5 million are replacements.
The same comscore report shows (if you do the math) an increase in Android-carrying smartphone users of 5.2 million users for the same period.
To be fair, analysis indicates that, to the extent that average number of Android activations per day for the quarter was north of 60K, there was a lot of churn in Android devices as well. But that does not negate the fact that, during the quarter, the number of iPhone users increased by 2.3 million, while the number of Android phone users increased by 5.2 million…
>But that does not negate the fact that, during the quarter, the number of iPhone users increased by 2.3 million, while the number of Android phone users increased by 5.2 million…
The Apple-idolater response to figures like this, of course, is….”Phthphthbbttt! Just you wait until Verizon ships an iPhone! Then we’ll show you, you big nasty mean person!” I think I need to do a blog post on the seven stages of anti-Android denial.
Okay, tmoney, let me clear this up for you, since you seem to lack a complete understanding of basic business science concepts. “Desirability” isn’t a product trait. You don’t get it by engineering anything. The desirability of a product is determined entirely by marketing and brand image.
That’s what Apple has: insanely great marketing, insanely great brand image. Not insanely great technology. Not insanely great engineering.
Steve Jobs’ model for running Apple is Sony. That works great in a market where brand differentiates a product more than features or compatibility. But in the technology sector, it isn’t about brand, it’s about features and compatibility. Apple will carve out a nice niche for itself, but it will not win the smartphone wars playing like Sony.
> 1. Jeopardizes their brand image of cutting-edge cool. When that’s your positioning being seen to play catch-up is deadly.
The Iphone was launched without 3G, copy&paste, tethering and multitasking. All features that were available on the smartphones of the day, crappy as they were. It didn’t seem to impact their brand image of cutting-edge cool. On the contrary, it set the standard that every other manufacturer is now trying to emulate. Apple will not worry about the short term lack of a feature that pretty much nobody outside Japan can currently use.
> See this list.
That list is out of date. Canadian providers have all started to switch to GSM standards, NZ Telecom is supporting their cdma network as “legacy” while building up their UMTS infrastructure and Zapp Mobile in Romania was acquired and their network is phased out. That’s only places I know about from personal experience. I hear many others are phasing out cdma.
@J. Jay
> Eric will celebrate the coming of the $99 android candybar phone, without realizing that death of the platform rides in, as these $99 phones will be extremely limited (poor screens, no gps, limited memory, cheap build quality.)
I think Eric rightly views it as a continuum. Right now, only 26% of US cellphone users carry a “smartphone” That leaves a potential market of 173 million cellphone users in the US alone who are ripe for an upgrade. Are they going to go straight for the iPhone. Sure, some of them will save up for it. The top-of-the-line Samsung? Maybe a few. The $50-more-than-their-current-phone, but kinda acts like a computer? That’s a decent number. The same-price-as-your-current-phone-but-Moore’s-law-let-us-add-a-computer-and-the-OS-is-free? Now you’re talking.
@Winter:
> My thought about the smartphone wars is that the Android producers have no choice. For them, it is android or bankruptcy.
And here we agree. Apple’s unwillingness to license either its computer OS or its cellphone OS to third parties will continue to relegate them to low two-digits marketshare in both cases.
> Apple will carve out a nice niche for itself, but it will not win the smartphone wars playing like Sony.
I don’t think they even want to win the smartphone wars. I believe they’ll be happy with a 20-25% market share, as long as it is the upscale, high margin end of it.
>I believe they’ll be happy with a 20-25% market share, as long as it is the upscale, high margin end of it.
They’re not going to keep that.
The reason I state with such confidence is that “retreat upward” is about the oldest losing move in the tech-industry playbook. History is littered with the corpses of companies who tried it. The only way it ever works is when transition costs out of your product are absurdly high, which isn’t the case for cellphones or any other item of consumer electronics.
Retreating upward looks smart because it improves margins; shareholders love this. It can buy you a few good years. Trouble is, there are toxic side-effects of the strategy that accumulate and eventually kill you. For detailed discussion of one of the most imporant, go read Clayton Christensen’s The Innovator’s Dilemma.
> The Apple-idolater response to figures like this, of course, is….”Phthphthbbttt! Just you wait until Verizon ships an iPhone!
Yeah, but obviously, iPhone on Verizon will be great for Apple, so you have to keep restating your thesis, and some of the consequences of your thesis. For example, that Apple can do well financially with phones, just like they do with computers, but that they shouldn’t be allowed to win the war that lets them define what a phone is, and that Android has taken that ability away from them.
Because (Exhibit A: 5.5 million iPhone units in quarter, and some of the comments on this post) every time something good happens to Apple, those same people treat it like it invalidates your thesis.
I noticed this week that the line Eric proposed for the carriers losing control has been reached: T-Mobile is now selling the LG Optimus for $160 and the T-Mobile branded Comet for $140. That is the unsubsidized price, free with contract. The Comet is also available with prepaid plans (although it appears that prepaid plans don’t have data, you may have to use WiFi for data, but that isn’t completely clear).
Both are pretty wimpy phones (but do have GPS), and the Inquirer.net review of the Optimus was not very positive, mainly because the touch screen was not very responsive and missed touches a lot.
What I can’t quite figure out is whether the predictable Moore’s Law style improvements in the IC’s will help the situation with the screen. Screen resolution doesn’t follow the same kind of curve, but that is separate from the responsiveness of the touch function, which seemed to be the biggest complaint about the cheap LG phone. If cheaper IC’s or improved firmware help the screen responsiveness, then processor speed and memory size will take care of themselves over the next several months (i.e. phones for the same price next year should have more memory, or the equivalent phone should be a little cheaper next year).
http://www.theinquirer.net/inquirer/review/1900799/lg-optimus-review
http://www.theinquirer.net/inquirer/news/1928949/lg-optimus-video-demo
>I noticed this week that the line Eric proposed for the carriers losing control has been reached:
Not quite. I picked that idea up from Jean-Louis Gassée, who thinks the magic price point that kills off the contract system is $89. That seems a bit low to me, but $140 is still a little high. Besides, you said “free with contract”. That kind of negates the premise.
CDMA is like inches and pounds.
The rest of the world has already standardized on something else.
Huh, there seems to be some confusion there.
I, personally, am not saying “Phthphthbbttt.” I am also not saying “Just you wait.” I will also refrain from saying “you big nasty mean person.” I’m not a fan of demonizing people I disagree with; I find that to be a profoundly irrational act which, rather frequently, prevents me from learning anything. Your mileage may vary.
What I will say is this: I expect iPhone market share to increase if and when the iPhone is available from Verizon. I hope that’s not a wildly controversial statement, but if it is, let me know and I’ll explain why I think that’s true. I don’t think Android would vanish, either immediately or over time. I think Android will be a great smartphone and tablet OS for a long time, and I think iOS will be the same. I certainly hope both of those things are true. Android would be much, much worse if Apple hadn’t demonstrated what a smartphone could be, and I believe that the very minor improvements Apple has made to the App Store process are probably thanks to Android. Monocultures suck.
>I, personally, am not saying “Phthphthbbttt.” \
No, you aren’t. My snark wasn’t mainly aimed at you.
The iPhone is available on all carriers in Canada.
It may be useful to look at Canadian iPhone vs. Android uptake numbers, rather than simply hypothesizing without basis that multi-carrier iPhone will/will not overtake Android.
@esr:
I don’t think they have that. It’s even not about “keep”
If you take the view (that Android will fully support) that cellphones live along a price/functionality continuum, then you have to consider that the first prerequisite for wanting a smart phone is that you, in fact, want a phone. Sales statistics show that, last quarter, worldwide, less than 20% of people who bought a new phone thought it was worthwhile to buy any kind of smartphone, and only 3% thought it was worthwhile to buy an Apple smartphone.
iOS will probably wind up with a much smaller share of the smartphone OS market share than MacOS has in the computer market. At least in the computer market, Apple is primarily competing with box manufacturers who have to pay something for the OS. But in the cellphone market, Apple is competing with people who get the OS for free, who build their own touchscreens and chips. The Gartner report I just quoted shows that Apple and Android are both doing very well in the US smartphone segment (where phones are heavily subsidized), but in the rest of the world, unsurprisingly, cheap Symbian phones are still in the lead. Moore’s law is increasing Android’s competitiveness vs. Symbian (and all smartphones vs. featurephones or dumbphones), but it’s hard to imagine, as smartphones take over the more price-conscious general purpose phone market, that Apple has any interest in dropping prices enough to even get to 10% of the overall phone market. Conversely, it’s very easy to imagine Android eventually exceeding 50% of the overall phone market in a few years.
Patrick M:
The $50-more-than-their-current-phone, but kinda acts like a computer? That’s a decent number. The same-price-as-your-current-phone-but-Moore’s-law-let-us-add-a-computer-and-the-OS-is-free? Now you’re talking.
And that’s just US and Europe. Right now, for some definitions of smartphone (“Has a webbrowser”), the most common smartphone user worldwide is the middle-class resident of a ‘third-world’ country with a Symbian phone that’s probably their only computer.
Unless Nokia (or Microsoft?) really pulls a rabbit out of their hat, that market seems likely to go Android in a big way, although the churn rate there is a little lower than the US or Europe, so it will take a few years.
>as long as it is the upscale, high margin end of it.
Eric already beat me to the Innovator’s Dilemma reference. But I also wanted to point out , that from what I read at the time, in the late nineties, Apple almost died from a similar strategy when Windows 98 and Adobe software bumped it from its last refuge in desktop publishing. At that point, you could do anything that a Mac could do on a Windows machine at a lower cost. Please note that I have never used an Apple product, nor any major Windows applications, this is from what I remember reading in books and magazines at the time.
@ Morgan
Someone piss in your corn flakes today?
> The desirability of a product is determined entirely by marketing and brand image.
>That’s what Apple has: insanely great marketing, insanely great brand image. Not insanely
>great technology. Not insanely great engineering.
>But in the technology sector, it isn’t about brand, it’s about features and compatibility.
>Apple will carve out a nice niche for itself, but it will not win the smartphone wars playing like Sony.
And none of that has anything to do with my assertion that the iPhone’s lack of a single given (limited, mostly unknown and by no means essential) feature spells the beginning of the end of iPhones “cutting edge”. The fact of the matter is, the iPhone has been and is “cutting edge” in spite of, not because of its feature list. And the lack or appearance of NFC is unlikely to sway any large swaths of potential phone buyers for years to come.
@esr
>The only way it ever works is when transition costs out of your product are absurdly high,
>which isn’t the case for cellphones or any other item of consumer electronics.
This may not be true for previous generation cell phones, but tell me, what are the transition costs for someone with a few hundred dollars invested in iPhone / Android / Blackberry apps? I imagine some developers might try to make it so that if they develop for both platforms you can switch versions at no cost, but there are absolutely new costs of switching phones in the new smartphone world.
>My snark wasn’t mainly aimed at you.
I have to wonder at who then, because to me, the only people I see snarking and deserving of snark right now are you and Morgan (and myself as of this post)
@ Patrick
>For example, that Apple can do well financially with phones, just like they do with computers,
>but that they shouldn’t be allowed to win the war that lets them define what a phone is, and that
>Android has taken that ability away from them.
I would argue that while it is unlikely it will remain this way, Android has not yet taken away Apple’s ability to define what a phone is. While the details are different, from a overhead view, all the successful Android phones are really defined by their similarity to the iPhone. That isn’t to say that Apple will keep this sort of control into the future, but for the time being, I would bet my money on a radical change in the iPhone showing up similarly in Android phones before I would bet a radical change in any given Android model showing up in the iPhone. Arguably, Android can’t control what a phone is unless Google stays in the hardware game (or at least continues to produce a “pure google” phone, simply because Android can’t dictate to either manufacturers or carriers the terms they want, Apple can as they are their own manufacturer and the iPhone is still a major phone.
>Conversely, it’s very easy to imagine Android eventually exceeding 50% of the overall phone market in a few years.
You’re likely right, but I have to wonder what happens at that point to allow manufacturers to differentiate themselves? In some ways, I fear us seeing Android on multiple phones the way that computers running Ubuntu, RedHat and SuSE are all running linux. It’s all the same at the low level for the geeks, but for the consumer, they might as well be different OSes entirely.
Do not take me for saying that Apple will continue to dominate the phone market as they currently do. But just because they don’t hardly spells the end of Apple or its influence in the mobile computing world.
>This may not be true for previous generation cell phones, but tell me, what are the transition costs for someone with a few hundred dollars invested in iPhone / Android / Blackberry apps?
Might be an interesting question if any such person existed. At the prices I’m seeing for apps, I doubt it.
>The fact of the matter is, the iPhone has been and is “cutting edge” in spite of, not because of its feature list.
Uh huh. This is why Jobs hyped the “retina display” so hard. Now you’re talking nonsense.
> $140 is still a little high. Besides, you said “free with contract”. That kind of negates the premise.
Sorry, a bit of wooly thinking on my part led to mixing two different ideas together.
To separate out the two ideas, the first would be that $140 gets into the range where people like me who previously spent $70 to get a plain jane phone without contract will now be tempted to pay a bit extra and get a smart phone without contract. The extra $70 becomes relatively easy to justify when considering how much I save with my pay per use plan rather than a contract with fixed fees. That only works for people who don’t spend a lot of time on the phone.
The second is that with the unsubsidized price of the phones that low, the subsidized price becomes free, at which point the comparison for a customer becomes get a “free” (with contract) phone that is just a phone, or get a “free” (with contract) pocket computer that makes phone calls, can access the app store, can browse the web, get email, etc. I see it as the point where a smart phone running android becomes the easy default choice, and you have to actively look for a dumb phone. (Could still be a dumb-phone market there for techno-phobes, or people who value battery run time above other features. Maybe blind people as well if smart phones don’t get sophisticated accessibility features. All a very small market, however.)
Once you pick up cheapskates like me, and people who just want whatever cool phone they can get free with their contract, what else is there? At least in the North American market that pretty much sews up the market: looking for a free phone, just want a cheap no-contract phone, and willing to pay more for a better phone (with or without contract). It seems like the network effect at that point would be so strong that the whole argument will be over sooner rather than later, and will devolve to a few on the edge arguing about whether this year’s iPhone is better at this or that than the equivalent high end Android phone from Samsung, or HTC, or Motorola, or whoever has the popular highest-end phone that particular year (which will probably rotate among vendors, since it doesn’t seem that any one hardware vendor is really pulling ahead of the others).
“The $50-more-than-their-current-phone, but kinda acts like a computer? That’s a decent number. The same-price-as-your-current-phone-but-Moore’s-law-let-us-add-a-computer-and-the-OS-is-free? Now you’re talking.”
I just checked AT&T’s website. As a non-logged in user looking at smartphones with a new contract, I see refurbed Arias and Captivates for $10. That’s an interesting price point. For that matter, the Backflip is $50 new, and while it’s probably not that exciting to most people here, for someone who doesn’t already have a smartphone, a $50 android phone with a 3″ screen and a keyboard is probably also interesting. For that matter, there’s now two variants on the backflip: the flipout, which is a rotating device at $80, and the flipside, which is a backflip in a regular slider format for $100.
tmoney, all I’m saying is that I’m not convinced by your argument. Being first-to-market with NFC means that Google has a shot at owning this market: stores across the country will install NFC receivers and they’ll be compatible with Android, but probably not with iPhone. It’s features and compatibility all over again, and Apple has set itself up for failure — again.
@ Morgan
Just like the original iPhone was a failure without its 3G connectivity or MMSMS right? It’s like I was saying before, NFC is a tech chicken and egg problem, and your your scenario to come true, NFC receivers + NFC cards / bank access have to appear, and grow in popularity, and demand to the point it is an essential or highly desirable feature, and all before Apple releases a phone with NFC tech on it. Right now, we have NFC on one Android model on one carrier, virtually no NFC POS terminals and no confirmed agreements with Google and any card distributor or bank.
According to the numbers http://www.gartner.com/it/page.jsp?id=1388914 mobile payment makes up a mere 3M users in north america, compared to 62M in asia, where even without NFC, the iPhone is successful. As I said, without a massive shift, I don’t see NFC availability being driving feature here for a few years.
Eric: I’m not sure that the Innovators Dilemma applied to Apple just yet. That phenomenon applies assuming a fix good ceiling. You can retreat into the high-end as long as the high end is growing. Compare this to Intel processors in the 80s and 90s. Just about anybody else can come out with a product which is competitive with their current offering, but as long as they still have a large area to grow into, they can hold off. Note that though Apple still sells desktop machines, almost nobody cares any more – they aren’t competitive except in certain niches.
tmoney: Credit card companies make money off of three main things: transaction fees charged to the vendors, interest charged to cardholders, and fees charge to cardholders. All of these have different target markets and risk rates. The transaction fees, which are currently about 2-4% but which will probably eventually settle at or around 1% are based on volume and are very low risk. Remember that they get the money by turning over less to the merchant. This is a near zero-risk income stream. Fractionally small, but the aggregate is huge. Interest payments target customers who need to manage cash flow, but are otherwise in good shape. This is income which is priced according to risk and has to be competitive with bank loans and check cashing places. These interest rates are pretty high, but it is unsecured credit. Finally, the fees are how credit card companies make money off of people who probably shouldn’t be issued cards and are charged $20/month in late payment fees based on a $200 balance that can’t be paid off in a reasonable basis.
If the transaction cost for credit card companies can be reduced by reducing wear and tear on CC machines, and the costs can be reduced for customer support such as replacing lost, stolen or worn cards, the credit card companies would be interested in at least checking it out. They would be interested even if it were only to bring in the transaction fees. 1% of something is better than 100% of nothing.
>Eric: I’m not sure that the Innovators Dilemma applied to Apple just yet.
You may be right. On the other hand, nobody thinks the innovator’s dilemma applies to them, until their business gets dry-gulched by it.
@tmoney:
I was using a slightly different meaning of “define” as in, for awhile there a couple of years ago, it looked as if Apple were going to own the smartphone category to the extent that, if it wasn’t running iOS, it couldn’t be a smartphone. We’re past that point, but you are of course right that Apple will continue to help to define valuable phone features, just as they have with computers.
Exactly, but I don’t think that contradicts anything Eric’s been saying about how this goes down. It’s certainly not a statement I take issue with. Apple making good money on a small percentage of the cellphone market is fine and, absent any screwups on their part, should continue indefinitely.
I think the excitement I personally feel about Android is because it appears that it will help to make the following two statements true:
1) There should be a single obvious candidate for dominant smartphone OS, to drive development; and
2) That dominant smartphone OS should be FOSS.
If you look at the history of PCs and DOS/Windows, you can see all the good that (1) caused there, while simultaneously seeing all the problems that the lack of (2) caused there. Since the battle for control of how users compute is now moving from the desktop/laptop down to the phone, it is great for FOSS that, on this wildly successful new computing platform, the wildly successful new OS is open.
The most optimistic stats I have seen on this indicate that perhaps 50% of iPhone users actually purchase apps, and those users spend maybe $10/month on apps. So for some people, that could be a problem, but it’s doubtful — most of those apps are probably games; very few people are likely to have more than $20 worth of “must-have” apps. OTOH, you have hit the nail squarely on the head why marketshare matters for Android. If, as you admit is possible, Android could take 50% market share within a few years, that is a market size that developers ignore at their peril. Sure, there will be specialized apps that can demand a premium phone, and some app developers will realize that the price-insensitive portion of the market is on iPhone, so the iPhone may be a better bet as far as per-copy revenue goes, but that only works until some competing developer aiming for the mainstream does something really compelling, and doesn’t bother to port it to iPhone.
Btw, when can I get my smart phone/stun gun combo device? As soon as that comes out I’ll spring for a smart phone.
> Btw, when can I get my smart phone/stun gun combo device? As soon as that comes out I’ll spring for a smart phone.
You won’t want one of those from Apple. The wear and tear on the non-replaceable battery is a PITA if you have to stun more than one or two people a month.
I see that one phone (ok, technically 4 phones) on one carrier holds the same level of command as 50+ phones on 4+ carriers.
Bingo. I’m still buying AAPL shares.
Apple would have found some way to convincingly leak it to their investor base by now.
Nope. First of all, Apple doesn’t do that, and secondly, if they did do that kind of thing, there’s no reason to do so when AAPL is trading within a buck or two of its all-time high.
Steve Jobs’ model for running Apple is Sony
You’re about a decade out of date on that. SJ’s model for running Apple is Apple.
@Some Guy:
Hey! That’s MY patented buy-high/sell-low strategy!
@esr
>Might be an interesting question if any such person existed. At the prices I’m seeing for apps, I doubt it.
You don’t think in the future that people will have similar amounts of money (relative to the cost of the unit) tied up into their phones as they do for their computers? Further, it doesn’t take much to reach that point. If we go with the stats that patrick gave, $10 / month puts you $240 into applications within a single contract term. And if any of these people are iPod touch / iPad users, I imagine there’s even more tie in.
>Uh huh. This is why Jobs hyped the “retina display” so hard. Now you’re talking nonsense.
How much of Apple’s marketing talks about the retina display? How many users do you hear saying “Man, I’d get an Android, but they just don’t have a retina display?”
@Garrett
Do the CC companies bear the costs of the POS terminals? It was my understanding that merchants and the merchant banks are the ones that bear those costs.
While I agree that the CC companies are would gladly take any percentage they can get, and I admit (even if I didn’t make it clear before) that they are the most likely group to jump on board first, they are not the only hurdle facing american NFC adoption.
@ Patrick
Ah, that makes a bit more sense now, though I really can’t imagine there are people who truly think or thought there would ever be a point where if it was Apple, it wasn’t a smart phone, especially when we’re talking a company who prides themselves on “the one true way” for most of their products. Then again, there are people who believe Germany was invited over for tea by Poland in the early 1900’s so I guess it shouldn’t surprise me.
As to dominant open source OSes, I remain unconvinced of the damage that a lack of open source operating systems caused to computers. Open source programs, and open data formats, absolutely I buy that argument; as to the OS, I’m not so sure. I think to most users (phones and computers alike) the OS is that thing that lets them get started, and whether it’s OSS or not really wouldn’t make a hill of beans difference to them, or their experience. This is especially true when we consider how the people who dominate OSS could generally not give a damn about doing the boring design, UI, ease of use and documentation stuff that closed source vendors pay their guys to do. Not saying there aren’t exceptions, or that OSS guys couldn’t make it happen (a la Google), but as a rule, that sort of stuff is boring to the guys who tend to get into OSS. A fusion model seems the most likely to make the world go round, and two of the best examples are found in Google and Apple with Android and Mac OS X respectively. It’s also worth noting that even OSS hasn’t prevented carriers from locking things down and making it necessary to root your Android phone just like someone would an iPhone.
>If, as you admit is possible, Android could take 50% market share within a few years, that is a
>market size that developers ignore at their peril.
To me it all depends on what that 50% is made of. Vanilla Android clones? Absolutely, ignoring it would be like ignoring windows. Disparate distributions, all trying to differentiate themselves from the competition? Now it’s a bit tougher. What if the majority of that 50% are dumb(er?) phones that “run android” the same way your DVD player “runs linux”? How many of those people can you push your app to? How many of them will be allowed to upgrade to newer adroid versions? For now, the carriers still seem to strangle that option, so even if 50% run android, if only 10% run the version you developed for? Admittedly Apple has this problem starting with iOS4 so we’ll have to see how that shakes out.
>You don’t think in the future that people will have similar amounts of money (relative to the cost of the unit) tied up into their phones as they do for their computers?
No, because I have yet to buy an app. So far, everything I’ve been able to identify that I wanted has been available for free. And I am probably more willing to spend money on such things than most.
If you’re evaluating product stickiness, it’s absolute cost not relative that matters.
>How much of Apple’s marketing talks about the retina display?
More than enough of it. They worked hard at making “retina display” a term everyone would recognize.
>I remain unconvinced of the damage that a lack of open source operating systems caused to computers.
You haven’t been paying attention, then. Closed-source operating systems are lock-in devices. That’s damage in and of itself.
@tmoney:
It’s the hidden tax that raises the price of the computer, allows Microsoft to fund anti-competitive forays into other types of software, and made it difficult for someone who created Window apps to offer those on lower-priced systems — look how long it took Wine to get reasonably functional, and it’s arguably too late now — Windows is quite entrenched, and any change to this will happen quite slowly, I think (although a big box that was just a faster computer for your Android cell phone to control could be interesting…)
If you take the OS in a vacuum, you might have the disparate distribution problem. But if we segregate apps into free and paid apps, and realize that, at least on Android, the vast majority of apps that affect the vast majority of people are going to be free, then any carrier that somehow changes Android so much that it doesn’t work with most of the free apps in the app store will be cutting off his nose to spite his face. The google app store is an indispensable part of keeping Android from fragmenting. Combine that with the rapid pace of Android development, and the cost of programmers, and it becomes very costly for a carrier to attempt too much differentiation.
> Uh huh. This is why Jobs hyped the “retina display” so hard. Now you’re talking nonsense.
The thing I noticed is that switching *to* the retina display wasn’t that big a transition. But switching back (as I do whenever
I start a test cycle on my Nexus One (I do software dev on both iPhone and Android) … switching back makes the Android screen
look ugly and tawdry. Really gross.
So, no hype necessary.
> No, because I have yet to buy an app. So far, everything I’ve been able to identify that I wanted has been available for free. And I am probably more willing to spend money on such things than most.
Sample randomly, much, Eric?
First, most apps on Android *are* free. Second, you probably think, “I could write that.” for any given app you’re interested in.
Third. Are you now an advocate for … Free Software? :-)
@tmoney:
Anything that reduces payment friction is going to be adopted by credit card companies and retailers in droves. And it doesn’t matter who bears the costs of POS terminals since the nation’s largest retailer, Walmart is its own bank. That matters because no matter how you slice it, it’s in their best interests to have state-of-the-art payment transaction processing. Even NFC receivers were only present in every Walmart store in the country, they’d come into the public view immediately. (Google, are you paying attention here?)
Once you start seeing a big retailer like Walmart adopting this, everyone else will follow suit.
@TJL:
Me, too. Sure. I could write a lot of apps. One of the most popular free apps is called “Just Compass,” which is, in fact, just a compass, for example. It would be trivial to write that app. But why do that when there are already a hundred apps like it in the Android store?
Similarly, I could change my own oil. But the shop down the street does it for ~$30 including the oil in 15 minutes or less, so why would I?
@ esr
>No, because I have yet to buy an app. So far, everything I’ve been able to identify that I wanted has been
>available for free. And I am probably more willing to spend money on such things than most.
I would imagine the apps and things you want are significantly different from average joe consumer.
>More than enough of it. They worked hard at making “retina display” a term everyone would recognize.
And yet, the question remains, how many people do you hear talking about how they would buy an Android if only they could have a retina display? How many people do you really think the “retina display” caused to drop their current phone and buy an iPhone? Marketing of features != success because of features
>You haven’t been paying attention, then. Closed-source operating systems are lock-in devices.
> That’s damage in and of itself.
They’re only lock in if the data format being used is not open. But the fact that windows is closed source does absolutely nothing to prevent me from taking my .rtf files and opening them on a mac, a linux box, a blackberrry or an iPhone or Android tablet. Application lockin comes from closed source apps, unless we believe that windows closed source makes it impossible to run Apache on other systems? Now, agreed a closed source OS with lots of fun API hooks can make porting even an open source app difficult, but even then, it would only be easier if the close source API hooks could be and were ported to another OS, and even then, ask some Java devs how well write once run everywhere works.
@Patrick,
That’s a better argument, but still not all that convincing. The added cost is not because it’s closed source, but rather because it comes with all of those things that most open source OSs don’t. See also the cost of buying RedHat. As for porting apps, see what I said above to esr. Not to say that closed source doesn’t have it’s costs, but I’m saying that open source has it’s own, and there’s no evidence currently to suggest that open source OSs would have been an improvement for computing as a whole. A cursory glance at the current state of Ubuntu will demonstrate the real possibility that open source OSs could instead discourage computer use and regulate it back to the dungeons of the nerd.
>then any carrier that somehow changes Android so much that it doesn’t work with most of the free apps in
>the app store will be cutting off his nose to spite his face.
It doesn’t have to be most of the apps, just whatever is the newest and greatest. Let’s take this NFC we’ve been discussing as an example, let’s say a year from now, it’s in a bunch of android phones but it hasn’t really caught on. HTC or Samsung, deciding is isn’t going anywhere, leaves it out for their new UbarDroid, flash forward 6 months and someone finally figures out a real killer use for NFC (maybe it integrates with facebook and 4square and tells reminds you to pick your nose when no one is looking all at once, pick any sort of out of nowhere advance that tends to happen in the new internet/meme age). Now the UbarDroid and all it’s users can’t use the one new must have app.
Or looking at it another way. Look at the list of Android phones that the Angry Birds guys were having trouble with. Can you pick out the one factor in all of those phones and ensure your new Droidinator is going to work when you release it?
>Combine that with the rapid pace of Android development, and the cost of programmers, and it becomes
>very costly for a carrier to attempt too much differentiation.
But too little and they run the risk of being replaceable. But I’m not so much envisioning carrier differentiation as I am manufacturer differentiation.
>I would imagine the apps and things you want are significantly different from average joe consumer.
You imagine a lot of things that aren’t true. I am a very typical smartphone consumer in this respect – most of my apps are games.
>Marketing of features != success because of features
You’re entitled to that opinion, but judging by how they market Apple does not appear to share it.
>But the fact that windows is closed source does absolutely nothing to prevent me from taking my .rtf files and opening them on a mac, a linux box, a blackberrry or an iPhone or Android tablet.
Quite right. Now try this with any application where Microsoft hasn’t been forced kicking and screaming into offering a portable format. One of the many, many problems with closed-source operaing systems is that they encourage data jailing by making obscure, undocumented data formats the rule rather than the exception.
@Morgan,
But, that requires Wal-Mart believing it will make a difference, or people will take it up. Again, we’re talking chickens and eggs. One android phone is not the tipping point here. Nor is one major retailer, as I recall, most McDonalds have some form of wireless payment system (not sure if it’s NFC or not), but I don’t see any recommendations from my bank to upgrade to an NFC device.
Again, I’m not arguing it can’t or won’t happen, I’m arguing that by the time it does, and by the time it becomes a must have feature, the iPhone will have it.
By contrast, even between proprietary Unices, data file formats have been very open. Of course, most of the formats have been plain text, with various layers of structure placed upon them (colon-separated fields in passwd, etc. all the way up to XML), but even things like compressed tarballs have been portable since forever.
I am such a person. I own ~220 iOS apps, most of which cost at least 99 cents. When I have a need for some category of application I usually find there’s a range of options available from free up to a few bucks up to, say, ten bucks or so. The most attractive and well-designed apps tend to cost at least a couple of bucks, which I don’t mind paying for a better product/user experience. So for instance, the built-in iPhone calendar app is free but sucks compared to earlier datebook apps (including those on Palm or Newton). But there’s a third-party app called “Calvetica” which is much more functional and really gorgeous in a minimalist Edward Tufte sort of way – just a joy to use. (I like the landscape mode layouts best – the built-in app doesn’t even *have* a landscape mode layout!). So I paid the three bucks, why not? I’m essentially upgrading my phone piecemeal, app by app, as I see a need for a better weather app, a better notes app, a flashlight, a to-do manager, a multitrack recorder, etcetera. But the real lock-in value there is the sunk cost in terms of search time to discover the best apps and fold them into my workflow – that would be more of a pain to replicate on another platform than the mere cost of buying all the software again.
I haven’t yet had an excuse to really try Android but having gotten my iPhone 4 just the way I want it does make it a lot less tempting to switch. Android would have to provide some really compelling benefit to make up for that.
That makes sense, but I think you’re at the high end of the market. The Mac held on to those people, even during the most dismal times, and I see no reason why the iPhone would be different. But Apple could hold on to all its customers, and then triple that, and it would still have under 10% of the total phone market.
If Apple integrated NFC into the iPhone…
a) they would do it right
b) vendors and credit card companies would be on board yesterday, and within a few months iPay would surpass American Express in dense urban metro areas.
Being a few months late to the party will not really diminish the advantage to Apple of having the definitive implementation and the one that Just Works.
>Being a few months late to the party will not really diminish the advantage to Apple of having the definitive implementation and the one that Just Works.
Sorry, “Just Works” is a joke in this context following the antenna fiasco. If Apple lives up to its current high standards for wireless implementation, their NFC implementation will only work for transferring money to phishing sites and Nigerian scam artists.
@tmoney:
Every McDonald’s here in Florida has them. Ever been to McD’s recently? They have those POS card readers that say “Visa PayPass” on them? Those are ISO IEC 14443 compliant. That’s what the Nexus S has. Now how many places have these? I don’t know. But I suspect that since credit cards already have this, there will be more and more at retailers near you.
@Winter
> You yourself pointed out before that 9 million Android phones were activated a month (worldwide). This would come down to more than 25 million a quarter.
I pointed out that Andy Rubin tweeted, “There are over 300,000 Android phones activated each day.” Assuming Andy isn’t lying, the *current rate*. A local maxima, if you will. That doesn’t mean or even imply that Google activated 25 million phones during Q3. What are you smoking?
> Assuming Andy isn’t lying, the *current rate*. A local maxima, if you will. That doesn’t mean or even imply that Google activated 25 million phones during Q3.
Andy probably wasn’t lying. The Gartner article I quoted from earlier says Android shipments for the quarter were 20.5M.
Interestingly, there were 2.2M shipments of Windows Mobile (not broken out by version), and 1.7M shipments of “Linux.” Probably already more linux phones than desktops, even without Android…
@tjl>>(which Eric now apparently now acknowledges),
@esr> No. I’m not expecting a Verizon iPhone until 2012.
Yet, you stated, ” If the Nexus S and iPhone sell at rough parity Apple has to do something – another product rollout, a multicarrier announcement ”
Exactly which US carrier do you think might be announced? T-Mobile? Don’t make me laugh so hard, it hurts.
@esr> I refer of course to the SCO lawsuit against IBM – massive fail on epic scale.
This failed because, as it turns out, SCO didn’t own the copyright to Unix. Novel did.
Are you claiming that Oracle didn’t acquire a valid copyright on Java, or are you suffering the same dimwitted thinking as an ordinary layperson, “Well, SCO lost, so Google will lose!” As you must know (or, barring that, at least, as Cathy must know, she being a lawyer), the SCO lawsuit(s) establish no president in the Oracle / Google action.
Or maybe you’re referencing (without naming) the new action against Google by Myriad?
http://www.theregister.co.uk/2010/12/14/oracle_versus_myriad_on_java/
If so, Florian thinks their chances are slim.
http://fosspatents.blogspot.com/2010/12/google-ally-myriad-group-sues-oracle.html
@esr “I noted that large investors seem already to have priced into Oracle stock the expectation that its lawsuit vs. Google will fail. ”
Is this where I point out that at the end of Aug 2010, Oracle’s stock price was 24, and today it closed at 30.51, and that the 30.51 is as high as its ever been since January 2001?
Do you really mean to say that Oracle’s investors are long on ORCL due to the Google lawsuit?
>Yet, you stated, ” If the Nexus S and iPhone sell at rough parity Apple has to do something – another product rollout, a multicarrier announcement ”
Quite. They’ll have to do something in that scenario. But I don’t think it’s going to involve Verizon, and have never implied otherwise.
>This failed because, as it turns out, SCO didn’t own the copyright to Unix.
No, it failed because Judge Kimball disallowed two thirds of SCO’s evidence exhibits and required SCO to actually specify what they thought IBM had copied. My knowledge of this case is close-up and firsthand; it would still be imprudent for me to reveal everything I know, but you will not find a more knowledgeable source than me short of some lawyer who was on the Cravath Swaine & Moore case team at the time. The copyright finding was nice, but SCO was utterly, utterly doomed well before that.
>“Well, SCO lost, so Google will lose!”
You appear to have missed the part where I said this would be an unjustified inference. I think investors are reaching a correct conclusion (Oracle’s odds are very poor) but I think they may be getting there by incorrectly generalizing from the SCO case.
>Is this where I point out that at the end of Aug 2010, Oracle’s stock price was 24, and today it closed at 30.51, and that the 30.51 is as high as its ever been since January 2001?
Go right ahead. I was watching Oracle’s stock price just after they announced the lawsuit. That is, the actually relevant period.
@Patrick> Andy probably wasn’t lying. The Gartner article I quoted from earlier says Android shipments for the quarter were 20.5M.
I assumed so as well. But 20.5M isn’t 25M, as @Winter wanted to state. @Winter also states, “AT&T sold 5 million iPhones in Q3.”
This figure is actually 5.3 million.
For the quarter ended September 25, 2010, Apple sold 14.1 million iPhones, world-wide. (ref: http://www.apple.com/pr/library/2010/10/18results.html) in that same time period
The issue is this: In an straight-up comparison, in Q3 the iPhone at AT&T outsold Android at Verizon by a factor of 2.5. The point here being that these two carriers represent approximately the same addressable market. That’s a fact that not even ESR can argue away.
@Morgan> Every McDonald’s here in Florida has them. Ever been to McD’s recently? They have those POS card readers that say “Visa PayPass” on them? That’s what the Nexus S has. Now how many places have these? I don’t know. But I suspect that since credit cards already have this, there will be more and more at retailers near you.
PayPass is a well-known MasterCard standard, and many banks already issue PayPass credit cards: http://en.wikipedia.org/wiki/PayPass#PayPass
Visa’s equivalent is PayWave: http://en.wikipedia.org/wiki/Visa_Inc.#PayWave
Unfortunately, “That’s what the Nexus S has.” is an incorrect statement.
You are missing a couple things to be able to do a Visa PayWave or MasterCard PayPass transaction with you shiny new Nexus S:
1. A compatible SIM card (or other Secure Element) containing the PayWave/PayPass application
2. Your payment credentials securely stored in that SIM so that the transactions hook into your account.
Note that the NXP part can do “Card Emulation”, so technically, with these missing pieces supplied, it should be able to access PayWave/PayPass.
>>> Sorry, “Just Works” is a joke in this context following the antenna fiasco. If Apple lives up to its current high standards for wireless implementation, their NFC implementation will only work for transferring money to phishing sites and Nigerian scam artists.
This is such bs. Most people have already forgotten about “Antennagate.” It was a non-story hyped by a lot of people who have an axe to grind against Apple.
I gotta go by what I see, not by what some guy says on his blog. And what I see is that my phone works just fine; never dropped a call. I suspect the vast majority of users would report the same.
Android is weaker than you think, Android has nothing as good as the Apple store and is particularly weak in games.
And I would be willing to bet that Apple’s implementation of NFC will work just fine.
@esr> What’s even worse for Apple is that with the release of NFC capability in the Nexus S the iPhone is now playing technological catch-up.
Really, Eric?
The iPhone 4 (and latest generation iPod) shipped with a gyroscope months before the Nexus S shipped with one. Was Google “playing catch-up?”
Vendors were already announcing NFC ‘add-ons’ for iPhone over the summer: http://www.nearfieldcommunicationsworld.com/2010/05/04/33575/breaking-visa-to-bring-nfc-mobile-payments-to-the-iphone/
Apple has filled probably a dozen patents for using NFC on a phone. I know of 3 in September 2008, (20100082444, 20100082485 & 20100082784), one in August 2009, and one in September 2009. There have to be others. You can google for them if you like.
Now, don’t go gettin’ all butt-hurt when Apple shuts down NFC on Android for patent infringement, Okay?
“What’s even worse for Apple is that with the release of NFC capability in the Nexus S the iPhone is now playing technological catch-up.”
NFC is a solution looking for a problem. Need to be able to bump phones to exchange contact information with one or two clicks or a spoken passphrase, need to be able to bump a POS device to make a payment with a spoken passphrase.
NFC on the Nexus is half a solution.
> It was a non-story hyped by a lot of people who have an axe to grind against Apple.
Agreed.
> And what I see is that my phone works just fine; never dropped a call. I suspect the vast majority of users would report the same.
To be perfectly fair, Apple did admit that they’d rigged the signal indicator wrong. And Jobs did his ‘mea culpa’.
> Android is weaker than you think, Android has nothing as good as the Apple store and is particularly weak in games.
Agreed. I know of very few people who own both a current-generation Android phone and current-generation iPhone. I do (and I do software dev on both.)
The iPhone is a clear winner. I only carry the Android phone (a Nexus One) for testing purposes, oh, and it makes a better car GPS.
Hey, Jezus Lizzard. What esr says here:
is 100% true. You don’t even need to take esr’s word for that. You can go look for yourself over at Groklaw. where they have every single filing ever posted to the legal database services. I can’t be quite as familiar with the case as esr is, since he was personally involved, but I followed the blow-by-blow coverage on Groklaw daily up until the point that it was pretty clear that SCO would lose. Every statement esr has made here is 100% verifiable over there.
As for the PayWave/PayPass thing, yes, I knew all that. Regarding Apple’s NFC patents, does Apple really, really want to go toe-to-toe with Google in an all-out patent war? I’m sure we all remember the last time Apple tried to pick a patent fight. *blink*
> You can go look for yourself over at Groklaw. where they have every single filing ever posted to the legal database services.
I’m more than familiar with the Groklaw postings, thanks. There were many facets to the suit, and Eric only addresses one.
Still, SCO attempted to assert that it owned Unix. At one point, SCO “canceled” IBM’s Unix license for AIX. Eventually Novell interceded, based on both the court’s findings that Novell still owned the copyrights to Unix, and a section of the asset purchase agreement between Novell and SCO.
On August 10, 2007, Judge Kimball issued a ruling which says in part “the court concludes that Novell is the owner of the UNIX and UnixWare copyrights.” The ruling stated that Novell, “is entitled, at its sole discretion, to direct SCO to waive its claims against IBM and Sequent”. A month later, SCO filed for Chapter 11, and a month after that, they terminated Darl. http://www.sec.gov/Archives/edgar/data/1102542/000114420409053428/v163103_8k.htm
So I stand by my former statement. “This (SCO .v IBM) failed because, as it turns out, SCO didn’t own the copyright to Unix. Novel did.”
Because when the Copyright holder shows up and says, “You’ve got a problem”, its almost always game over. This is true both when the (true) Copyright holder shows up and tells the judge, “we don’t have a problem”, as in SCO .v IBM, as well as when the true Copyright holder shows up and says, “we didn’t give them permission for that” (Oracle .v Google).
>So I stand by my former statement. “This (SCO .v IBM) failed because, as it turns out, SCO didn’t own the copyright to Unix. Novel did.”
ROTFL. Lizzard, you’re precious; don’t ever change, you’re far too entertaining as you are.
Someday, the world will know why I am laughing. I must have some conversations to find out how soon that can be.
And hey, since you brought up Groklaw, let’s quote today’s front page there:
Eric is about to get his head slammed against the lockers by the school bully.
Which, BTW, is my killer app for the device and the main reason I have one.
Of course, it would be really nice if there were a port of gpsd for Android so I could interface my laptop. If only I knew where to find gpsd’s chief software architect ….
…Now, don’t go gettin’ all butt-hurt when Apple shuts down NFC on Android for patent infringement, Okay?…
Seriously? You want to go on record with this kind of childish gumph?
Obviously Apple doesn’t mind, since it was quite willing to file suit against HTC this spring. Everyone’s suing everyone in the mobile space right now.
> Go right ahead. I was watching Oracle’s stock price just after they announced the lawsuit. That is, the actually relevant period.
Right… lets look at some data, Eric. What ‘cha say?
http://finance.yahoo.com/q/hp?s=ORCL&a=07&b=10&c=2010&d=11&e=14&f=2010&g=d&z=66&y=66
As you can see, Oracle’s share price was already falling, prior to the suit being filed.
On 10 Aug, ORCL closed at $24.29. 11 Aug: $23.66, 12 Aug: $22.94, and $22.66 on 13 Aug (the day after the suit was filed, though the markets had closed on 12 Aug prior to the filing of the suit. http://www.marketwire.com/press-release/Oracle-Files-Complaint-Against-Google-for-Patent-and-Copyright-Infringement-NASDAQ-ORCL-1304265.htm)
Subsequent to 13 Aug, the stock closed *UP* for 3 subsequent days, ($22.72, $23.05, $23.09), then fell off a bit, reaching a “low” of $21.85 on 31 Aug, and has been on an over-all gain ever since.
Again, the suit was announced after close, so all you can really be talking about is 13 Aug, when yes, the stock continued it’s gradual slide, and then it closed *UP* in three subsequent sessions.
>Again, the suit was announced after close, so all you can really be talking about is 13 Aug,
Well spotted.
…it would be really nice if there were a port of gpsd for Android …
There are some hinky issues, but supporting Android & Maemo/Meego is totally on the radar
>There are some hinky issues, but supporting Android & Maemo/Meego is totally on the radar
Dan: I know you’re affiliated with Meego. The GPSD project would be happy to cooperate with you guys on the port.
> Which, BTW, is my killer app for the device and the main reason I have one.
There are similar apps for the iPhone, but they cost real money, and aren’t quite as good. Face it, the iPhone and Android are going to kill consumer GPS.
> Of course, it would be really nice if there were a port of gpsd for Android so I could interface my laptop. If only I knew where to find gpsd’s chief software architect ….
I think you’d get farther, faster if you managed to write an Android app that makes the GPS available over Bluetooth, and just use GPSD on your little linux laptop. There is some decent work in mytracks to borrow/copy/interface to. (Mytracks provides an intent for a recorder service.)
> Someday, the world will know why I am laughing. I must have some conversations to find out how soon that can be.
Do let us know when you can de-cloak, dear leader.
>Do let us know when you can de-cloak, dear leader.
None of your observed character flaws include stupidity, Lizzard, so I’ll just put it this way. If you imagine the absolute worst case for anyone arguing about the SCO litigation with me – that is, you form a theory about my knowledge and role that is as totally embarrassing to anyone else’s claim of superior understanding as it could possibly be (and I mean anyone else on the planet) – then you will probably come somewhere close to the actual truth.
@TJL:
Well, I did discuss earlier that a lot of those were probably upgrades, because the comscore survey only shows 2.3 million or so new iPhones. But let’s see what AT&T themselves have to say about it.
In July, they said At the end of the quarter, 53.2 percent of AT&T’s 67.0 million postpaid subscribers had integrated devices.
Last quarter, they said:
OK, what can we deduce from all this? (I’m going to assume that basically all iPhones are postpaid, unless somebody can show me how wrong this is…)
1) 1.24 million new subscribers for AT&T on iPhone. Makes sense to chalk these up to Apple. iPhone good; AT&T bad. Incidentally, that would be almost all their new subscriber growth in postpaid: 700K increase + 770K churn – 1.24M iPhones = 230K other.
2) Total postpaid integrated device subscribers increased by 3.1 M (by doing the math from the two press releases), but there were “more than 8 million” postpaid integrated device activations. This is eminently consistent with having lots of replacement activations — 5 out of every 8 activations were a replacement of some kind. Even if you completely discount the comscore report, and believe with all your heart that all of AT&T’s increase in integrated device subscribers is due to iPhones, you would still have to admit that over 2 million of them are replacements, AND you would also then have to admit that on AT&T’s network, 3 million non-iPhone smartphone customers upgraded to non-iPhone smartphones…
3) On the only network where you can go if you want an iPhone, on the network that doesn’t even try to sell non-iPhone smartphones, during the quarter with the iPhone 4 rollout, where AT&T let people upgrade iPhones early, over 35% of integrated device activations were not iPhones. If you subtract out the new subscribers, over 41% of integrated device activations were not iPhones.
4) If we believe the comscore report, and subtract 1.2 million new AT&T iPhone customers from the total 2.3 million increase in iPhones for the quarter that comscore reports, that means that around 1.1 million internal customers are newly-minted iPhone carriers. But 800K internal customers are newly-minted NON-Apple smartphone carriers, and then you still have around 2.2 million smartphone -> smartphone upgraders who opted not to go for the iPhone.
5) Comscore (which is, AFAIK, pretty well-respected, and polls actual people who carry phones) shows total US Android growth (users) for the quarter at over 2X iPhone user growth. You have pointed to a couple of articles that all point back to a report by ITG, which is based on a survey of independent retail outlets, which is really pretty bogus. Even ITG itself disclaims the last article you pointed to and pointedly claims not to have any inside information from Verizon, which doesn’t seem to have released any phone sales information itself.
BTW, when you say stuff like “And I gotta say, his analysis seems a bit more based in fact than yours” and “That’s a fact that not even ESR can argue away” with zero evidence other than articles that even the slanted study authors don’t like, it doesn’t really help your credibility any…
> Even ITG itself disclaims the last article you pointed to and pointedly claims not to have any inside information from Verizon, which doesn’t seem to have released any phone sales information itself.
Please. That wasn’t a disclaimer, that was restatement. “”To be clear, we do not misappropriate or improperly obtain nonpublic information,”
You just don’t like the results, or something.
@esr> Once again, the fundamental problem Apple has is not the Nexus S itself; rather, it’s that they’re stuck in a comparatively fragile single-vendor/single-product strategy, facing an multi-vendor army that in aggregate has more financial mass, more capacity to innovate, and more freedom to innovate.
Bill Joy said it best:
The open question is if its better to have people innovate in the breadth of the platform (Android), or on top of a single platform (iOS). This really resolves to who makes money in the ecosystem, the app vendors (iOS) or the handset makers (Android).
Perhaps if Google comes up with an application architecture better than java on dalvik over linux, things will change. For now, developing on Android is worse than developing on linux. Restated (heh), subject only to the issues with Apple’s app-store restrictions, development on iOS far exceeds development on Android.
But for any number of reasons, Google will have to move away from Android (Java on Dalvik). How they do that without killing the momentum they’ve found is perhaps the most interesting subject the assembled here on this blog could discuss, if it weren’t for our hosts propensity to crow about his brilliance.
>This really resolves to who makes money in the ecosystem, the app vendors (iOS) or the handset makers (Android).
You came in late, so maybe you haven’t seen my analysis of Google’s grand strategy. You should probably read this and this for starters.
Yes, Eric, we’ve all read halloween9.html
Are you saying you know more about the suit than a certain D. Boies?
>Are you saying you know more about the suit than a certain D. Boies?
Yes, that’s actually likely – not certain, but likely.
I won’t explain the positive reasons just yet, but the negative one is that Boies was more a marquee name than an actual player. You bring someone like him in to have an impressive signature on the key briefs and in order to put a certain sort of spin on your legal posture; in this case, SCO wanted to both associate itself with antitrust action and have a PR card to play against those who understood how much of a cat’s paw for Microsoft they were being (Gates & Co. spotted an opportunity and slipped ’em somewhere between $66 and $166 million under the table to fund the continuing litigation).
You do not expect a Boies to get deep into – and I’m being entirely hypothetical for now – the code forensics of the evidence. You don’t even necessarily expect him to have specialist knowledge about IP law; that’s what the actual specialists on the team are for. It can easily be the case that someone else who is (again hypothetically) code-literate, legally literate, an expert on several relevant areas of history, and up to his eyeballs in the primary evidence evaluation would end up with a better grasp of the litigation’s nuts and bolts.
In fact, I think I know of more than one person who could fit that description. One of them was getting paid by SCO, and if I ever meet him FTF I’m going to buy him a beer for doing so clever a job of sabotaging “his” side’s attempts to flimflam the court that the lawyers at SCO failed to notice that his testimony was actually damaging them. And no, you don’t expect a Boies to notice that, either, necessarily; he lacks the required domain expertise. (I have a low opinion of Boies’s general competence, but that’s for unrelated reasons.)
If any of this surprises you, it just means you haven’t had enough experience of how litigation at that scale is actually run. I…have.
@TJL
“I pointed out that Andy Rubin tweeted, “There are over 300,000 Android phones activated each day.” Assuming Andy isn’t lying, the *current rate*. A local maxima, if you will. That doesn’t mean or even imply that Google activated 25 million phones during Q3. What are you smoking?”
So, Andy says 300,000 activations a day, you make that 9 million a month, and I make it 25 million a quarter. I am perfectly aware that 300,000 on a day does not make 27 million a quarter. I was simply rather amused you were calling out 9 million Android activations a month then and are now claiming Android is a complete failure and doomed here.
I read in another comment it was 20.5 million. So 25 million was just 25% off, not bad given the paucity of data.
And I do not smoke, thank you. Any hallucinations are purely organic.
Another question. Ever seen the streets in China? You get the impression that you see more cell phones on sale than Chinese in the streets. Any number of wannabee iPhones with very creative spellings of iPhone and Apple. But many, many others. On the streets mostly cheap stuff, in the shops more expensive ones.
At some point it will be cheaper to make a low powered Android phone than any other type. Simple mass-market effects. I can perfectly see how Apple will remain the ultra-cool phone, and the other 95% of the Chinese market going to cheap and ultra cheap Android.
If, on the contrary, Apple takes 60% of the market, it will not be ultra-cool anymore. Simple sociology. So what will become ultra-cool then?
Maybe I should again to point at the elephant in the room: Nokia.
Symbian is still the biggest selling smarphone OS…
http://www.gartner.com/it/page.jsp?id=1466313
From your article :-
Symbian has no clear vision and has just closed down the closest thing it had to a vibrant multi-player ecosystem.
>Symbian has no clear vision and has just closed down the closest thing it had to a vibrant multi-player ecosystem.
Yes. Symbian is in about the the worst possible fix for a sustaining technology: it’s aging, inflexible, an attempt to fix its problems by open-sourcing it has failed, and Nokia has shown no sign of any clue about integrating it with Maemo/MeeGo – though, to be fair, I suspect this this is not an indication of incompetence but rather a limitation arising from the technology itself. It’s got “dead end” written on it in letters of fire.
> Yes, that’s actually likely – not certain, but likely.
I continue to enjoy how your answers are always couched in a hypothetical, or some other rhetorical device that provides plausible deniability or otherwise evades making any definitive statement.
> Symbian is in about the the worst possible fix for a sustaining technology.
And it don’t run Unix. Unix uber alles!
>I continue to enjoy how your answers are always couched in a hypothetical, or some other rhetorical device that provides plausible deniability or otherwise evades making any definitive statement.
The rules of civil procedure require it in this case. You do not understand how, and it is not necessary that you understand, but someday all will be explained.
If you want to predict the future, and cannot make the future yourself, look at Asia.
And I do not think these “Asians” will be intimidated by the USA patent “protection” racket.
50% of Smartphones Sold Last Quarter in China Running Android (December 3, 2010)
https://webcache.googleusercontent.com/search?q=cache:0dNE66LbZrEJ:www.androidguys.com/2010/12/03/50-smartphones-sold-quarter-china-running-android/+&cd=1&hl=en&ct=clnk
Google cache, original not available at the time of writing
Smartphone Sales In Asia On The Rise, Android Tops Symbian (November 24, 2010)
http://www.mobilecrunch.com/2010/11/24/report-asia-android-smartphones/
50 Percent of Smartphones Sold in China Last Quarter Run Android (December 2, 2010)
http://digitaldaily.allthingsd.com/20101202/50-percent-of-smartphones-sold-in-china-last-quarter-run-android/
iPhone: 5% of Third Quarter Smartphone Sales in China (2010.12.0)
http://www.displayblog.com/2010/12/03/iphone-5-of-third-quarter-smartphone-sales-in-china/
There are some rough and tumble works on what is and is not a smartphone, but the trends are clear, it seems.
@esr> You came in late, so maybe you haven’t seen my analysis of Google’s grand strategy. You should probably read this and this for starters.
I just quote you on the first link: “[Google’s] goal is to create the business conditions that will maximize their ad revenue not just two years out but ten years out.”
Didn’t I just say that? Didn’t you just object?
ESR says: No, you didn’t, and no, I didn’t. You’re hallucinating again.
> And I do not think these “Asians” will be intimidated by the USA patent “protection” racket.
In this, I’m sure you are correct.
> At some point it will be cheaper to make a low powered Android phone than any other type. Simple mass-market effects.
Hey, wait, Android is *software*. Let me change two words of your statement, and see how false it rings.
At some point it will be cheaper to make a low powered Linux computer than any other type. Simple mass-market effects.
Could Apple, *IF* it wanted-to, run a version of iOS on nearly any currently-produced ‘Android’ phone? Of course it could.
Eric does have something right. Google is playing for ubiquiti, and Apple… is not.
Apple makes revenue on each and every iPhone sold. Google makes nearly nothing, if not nothing, on most Android-powered phones. (They probably do make some on the Nexus series, and who knows what kind of support contracts they’re managing to extract from the OHA members.)
Most android apps are free. Google wants it this way. RIM doesn’t even have a decent developer program.
@esr> I am a very typical smartphone consumer in this respect – most of my apps are games.
You run Windows? :-)
In the end what matters is applications, not how open the platform is. Who will be able to woe most developers?
Currently I still see apps come out of iPhone first, and for the moment not a single one of my favorites is even available on Android. The comments I hear from developers is also that the iPhone is a lot easier to develop for than Android, as the fragmentation of Android poses quite a few challenges.
Eric, you’re willfully ignoring various important facts.
1) iPhone owners download nearly 2x as many apps (40 vs. 25)*, and are much more likely to pay for them (something like 8:1 IIRC) We can speculate as to the reasons for this disparity, but it’s real.
2) On the app store you’re also selling to iPod touches (of which there are well over 50 million) and iPads (~10 million).
3) Developers who have released identical apps for both platforms (such as Angry Birds) reveal that Android nets a small percentage of their profit. This is true EVEN THOUGH the size of the Android market is much smaller, so you would expect less dilution for popular apps – in other words, the android version should make more profit but it doesn’t.
4) If every developer switches to android, the dilution benefit disappears.
It may eventually be the case that app developers are forced to switch to Android, and in fact I wouldn’t be surprised in the least, but it is not happening in the short term. A strategic bet is one thing, but right now it should be a hypothetical bet rather than a practical one. (Saying this as an app developer, rather than as a phone owner).
*http://www.appconsumer.com/app-downloads-per-user-on-the-rise-games-most-popular-app-category/
>Eric, you’re willfully ignoring various important facts.
You misspelled “evaluate the significance of the facts differently”.
@TJL
“At some point it will be cheaper to make a low powered Linux computer than any other type. Simple mass-market effects. ”
You nailed it. That is it indeed.
So now we see masses of cheap routers, set-top boxes, navigation system etc. all running Linux. The netbooks started that way. There was no iOS or Windows eeePC, but a Linux one, until something happened to the producers that made them apologize in public for even showing Linux on some of their devices in public.
http://blogs.computerworld.com/microsoft_strikes_back_at_linux_netbook_push
http://www.zdnet.com/blog/open-source/sorry-linux-but-the-chicken-came-first/4311#more-4311
Obviously, Apple can give away iOS to cheap phone producers for free. And then there might be ultra cheap iPhone clones running iOS. Any bets that will happen?
So what will be cheaper in a year? Buying OTS components and assemble them into a phone that runs stock Android, or doing so to make a Windows/Symbian/iOS phone?
But you did not get that point, I think.
@ Eric
>You imagine a lot of things that aren’t true. I am a very typical smartphone consumer in this respect – most of my apps are games.
Even if we assume that the applications you have on your phone are representative in genre to a typical smart phone consumer, by your very nature of being who you are, you are not a typical smartphone consumer. The process by which you determine both what phone to get and what applications to download and possibly pay for is a different process from most consumers. The simple fact you would never consider a iPhone purchase on political reasons alone puts you outside “typical smartphone consumer”
>You’re entitled to that opinion, but judging by how they market Apple does not appear to share it.
So it is your contention that there is a sizable segment of iPhone owners who would not have bought the iPhone were it not for the Retina Display? Or alternatively, they would have bought an Android phone had the android device had a retina display?
>Quite right. Now try this with any application where Microsoft hasn’t been forced kicking and screaming into offering a portable format.
>One of the many, many problems with closed-source operaing systems is that they encourage data jailing by making obscure,
>undocumented data formats the rule rather than the exception.
Do you have any evidence of this at all? I just don’t see how it logically follows that the use of a closed source OS encourages the use of closed / incompatible data formats. I do see how being the dominant OS encourages this, but dominance != closed source.
>So it is your contention that there is a sizable segment of iPhone owners who would not have bought the iPhone were it not for the Retina Display?
Why yes. Yes, actually. And we can tell that Apple agrees with me by their actual behavior. You don’t pour marketing dollars into propagating a meme unless you think it’s going to net you sales.
>Do you have any evidence of this at all? I just don’t see how it logically follows that the use of a closed source OS encourages the use of closed / incompatible data formats. I do see how being the dominant OS encourages this, but dominance != closed source.
Wow. Are the kids forgetting so quickly?
What you just told me is that you have no experience of the variety of operating systems before 1990. Before the Windows monopoly and the PC monoculture, there was a much larger variety of operating systems than today. I worked with at least six different operating systems on early PCs alone, not counting the mainframe and minicomputer ones – tot those up and the number rises to around 20. And I mean really different, not ports of genetically related codebases. None were “dominant” in the modern sense; they couldn’t be, the hardware space was too fragmented.
And yes, there was a very direct and obvious correlation between the extent to which you could get source for the OS and the extent to which it used inscrutable proprietary data formats. Did you think RMS’s complaints and my later activism came out of nowhere? We were reacting to real, existing conditions that sucked. Your blithe ignorance on this score is rather dizzying, but I suppose I will have to get used to encountering it as those years recede further into history.
The desirability of a product is determined entirely by marketing and brand image.
If you actually believe that, then you should never, under any circumstances, attempt to run a business.
Obviously, Apple can give away iOS to cheap phone producers for free. And then there might be ultra cheap iPhone clones running iOS. Any bets that will happen?
Apple isn’t going to cut their throats by licensing iOS. What they will do, is lower the entry-level price point over time, as they did with the iPods, while keeping the price for the top of the line at around the same level as it is today.
That seems very, very unlikely. Apple’s been down that road before. It’s not one Steve Jobs likes. Windows Phone could fill that role, except it doesn’t seem to be doing very well.
>[Licensed iOS] seems very, very unlikely. Apple’s been down that road before. It’s not one Steve Jobs likes.
I agree. It would be completely incompatible with Apple’s strategic positioning, too.
>So it is your contention that there is a sizable segment of iPhone owners who would not have bought the iPhone were it not for the Retina Display?
That was the biggest draw for the phone to me. Unfortunately, it’s still not enough to overcome my strong personal dislike of AT&T.
I still considered buying one for a week, just for that display.
> You misspelled “evaluate the significance of the facts differently”.
Call it what you like; either way, you’re making a significant mistake.
>>I continue to enjoy how your answers are always couched in a hypothetical, or some other rhetorical device that provides plausible deniability or otherwise evades making any definitive statement.
>
> The rules of civil procedure require it in this case. You do not understand how, and it is not necessary that you understand, but someday all will be explained.
More ducking! You don’t even explain which part of the RoCP are in-force here.
Love it. Don’t ever change, Dude!
>Retreating upward looks smart because it improves margins; shareholders love this. It can buy you a few good years. Trouble is, there are toxic side-effects of the strategy that accumulate and eventually kill you. For detailed discussion of one of the most imporant, go read Clayton Christensen’s The Innovator’s Dilemma.
Can you escape this by constantly reinventing yourself every 5 years or so? A “invent, retreat upward, abandon” cycle.
>Can you escape this by constantly reinventing yourself every 5 years or so? A “invent, retreat upward, abandon” cycle.
It’s theoretically possible, I suppose. Apple has pulled it off twice (at much longer than five-year intervals); most companies that try can’t make it happen even once. The practical problem is that reinventing your company is difficult; it has to be both willing and able to abandon the corporate knowledge at the center of its value proposition in order to learn an entire new set of competences. This requires exceptionally strong leadership, risk-tolerant investors, and quite a bit of plain luck.
Pay attention to me. I’m lonely.
ESR ESR ESR ESR ESR
Look at me. I’m pretty. Talk to me. Indulge my open-ended demands for detail. Jump into my rabbit hole. I’m important. I’m significant. You need me in your life.
[/channeling JL]
>Look at me. I’m pretty. Talk to me. Indulge my open-ended demands for detail. Jump into my rabbit hole. I’m important. I’m significant. You need me in your life.
Yup, you’re describing a common variant of haterboy. I need to write that taxonomy-of-haterboys post
> So now we see masses of cheap routers, set-top boxes, navigation system etc. all running Linux. The netbooks started that way.
Sorry? Is this where I point out that Linksys came to strip the WRT54G down to 2MB of flash, with VxWorks on it, and then raised the price on
the 4MB flash variant, which ran linux?
The BOM add for vxworks on a $50 product is approximately $0.50/unit. If you can save more than that in flash/ram costs (shaved BOM), its a win.
Never mind the support costs from all of your customers constantly bricking the unit in their attempts to ‘hack’ it.
Or should we talk about Microsoft suing TomTom?
Thanks for playing.
> There was no iOS or Windows eeePC, but a Linux one, until something happened to the producers that made them apologize in public for even showing Linux on some of their devices in public.
I think that something was the ENORMOUS RETURN RATE! (Some 4X that of ‘normal’ laptops.)
Then you have to consider the entire “business model” of your normal notebook/netbook ODM.
Asus came out with the EEE 701 and it was very cheap, and very successful, and did what it was made to do (web access, skype, etc…. NOT run freaking Photoshop and Office). Everything you ‘needed’ was **pre-installed**, working as intended.
Then all other manufacturers jumped in the bandwagon, no one wants to miss the party.
But “they” were scared by the monster Asus had created. Where did the nice, fat margins go!? Where is the MS protection racket “money”? The money from all the “free” crapware loaded into every MS computer to protect it from itself (or not), where did it go?
And look, all these people complaining that their coffee maker does not wash dishes… add more CPU (and get an Intel part, so Intel will pay us for the stickers on the case!), add more RAM, more disk/SSD!
And then Microsoft came looking at a new, suddenly successful line of “PeeCees” that didn’t run Windows! Shock! HORROR!! Quick, ready the ‘special’ version of Windows XP. NO, NEVERMIND THAT WE WERE TRYING TO GET RID OF IT!!!
> So what will be cheaper in a year? Buying OTS components and assemble them into a phone that runs stock Android, or doing so to make a Windows/Symbian/iOS phone?
Someone doesn’t understand manufacturing at volume. It’s nothing like going down to the PeeCee parts store and selecting the components for your new LAN Party box.
@Dan> ESR ESR ESR ESR ESR
NOP NOP NOP NOP NOP NOP ….
http://bit.ly/jz93s
@TJL:
Well, I should have left that little dig out, for the simple reason that it allowed you to pretend like you answered me, while studiously ignoring the fact that your claim that 5 million iPhones is significantly more than whatever number somebody pulled out of their ass of Android phones is meaningless, for the simple reason that most of those iPhones were upgrades. In other words, your claim that “That’s a fact that not even ESR can argue away”” is bullshit. It is eminently possible to reconcile the huge domestic growth in Android usage vs. with 5 million new iPhones, and show that at least one of the conclusions that you parroted from that article: “Android is not competitive vs. iOS” — is simply incorrect. After all, if 40% of internal AT&T subscribers who activated a smartphone last quarter didn’t activate an iPhone, then something is competitive with iOS, and unless you can show me how RIM or Symbian is more competitive than Android, I gotta believe it’s mostly Android.
>In other words, your claim that “That’s a fact that not even ESR can argue away”” is bullshit.
Sometime the best way to win an argument like this is to just wait 90 days. Barring a dramatic reversal of trends in the last year, Android will have put enough daylight between its rising market share and iOS’s stable or falling share to be statistically significant.
@TJL
“Or should we talk about Microsoft suing TomTom?” And following
And your point is? Linux allowed cheap tech and that is good/bad/irrelevant? It has been banned as illegal by MS? I have no clue what you are talking about.
I must say it must have been an unreal experience, seeing a CEO of a company apologizing in public for showing a “must have” gadget at a big vendor show, more or less promising not to do it again.
@TJL
“> So what will be cheaper in a year? Buying OTS components and assemble them into a phone that runs stock Android, or doing so to make a Windows/Symbian/iOS phone?
Someone doesn’t understand manufacturing at volume. It’s nothing like going down to the PeeCee parts store and selecting the components for your new LAN Party box.”
No it is going to a Chinese warehuose and order a container of parts. And it matters if you can tick a web order form with parts you assemble on your production line and flash a stock image on with all drivers included. And those parts will be cheap(er) if they are produced in bulk anyway.
How many months did it take MS to get Windows booting on the OLPC XO?
Indeed, I do not understand manufacturing at volume.
> After all, if 40% of internal AT&T subscribers who activated a smartphone last quarter didn’t activate an iPhone, then something is competitive with iOS, and unless you can show me how RIM or Symbian is more competitive than Android, I gotta believe it’s mostly Android.
Maybe you missed the part where AT&T was deliberately pushing only Symbian as its only ‘smartphone’ for quite some time.
(other than, for some reason, iPhone. Oh yeah, right.)
Or maybe you missed the part where AT&T shipped its first Android phone, (the Moto Backflip) sans support for the Android Market?
Or that AT&T has 3 other Android phones now, priced at $99, $129 and $199? (Wait, how much is an iPhone again? http://bit.ly/9NqFrb)
But it pushes the ‘refurb’ (read: returned by customer) models at $9.99 or less? http://bit.ly/frbGCy while ‘refurb’ iPhones are $50?
Or that AT&T doesn’t even list iPhone in its smartphone category? http://bit.ly/gaqHJG
Or maybe you’re just desperate to make a point.
> I must say it must have been an unreal experience, seeing a CEO of a company apologizing in public for showing a “must have” gadget at a big vendor show, more or less promising not to do it again.
Follow the money! He was forced to capitulate!
(Winter may want to note that I own both a eee PC 701 and a eee PC 1000H. Neither has been as much as turned on in the last year.)
> Sorry, “Just Works” is a joke in this context following the antenna fiasco.
The logic of the “fiasco” is pretty clear in retrospect:
(1) IF my phone had bad reception today AND nothing has obviously changed recently in my hardware THEN I blame the overloaded network. Curse you, AT&T!
(2) IF my phone had bad reception today AND my antenna has a new design nobody knows anything about THEN I blame the new hardware. Curse you, Apple!
Choosing to make the antenna visually stand out rather than hiding it in a plastic shell like everybody else made it a potential focal point for complaints. Given millions of new customers in a week or two, some of them are bound to have complaints about reception; some of those complaints will fixate on the design; the internet rumor mill makes it hard to contain or react to that sort of story once it starts. It was a good story. But the story wasn’t true, which is why Apple stopped giving away the free cases after a few months.
Just Works is actually a pretty good description of the new antenna design *as an antenna*. New customers just wanted reassurance that any signal problems they happened to encounter weren’t caused by the new antenna design or could easily be mitigated if they were. Putting the phone in a case provided that reassurance; it defused the salience bias resulting from the new design just long enough for people to get used to it and stop thinking of antenna-on-the-outside as a Scary New Thing.
> How many months did it take MS to get Windows booting on the OLPC XO?
Booting? Probably no longer than your average linux hacker takes to get linux on a new PeeCee.
Tested & Supported? With support for all the languages, etc?
Well, son. Those things take time.
> No it is going to a Chinese warehuose and order a container of parts.
This isn’t how its done, sonny. “Inventory is waste.”
If you’re a big player, you make the sub-assembly manufacturers stock the parts at *YOUR* warehouse,
which is right by the assembly line. And you don’t pay them for 30 days after you’ve picked the part off the shelf in *YOUR* warehouse.
> And it matters if you can tick a web order form with parts you assemble on your production line and flash a stock image on with all drivers included.
Boy howdy. You are green. “Flash a stock image”? Of Android? Why are you burning flash storage on drivers you don’t need? Why didn’t you compile the firmware for the ARM part you have, enabling Thumb if you can support it?
Oh yeah, all the world is a PeeCee, and we all have N Gigabytes of RAM, and M Gigabytes of disk.
(punters…)
> And those parts will be cheap(er) if they are produced in bulk anyway.
Is this where I sit you down and show you the POs I’ve issued in the past year for 50K or more parts?
Jump into my rabbit hole. Please.
I thought I’d get a response on that. You’re still nibbling around the edges, refusing to admit that less than half of those 5.2 million iPhones were for new iOS customers.
What does that have to do with whether Android is competitive against iOS?
You seem to think this would be news to me. The whole point is that Android is going to capture all of the low end, along with perhaps some of the high end. Or maybe your definition of “competitive” doesn’t include the ability to sell lots of units. But then, why are you crowing about Apple selling lots of units?
In the quarterly report, they’re all in the “integrated device” category. But I see what you’re saying — AT&T didn’t sell many iPhones to people who weren’t already carrying them because people looking on the AT&T website for a smartphone wouldn’t find the iPhone. It’s a big secret that they sell it, and a shame that their marketing fucked up that way, otherwise Apple would be doing a lot better.
Keep writing — you’ve made most of my points for me.
And hey, these are intentional radiators. Emitters of RF energy sufficient to fall under the legal mandates of the Friendly Candy Company.
Y’all know ya gotta test these devices *on* the manufacturing line, for compliance, in order to not fall astray of several federal laws.
Right? RIGHT???
@TJL
“Winter may want to note that I own both a eee PC 701 and a eee PC 1000H. Neither has been as much as turned on in the last year.”
Winter has a eeePC 4G running XFCE Linux Mint which he carried with him while he traveled across Eurasia (albeit, it ran Cruncheee most of the time). He hooks it up to a screen and keyboard and happily runs everything from Firefox and Open Office to LaTex and git. He uses it to do his work when he is tired of lugging around his ultra-cool high end laptop.
> The whole point is that Android is going to capture all of the low end, along with perhaps some of the high end.
In this, we agree.
> Or maybe your definition of “competitive” doesn’t include the ability to sell lots of units. But then, why are you crowing about Apple selling lots of units?
Where was Android (Danger) prior to Google/T-Mobile attempting to fend off iPhone with Android?
AT&T, Verizon, Sprint and T-Mobile are all wary of the smartphone vendors. Both Apple and Google are attempting to wrest control of the mobile space away from the carriers. All the US carriers said, “No” to both Apple and Google/Android prior to Apple finally tipping AT&T.
Did Sprint and Verizon immediately jump to Android? NO, no they did not. There inherent fear of the Internet industries kept them from an effective response for many months. T-Mobile jumped, but they have a huge churn problem, and maybe they saw Android as a way to reduce the churn. (Its worked, btw.)
Only after the iPhone was a huge success, and T-Mobile started to succeed with Android and the G1 did Motorola finally gain purchase with Android in its ‘Droid’ line.
ESR: Symbian is in about the the worst possible fix for a sustaining technology: it’s aging, inflexible, an attempt to fix its problems by open-sourcing it has failed, and Nokia has shown no sign of any clue about integrating it with Maemo/MeeGo – though, to be fair, I suspect this this is not an indication of incompetence but rather a limitation arising from the technology itself.
Nokia is now pushing a version of QT toolkit that is portable across Symbian and Maemo; that’s probably about the best they can do.
@TJL
“> How many months did it take MS to get Windows booting on the OLPC XO?
Booting? Probably no longer than your average linux hacker takes to get linux on a new PeeCee.
Tested & Supported? With support for all the languages, etc? ”
OLPC in the News (Part 2)
“As part of this engineering effort, we have to design a new BIOS – the layer of software that runs between the hardware and an operating system — to have Windows boot and run off the SD card. For us this is new work and requires a design and processes for supporting the XO’s custom SD interface and for the installation of Windows on the SD card, both at the Quanta factory that manufactures the XO hardware and also in the field. ”
”
“Secondly, as we all know there are many innovative features in the XO computer that set it apart from other designs, and we are working with partners to write the driver software so that Windows can support all of them. This includes drivers for the XO’s wireless networking, camera, graphics processor, audio system, and the various user input devices (game pad, writing pad, touch pad, directional pad, and mouse pad.) There are ten custom drivers in all that we are writing. We also hope to support the XO’s mesh network design, its power-saving “e-book” mode, and its capability for excellent screen visibility in full daylight. ”
http://www.olpcnews.com/software/windows/no_microsoft_windows_xp_on_xo.html
In short: MS had to change the Bios to get it booting. There is much more about all those pesky drivers they did not have for windows, and could not write. And then there was some rumors of MS engineers not being able to understand hardware specifications.
@TJL
“If you’re a big player, you make the sub-assembly manufacturers stock the parts at *YOUR* warehouse,
which is right by the assembly line. And you don’t pay them for 30 days after you’ve picked the part off the shelf in *YOUR* warehouse.”
Obviously. That is the real world.
But if you want to be the first to produce that CheapChineseAndroidJunk, you could get by by emptying a container of last months surplus stuff into an empty production line and assemble another container of gadgets. Which only works if you can do it FAST, and without any real design effort. I was thinking about:
http://www.wired.com/gadgetlab/2010/08/android-epad-tablet-reviewed-verdict-junk/
But again, I would not know whether this can happen. Maybe in my dreams. I am a complete n00b in producing stuff.
> Where was Android (Danger) prior to Google/T-Mobile attempting to fend off iPhone with Android?
You make my head spin. This started off with you claiming that iOS market share gains last quarter, which IMO were mostly illusory, help to prove that Verizon is toast. I agree that iOS on Verizon would probably help Verizon some, but I think the advantage of iOS is rapidly eroding. A) most people who really want it probably already have it; and B) if you read the fine print on the quarterly reports, you will realize that AT&T’s phone subsidy costs per subscriber are actually not all that great for the bottom line. It’s good for Apple, though.
> Only after the iPhone was a huge success, and T-Mobile started to succeed with Android and the G1 did Motorola finally gain purchase with Android in its ‘Droid’ line.
Agreed, but in life a lot of problems carry the seeds of their own solutions. There are countless examples of markets where entrenched players all sit on their thumbs until one of them flinches, and then there is a flurry of activity. You have to give Android/Danger credit for hanging in there until they became relevant.
>B) if you read the fine print on the quarterly reports, you will realize that AT&T’s phone subsidy costs per subscriber are actually not all that great for the bottom line. It’s good for Apple, though.
This is the only reason not to dismiss AT&T letting Apple out of its exclusive early as a drug fantasy. I still don’t think it will go down that way, because AT&T is behaving as though it has decided iPhone is a net advantage despite the subsidy. But if it does, that’ll be why.
Accordingly, Apple fans should be careful what they wish for. If the exclusive does break early, it will be because AT&T has concluded that the competitive advantage of fielding iOS no longer justifies the subsidy payments.
@esr:
Interesting little juicy tidbit here: So you know about a connection to Microsoft other than BayStar Capital and Vulcan Ventures? IIRC, a connection between RBC and Microsoft was not shown. The Sun payment is also a bit questionable, but I can’t imagine Sun being Microsoft’s minion, before or after Oracle. The original BayStar investment was only $50 million. If you can’t go on record, I understand.
>IIRC, a connection between RBC and Microsoft was not shown.
Not in court, no, but none of the VCs and investment-banking types I know were in much doubt about it.
@TJL:
Many other router manufacturers are still using Linux and there is at least one router manufacturer (Buffalo) that I know of specifically using DD-WRT.
BTW, my 2MB WRT54G v5 runs DD-WRT micro, so there. :-P
@ esr
>And yes, there was a very direct and obvious correlation between the extent to which you could get source for the
>OS and the extent to which it used inscrutable proprietary data formats. Did you think RMS’s complaints and
>my later activism came out of nowhere? We were reacting to real, existing conditions that sucked. Your blithe
>ignorance on this score is rather dizzying, but I suppose I will have to get used to encountering it as those years
>recede further into history.
And you and I both know that correlation != causation, so what exactly is the mechanism which makes a closed OS automatically mean closed data formats.
>And you and I both know that correlation != causation, so what exactly is the mechanism which makes a closed OS automatically mean closed data formats.
I should think that’s pretty obvious. If you’ve already decided that some secrecy is good for your business, it’s easy to conclude that more must be better. Concealing technical information becomes a habit that you keep up even when it’s batshit insane – like, making access to interface specifications for your hardware difficult even when more openness would increase its attractiveness to customers.
> what exactly is the mechanism which makes a closed OS automatically mean closed data formats.
Rent-seeking behavior is natural. An environment where significant things are allowed to be closed signifies that it will be tolerated.
A closed monopoly O/S can spin off significant profits to be used to develop closed applications that run on it (and only on it), which can help it maintain its monopoly position.
But even before it attains monopoly status, the customers of the OS have shown that they accept closed, so if there is a sufficient customer base, then app vendors will push closed as hard as they can.
>An environment where significant things are allowed to be closed signifies that it will be tolerated.
Correct. And the flip side of this is that open source in an OS creates significant pressure for application vendors not to be evil.
@ esr
>I should think that’s pretty obvious. If you’ve already decided that some secrecy is good for your business,
>it’s easy to conclude that more must be better. Concealing technical information becomes a habit that you keep
>up even when it’s batshit insane – like, making access to interface specifications for your hardware difficult even
>when more openess would increase its attractieness to customers.
Again, we’re conflating applications and their data with the OS. Android is OSS, all of the apps that google ships? Not so much:
http://androidandme.com/2009/09/hacks/cyanogenmod-in-trouble/
http://android-developers.blogspot.com/2009/09/note-on-google-apps-for-android.html
Similarly, Adobe Acrobat? Decidedly closed source application. The PDF specification? Open: http://en.wikipedia.org/wiki/Portable_Document_Format
Noting however that PDF was proprietary and later made open, might I suggest that it is the demands of the consumer that create pressure on vendors to use open formats, not the nature of the OS. And that it is once again a correlation between the types of people who use open source operating systems and their demands that produces the correlation between open source OSes and open data formats?
>Again, we’re conflating applications and their data with the OS.
Yes, and doing so quite deliberately. Culture and expectations matter. So do legal constraints; it is often impossible to open-source applications effectively even when the vendor wants to, because they have critical dependencies on closed-source system libraries. To take a contemporary example, a lot of Mac OS X applications are in this boat.
>Android is OSS, all of the apps that google ships? Not so much:
And yet, Cyanogen still ships. Google’s response was revealing: they have an open-source-centric OS and culture, so they do business case analysis and close only what they think they need to. By observation, vendors in the ecologies around closed-source operating systems are less rational; having formed the habit of thinking of secrecy as a good in itself from the operating system’s example, they often close access to source and file formats even when it damages their value proposition to do so.
>I suggest that it is the demands of the consumer that create pressure on vendors to use open formats, not the nature of the OS.
Your error is in the belief that these can be cleanly separated. Unlike you, I have spent more than 50% of my years as a consumer and programmer in pre-open-source conditions; thus, I know by experience that the arrows of causation between OS practice and the expectations of the userbase go in both directions. It is really quite interesting to encounter someone too innocent to know this.
@TJL:
For someone who claims to know quite a bit about the history of computing, you are utterly ignorant of it. Ever wonder why there used to be, oh, I don’t know, maybe about 50-60 viable word processing programs, but these days you can count the ones that matter on the fingers of one hand?
Think about that for a bit and get to back to me.
Morgan Greywolf Says:
>Ever wonder why there used to be… about 50-60 viable word processing programs,
Cambrian explosion?
Another way of putting that would be that if an OS is open source there’s no technical reason preventing another party perceiving the closed solution as damage and routing around (which just leaves legal and resourcing reasons). A closed solution on a closed OS can reference poorly documented interfaces of the closed system (E.G. Internet Explorer).
@morgan
Mis-attribution. That was said by tmoney not TJL.
My bad. Yes, my last comment was directed at tmoney, not TJL.
@Jessica Boxer:
Are you seriously trying to draw a parallel or were you just trying to funny?
Well, I thought it was funny. And, perhaps, apt. An environmental event that many couldn’t adapt to — you know, like having their oxygen cut off.
>Well, I thought it was funny. And, perhaps, apt. An environmental event that many couldn’t adapt to — you know, like having their oxygen cut off.
I thought it was funny, too. That and labeling herself “Corporate Code Monkey” in ironic response to Morgan’s fit of temper makes me suspect that this crew may be infecting Jessica with the hacker nature.
No, I’m not entirely or even mostly joking.
I was wondering whether anyone here as read Paul Graham’s take on the issue (warning: it’s from 2009):
http://paulgraham.com/apple.htm
@Frederico: Yes, and I believe I used it in a discussion elsewhere within the smartphone wars saga on the relative merits of the App store a long, long time ago.
Apparently even Paul Graham, an admitted Apple fanboy, gets it. (To be fair, calling someone like Paul Graham an Apple fanboy is a bit of mis-justice considering his accomplishments and the work he does incubating Web 2.0 startups)
@Patrick/@esr:
Yes, I thought it was quite funny, but I took it to be perhaps HHOS rather than purely humorous and was interested to see if she had any more detailed thoughts in that direction. I agree, esr, Jessica seems to be fitting in just fine.
Code Monkey get up get coffee
Code Monkey go to job
Code Monkey have boring meeting
With boring manager Rob
Rob say Code Monkey very dilligent
But his output stink
His code not “functional” or “elegant”
What do Code Monkey think?
Code Monkey think maybe manager want to write god damned login page himself
Code Monkey not say it out loud
Code Monkey not crazy, just proud
…
http://www.jonathancoulton.com/wiki/index.php/Code_Monkey/Lyrics
> To be fair, calling someone like Paul Graham an Apple fanboy is a bit of mis-justice considering his accomplishments and the work he does incubating Web 2.0 startups)
Given how much of the browser world is webkit-based (including Android), and the drive to HTML5 (and how badly Gecko sucks at HTML5: http://en.wikipedia.org/wiki/Comparison_of_layout_engines_(HTML5)#Elements) on “Web 2.0″… and that Apple runs the open source project that *is* webkit,…
I think one can be an Apple fan and also promote “web 2.0” without any resultant cognitive dissonance.
And? I’m using a Webkit browser right now (Chrome on Linux). I find no fault with Webkit — maybe I think Apple’s treatment of the KHTML project was a bit rude — but Webkit is a fine open source browser engine. And giving Apple all the credit for Webkit ignores quite a bit of history.
Making Webkit open source was a very shrewd move for Apple and the widespread adoption of Webkit — it’s the browser engine for Chrome/Chromium, Android’s browser, non-Windows versions of Google Earth, Epiphany, Konqueror 4.x, WebOS’s browser, Symbian’s browser, etc. — has all but ensured that Safari and OS X cannot be ignored by Web 2.0 developers. This is, in fact, more proof that open source has become more fully entrenched. rather than less.
@esr
>So do legal constraints; it is often impossible to open-source applications effectively even when the vendor wants to, because they have
>critical dependencies on closed-source system libraries. To take a contemporary example, a lot of Mac OS X applications are in this boat.
Why not simply use entirely open source alternatives? This is why I can’t really get myself on board 100% with the open source is the answer to everything. Open source has been around for a long time, and yet the best open source systems are almost all hybrids of open and closed source pieces. Android, Ubuntu linux (where best is defined in this case as most used / most consumer friendly) and Mac OS all take huge chunks of Open Source and combine it with better value closed source to produce an ultimately better product. Of course, “better” is relative to your goal, but at least for the average consumer, I would argue they’re almost all better off with the closed source components, even with the price that entails.
>By observation, vendors in the ecologies around closed-source operating systems are less rational; having formed the habit of thinking of
>secrecy as a good in itself from the operating system’s example, they often close access to source and file formats even when it damages
>their value proposition to do so.
Alternatively, the vendors of close source operating systems have a different business model from Google or Canonical or Red Hat and have value in closing their source.
>Your error is in the belief that these can be cleanly separated. Unlike you, I have spent more than 50% of my years as a consumer and
>programmer in pre-open-source conditions; thus, I know by experience that the arrows of causation between OS practice and the expectations
>of the userbase go in both directions. It is really quite interesting to encounter someone too innocent to know this.
It is possible to separate them. As a whole, consumers are (admittedly slowly) becoming aware that their OS is not their Applications and that their Data is not their Application or OS either. There is more demand among consumers not only for compatible apps, but for compatible data as well. This is in large part a credit to both the recent successes of Apple and Mozilla (and to a lesser degree the OO.org folks) but also to the successes of Android and the iPhone in mobile computing. Multiple devices each running different systems, even proprietary ones, require open data formats if they’re going to succeed.
> Ever wonder why there used to be, oh, I don’t know, maybe about 50-60 viable word processing programs, but these days you can count the
>ones that matter on the fingers of one hand?
>Think about that for a bit and get to back to me.
Because programming resources are limited, and therefore development tends to become consolidated into fewer and fewer products, especially as product gain more features and become more alike. Projects also merge, developers lose interest, lose the ability to keep up with changes or advancements and sometimes simply go out of business. Also de facto standards tend to kill off those that can’t or don’t participate in the standard. Also, were there really ever 50 or 60 word processing programs that “mattered”, not that I can recall. Maybe back in the bad old days of the Commodore 64, but it’s a lot easier to matter when your feature check list can fit on a post it note.
A consolidation from 50 choices to 3 or 4 is what I would expect in any competitive software market with barriers to switching.
>Why not simply use entirely open source alternatives?
I do exactly this. That choice is not relevant to the error in your understanding. Your belief that the consumer lock-in of closed data file formats and applications can be readily separated from the technology and culture of the underlying OS is false. Concretely, you’ve ignored the issue about legal constraints arising from use of proprietary libraries.
>Alternatively, the vendors of close source operating systems have a different business model from Google or Canonical or Red Hat and have value in closing their source.
You’re not paying attention. I said “even when it damages their value proposition”. I’m telling you that the habit of closed source leads vendors to overprotect – that is, to fail to maximize profits on their closed-source-centric business by being open where that would be a net gain. I’m not arguing that the habit of closed source is wrong, I’m saying it makes them stupid at what they do.
>It is possible to separate them.
In a hypothetical future? Possibly, though for various (primarily psychological) reasons I am doubtful. You’d have a very, very hard time making that case about actual practice in the past.
I should mention that I really do think that in the long run the OS won’t, and shouldn’t matter. A focus on applications and data should be where the real push is. If open data formats and open source programs become more successful, an OS vendor trying to break these things would be shooting themselves in the foot.
In the same way of if it couldn’t run MS Office, a business wouldn’t buy it, mainstream success of open source applications and data means the OS vendor needs to make sure their system is open enough for the applications to run. I imagine that if you tried to introduce a new OS that was locked down enough to say keep Firefox from running or running well on it, you would see more consumers these days reject the OS rather than Firefox, and that is a Good Thing™
> And giving Apple all the credit for Webkit ignores quite a bit of history.
Gee, all I said is that they run the project. Which is a straight-up fact.
@tmoney:
Which, of course, is why permissive licenses matter.
That’s true in a lot of cases. Especially for niche markets. I don’t have problems with companies making proprietary apps, and I don’t think Eric particularly does, either; else, the OSS definition would be different. You have to watch out for lock-in, though, and although the data is a huge part of that (and getting bigger), for awhile lock-in meant knowing the OS-specific mouse/keyboard commands.
Maybe it’s just a visceral reaction to events past, but, to me, opening the lowest layer of the OS really seems most important. If the OS is open, it’s always possible to either reverse engineer whatever is going on above, or simply intercept data in flight from above. That becomes much more difficult with a closed OS. If the OS has activation codes and allows apps to have activation codes, that can be a huge problem. Of course, the downside is that we don’t currently have netflix on linux, but I think it’s a good tradeoff, and one that will eventually disappear. Also, and this is part of the visceral reaction, opening the OS should be the least contentious process. The OS is a true commodity. Everybody needs one; they all do about the same thing; and one in particular works fine on anything from a cellphone up to the world’s fastest supercomputer. So to the extent that opening the OS is contentious, I always have to wonder what the secret agenda is…
As far as the whole open source/proprietary spectrum, some people think that GPLed software will mean the death of proprietary software. And they are either cheering that or dreading it. I think GPLed software raises the bar for proprietary software, in that competing with free is hard to do, and permissive software also raises the bar for proprietary software, but simultaneously lowers the cost of development. This is why some companies will choose GPL for their flagship products, and why some companies will choose permissive licensing for those things which are obviously, or need to be, commodities. Apple is very good at separating core competencies from things that they will let the rest of the world have for free in exchange for help (webkit, BSD). They even drive some projects to make sure that they aren’t beholden to the GPL for anything (LLVM), because GPL V4 just might hold them hostage otherwise.
Last one I swear
@esr,
You mentioned that OS X make a good example of applications relying on critical closed source libraries. I’m guessing in particular you’re thinking of the Application Framework and the ties into the GUI yes? This is actually in my mind, a significant problem. As near as I can find, there are no high quality open source options for developing GUI applications for multiple systems. Swing development in java while improved is still painful at best, and while Python + Tk or wxWindows is a bit better, it’s still a far cry from the APIs that make developing for OS X or even Windows easier for new developers.
And while I understand that GUI programming is in general difficult, you don’t seem the type of grognard who insists that everyone start their programming by walking up hill in the snow both ways, so I’m sure you see the benefits in the systems that Apple and Microsoft provide for their developers to create apps quickly.
Are there that you’re aware of any such open source initiatives to make a quality, easily accessible GUI API, or even a sort of bridge API (a la Swing and it’s relation to OS native widgets)? I ask this out of all sincerity as I’m truly interested.
>You mentioned that OS X make a good example of applications relying on critical closed source libraries. I’m guessing in particular you’re thinking of the Application Framework and the ties into the GUI yes?
Yes. And this is exactly the sort of thing I mean when I say that consumer-visible openness cannot be separated from the corresponding property of the OS.
>you don’t seem the type of grognard who insists that everyone start their programming by walking up hill in the snow both ways,
Indeed, I am most certainly not. I actually derive considerable enjoyment from using my unassailable old-fart status to bust on other old farts who propagate that kind of thinking. There are enough real problems waiting to be solved that putting our novices through artificial crap-endurance tests would be a waste and a crime.
>Are there that you’re aware of any such open source initiatives to make a quality, easily accessible GUI API, or even a sort of bridge API (a la Swing and it’s relation to OS native widgets)? I ask this out of all sincerity as I’m truly interested.
For MacOS? No…and one reason it’s not happening is that it would be too easy for Apple to yank the rug out from under such an effort at any time through various technological or legal means. This is yet another illustration of the point I’ve been trying to get through to you; “openness” that doesn’t go all the way down to the OS level is a dangerous trap.
@tmoney:
Heh. Typed my previous reply before I saw this. I obviously agree, but these things are insidious. Look at the restrictions iPhone users put up with. And now, perhaps Apple is doing the same thing with the mac.
Ok, so I lied, can we have an edit feature yet eric? :)
@Patrick
>So to the extent that opening the OS is contentious, I always have to wonder what the secret agenda is…
Aside from old holdover attitudes (that habit Eric was talking about), I imagine not much … of course what an OS is is changing too. The entire base of OS X is open source, but I doubt anyone would argue that OS X itself is open source, and that’s because the other components (which make up the OS experience but aren’t necessarily part of the OS) are not.
Similarly with Android, although in this case we do call it Open Source, probably because much less is closed, but is a phone that runs android but can’t use the Application Store really an Android phone anymore than an OS X (Darwin) system that can’t run Aqua is OS X?
I’ll give you 3 for Python:
1. Python + PyGTK + gtkbuilder + glade Watch the videos. No, really. Watch the videos.
2. Python + Qt4 + Qt4 Designer
3. Python + Boa Constructor + wxPython
I could give you more, but that should keep you busy for a while. All of these are cross platform, but I’m not sure if wxPython works on Mac OS X yet. The most portable is probably Qt4.
I find python and wxWindows (wxPython) GUI development to be vastly less annoying than developing using Carbon on OS X.
And yes i’m sure Apple would rather you use ObjC and Cocoa… to which i’d reply i’d place working in Cocoa on ObjC at about the same level as MIDP on Java. Sure it’s possible but i’ve got better things to do with my life than work out why coding something in it feels like wading through granite.
@ esr
>Your belief that the consumer lock-in of closed data file formats and applications can be readily separated from the technology and
>culture of the underlying OS is false. Concretely, you’ve ignored the issue about legal constraints arising from use of proprietary libraries.
If you are using entirely open source alternatives to these libraries, the it’s irrelevant. I guess this is what I’m after, no one these days (for the most part) gives a damn about the fact that hardware is almost completely closed. Rarely do you find hardware that has open and available schematics and if it uses firmware, open source firmware, yet for the most part this doesn’t matter, because every OS pretty much abstracts away hardware access so regardless of whether your hard drive is running proprietary seagate firmware or GNU/Drive 0.2 firmware, as long as it provides and interfaces the right hooks, you can talk to and use your drive how you want. And any drive that doesn’t do so fins itself unused and out of business.
So where is the opensource layer that can abstract away the OS or at least the parts of it that are proprietary? Java sort of does this, but the OS bleeds through in places and the responsibility model is backwards, the java teams have to make java work with an OS. I’m talking about an environment where the OS vendor, to get access to good applications like Firefox, provides the right hooks to work with the environment.
>This is yet another illustration of the point I’ve been trying to get through to you; “openness” that doesn’t go all the way down to the
>OS level is a dangerous trap.
Yet at the same time, the reverse can be true as well. I wonder how many resources and time have been lost to forking wars.
@Morgan
>I’ll give you 3 for Python:
Thanks. I’ve only recently begun playing with Python so I hadn’t found much beyond the Tk option.
@JonB
>And yes i’m sure Apple would rather you use ObjC and Cocoa… to which i’d reply i’d place working in Cocoa on ObjC at about the same
>level as MIDP on Java. Sure it’s possible but i’ve got better things to do with my life than work out why coding something in it feels like
>wading through granite.
Eh, your milage may vary, but the little bit of ObjC / Cocoa programming I’ve done doesn’t feel any more or less painful that Java. Different? Sure, but that’s to be expected.
>If you are using entirely open source alternatives to these libraries, the it’s irrelevant.
That’s a theoretical possibility that may not actually be practical for various reasons. One is low-level APIs which are deliberately poorly documented or locked up by legal keep-off signs; this sort of thing happens a lot with (for example) graphics cards, and is the reason 3D acceleration is often unavailable in open-source drivers. It also used to be a problem with wireless cards, and has a strong tendency to recur in new classes of hardware as vendors seek competitive advantage through lockdown (a tactic that remains popular despite being extremely destructive of market growth and profits).
Unless the operating system is open source, there is little or no check on its tendency to grow opaque blobs that are in practice roadblocks to open-source development, with ripple effects that reach all the way up to applications and the user experience. Even if large parts of the OS are open source, if the vendor does not have an open-source-centric culture (Google and Apple are today’s defining contrast) noxious things will grow in the parts that aren’t.
Morgan,
There, ftfy.
tmoney,
No. You always want to use native GUI libraries.
Well, there are things like Qt, which requires the use of (grumble) C++ and a special precompiler, but is not terrible to use and does an acceptable job of abstracting away the platform details for POSIX and Windows, anyway.
Mac OS X is an entirely different beast. It doesn’t matter how convincing you think you are. The Apple hivemind can sniff out an outsider. If your OK button is just one pixel farther than the Jobs-approved distance from the cancel button, they will one-star you so hard on VersionTracker. “OUT, OUT, ACCURSED HEATHEN. THIS IS NOT A REAL MAC APPLICATION. HOW DARE YOU PROFANE OUR SACRED GUI WITH SUCH HORRIBLE DESIGN.”
Windows users, having endured decades of horrible bespoke applications cobbled together in VB, and immediate bending or breaking of the UI guidelines by every vendor including, not least of all, Microsoft itself, are much more tolerant of such things. As for (non-Mac) Unix… “mechanism, not policy” basically means anything goes, no matter how ugly. It’s 2010 and it still shows.
So yes, if you’re targeting Mac you always want to use the native Cocoa framework. It is literally the easiest GUI framework on the planet, and the consequences for not doing so are too great if you want to win the loyalty (and business) of the faithful.
It’s worth noting that the Cocotron on Win32 and GNUstep on Linux provide pretty good Cocoa-workalikes that enable cross-platform development while still targeting Cocoa natively.
Er?
Have you been consing up GUI widgets manually instead of using the Interface Builder or something?
Building Cocoa apps is the easiest thing in the world. Historically Apple has been user-focused and Microsoft more developer-focused, which is a factor (though not the factor) in why Microsoft won. With Cocoa, Apple lowered the barrier of entry for development to such an extent that kids are writing iPhone apps.
@ esr
>this sort of thing happens a lot with (for example) graphics cards, and is the reason 3D acceleration is often unavailable
>in open-source drivers. It also used to be a problem with wireless cards, and has a strong tendency to recur in new
>classes of hardware as vendors seek competitive advantage through lockdown (a tactic that remains popular despite
>being extremely destructive of market growth and profits).
Which leads me to another question. Where is the “open source” hardware? I realize that hardware is more capital expensive to produce than software, but at the same time, Chinese factory time is cheap, where is the Canonical or Red Hat of hardware? It seems in some ways, the OS is ultimately, a middle man, and that if Open down to the OS is important, it should be equally important to be Open down to the hardware level that the OS runs on.
BTW, I hope I’m not coming off too brash in these discussions. I do enjoy them and find them enlightening and interesting, especially given our differences in both experience and views on the computer experience as it relates to the user.
@ Jeff
>So yes, if you’re targeting Mac you always want to use the native Cocoa framework. It is literally the easiest
>GUI framework on the planet, and the consequences for not doing so are too great if you want to win the
>loyalty (and business) of the faithful.
In some ways, this is a good thing for the user. By having such a complete framework and having users and developers alike demanding its use, there are a bunch of things that applications “get for free” that not only makes the development process easier, but also make the user experience much more consistent. Just a simple example that frustrates me to no end on my linux work station. Because Ubuntu uses XWindows, it has the old multiple clipboards / copy-paste channels. Without a clipboard manager, behavior between apps is inconsistent at best. Even with a clipboard manger however, Firefox as an example still has inconsistent copy-paste behavior compared to other applications. You never have such a problem on macs or windows, in part because the application frameworks handle all the copy-paste stuff and decisions for you.
I wonder why there isn’t (at least that I’m aware of) an open source (even if just for Linux) analog to the Cocoa Framework. Is there such a beast and I’m just not aware of it; or is there something I’m missing that makes its development not possible? Or is this an example of where forking wars would make the development of such a product unlikely to succeed?
It’s coming. The home fabrication technology isn’t quite there, but give it about another 5-10 years. Possibly less.
I don’t know. The Maketbot and it’s ilk are interesting, but chip fabrication is a bit different from Lego blocks isn’t it? I mean, I’m sure the tech could be made (never bet against progress), but I doubt these businesses invest in clean rooms because they like the white suits, and well, have you looked at the inside of a consumer’s computer these days? Clean they ain’t.
Or is this something like static grounding where 95% of the time you’ll be ok, and therefore this will actually work for the hobby hardware maker?
Or are you talking a level up, where you have a sort of blank general purpose chip that you flash (and freeze) into a specific purpose chip?
tmoney,
Indeed, and this is why X11 needs to eat flaming death. Ubuntu has already made moves in this direction.
tmoney:
There exists some OpenSource hardware stuff, such as the OpenCores project. This is a helpful way to reduce the development costs for those who are already in the hardware business (except that no one wants to rely on it).
The problem that you run into with hardware is that it is really expensive to do anything interesting, and incrementalism doesn’t help you like it does in software. I talked with some of our local hardware folks and if you wanted to do a one-off custom motherboard, you’re looking at a cost of about $10k, give or take. Just checking online for some of the places that do the custom PCB work, when you start looking at an 8″x8″ PCB with 14 layers, it is >1.5k, and that’s without knowing much of the quality for what you get.
A Xlinix PCI-Express development board is $2200.
Anything that you want to do is going to be cheaper to do out of FPGAs than with ASICs (semi-custom hardware you are talking about). If you want to make your own embedded board, $200 will get you something with ethernet and USB connectivity. Have at.
Finally, hardware is really, really hard to do. If you want to be taken seriously, you need to find a niche that someone else isn’t in. All of the easy stuff is uninteresting – maybe you can make an open network card, for example. But I can buy one which I know is too stupid to have any back doors for $7. You want to do graphics? You’re looking at high-end FPGAs (at least), and then you’re competing with a lot of skilled engineers at major companies. And on top of that you still have to write drivers, and keep the two in sync. A *huge* headache, without a lot of return to show for it. So, assuming that you manage to make your device, whatever it is, and debug it (not just with your setup, with with dozens of different other systems) – then what? If you convert that to an ASIC, all you’re able to sell to people at that point is another baked-in solution, but this time you can see the system layout that’s supposed to be in the chip. Nice, but not
really helpful to anybody.
The Arduino project is neat, but that isn’t really doing a lot of their own hardware, and is primarily directed towards learning or fun projects, not, say, writing a automobile control system.
Unless or until Wayland supports both NVidia and ATI graphics cards, we won’t see Wayland replacing X on Ubuntu any time real soon, no matter what noise Shuttleworth is making in that direction.
The real question is ‘When will the Symbian marketshare run off the inevitable cliff?’ and ‘Whither RIM?’.
Symbian has no future. When it does run off a cliff, it’s almost assuredly going to see most of its share go to Android since the iPhone won’t meet the pricepoint necessary (note I also expect Android to eat most of the non-smartphone market with low-end/obsolete hardware).
RIM is a bigger question. They’ve blown their last few releases and it’s clear they’re in a bind. However they’re also clearly aware of this and unlike Nokia they’re making some serious steps towards resolving their issues. But that fix is 12-18 months out right now and they need to stay alive that long. If RIM goes down, I expect iOS will pick up the majority of their share due to preferences in Corporate IT policy (single-vendor sourcing most notably) with WP7 picking up much of the remainder via the Dell phones (for much the same reasons). Android can’t offer a single-sourced package for backend and handset and thus is starting well back from iOS and WP7 here.
Comparing Apple’s shipments against the sum tota of Android shipments is an interesting lark, but not terribly relevant. Apple isn’t competing against most of the Andoid handsets, but rather just a few (the Samsung Galaxy’s, Motorola Droid’s, the Nexus and whatever HTC’s top-end phones are) and I doubt that they are selling as well as the iPhone given the aggregate Android numbers.
Another thing to consider about Android’s viability in the high end of the market is the coming death of media players. Apple utterly dominates this space via iTMS and its lock-in to the iPod. And guess what the only Phone with iTMS support is? Android simply cannot even begin to compete with iOS as a multimedia device because it’s seriously lacking in content sources for Music, TV and Movies. Here’s the one spot other than corporate use where WP7 has some serious advantages. MS’s XBox team understands content in a way Google and the OHA simply don’t, and the XBox team is in charge of the WP7’s content sources and they’ve got a solid library of content already from Vuze.
Frankly, I used to think that the smartphone market was going to simply eat the entire phone market and leave it at that, but I’m starting to thing that we are going to see a segmented market, with smartphone’s taking over the phone/smartphone/GPS/PDA portion and ‘Media’ phones covering the smartphone/Media Player/GPS combo. I see Android dominating the first option and a combination of iOS and Windows Phone dominating the second, likely with an entry from Sony Ericsson with PSN integration competing here as well. It pretty much comes down to a split between computing functions and media consumption functions. Android and RIM are the players for the first and iOS and WP7 are the players for the second. Unless something happens that provides a significant content portal for music/TV/Film for Android which can compete with iTMS (and to a lesser extent Vuze) Android simply can’t compete here. And no Netflix isn’t the answer, it’s the default low-end for everybody.
@tmoney:
I know of at least one American company with chip fab capabilities that is slowly replacing some of their cleanrooms with technologies that don’t require them. For example, FPGAs, which have been around a long time, have gotten to the point that for smaller runs, they compete or even beat traditional ASICs on cost. There are also FPGA-hybrid and non-FPGA technologies that offer similar capabilities, but with capabilities and performance that compete with traditional ASIC designs.
@Morgan> I know of at least one American company with chip fab capabilities that is slowly replacing some of their cleanrooms with technologies that don’t require them. For example, FPGAs, which have been around a long time, have gotten to the point that for smaller runs, they compete or even beat traditional ASICs on cost.
This is because for smaller runs, you can’t “pay back” the additional NRE required to design ASICs, not because FPGAs cost less. FPGAs cost more and are normally 1/3 as fast (as well as drawing 5X or more the power of an ASIC).
@Adam,
Saying Symbian has no future is like saying Windows or Power PC have no future. Symbian is a product of Nokia, and Nokia, like Microsoft or IBM is too big to go anywhere soon.
Agreed about ‘media’ keeping iOS in the forefront.
>Agreed about ‘media’ keeping iOS in the forefront.
I’ll concur with you and Adam this far: if anything can keep iOS’s long-term share above the mid-single-digit range characteristic for Macs, the media tie-in is what will do it. This, however, presumes that Android finds no way to game effectively against that advantage, a bet I wouldn’t care to make.
@ Garrett> writing a automobile control system.
You mean, like the megaquirt / megashift?
tmoney> It seems in some ways, the OS is ultimately, a middle man, and that if Open down to the OS is important, it should be equally important to be Open down to the hardware level that the OS runs on.
There is a view that says that anything in your OS is a missing feature (bug) in your language/runtime.
Eric,
Current market share estimates for Macintosh put it slightly above 10%. That’s more than mid-single-digits and certainly a lot bigger than Linux’s numbers. Linux faces the mother of all uphill battles here, as Mac OS X has beat it at its own desktop game. Even hackers consider the Mac the workstation platform of choice.
>Even hackers consider the Mac the workstation platform of choice.
You know, you were almost sounding plausible until you said that.
@tmoney>If you are using entirely open source alternatives to these libraries, the it’s irrelevant.
@esr> That’s a theoretical possibility that may not actually be practical for various reasons. One is low-level APIs which are deliberately poorly documented or locked up by legal keep-off signs; this sort of thing happens a lot with (for example) graphics cards, and is the reason 3D acceleration is often unavailable in open-source drivers. It also used to be a problem with wireless cards, and has a strong tendency to recur in new classes of hardware as vendors seek competitive advantage through lockdown (a tactic that remains popular despite being extremely destructive of market growth and profits).
Wireless cards typically had closed-source components due to worries about excited young youth attempting to increase power or tune the card outside of regulatory limits, thereby violating federal law in the US and other geographies.
As for Open Source 3D graphics drivers, there is a whole wikipedia entry on the current state-of-play: http://en.wikipedia.org/wiki/Graphics_hardware_and_FOSS
Adam Maas,
I have a doddering old iPod. First gen — literally, the thumbwheel is a hunk of plastic mounted on an axle and it spins.
I haven’t sought to replace it for one reason: my huge collection of oggs still won’t play on an iPod.
My Android phone handles them just fine.
Android has a chance by squeezing into the niches and crevices that the big players (Apple Microsoft) ignore.
@TJL:
> There is a view that says that anything in your OS is a missing feature (bug) in your language/runtime.
It may be that the “OS” of the future is just something that virtualizes a bunch of other OSes. This is certainly happening in servers to a great extent. Just needs a bit more uptake on the desk. A couple more ugly Windows viruses ought to have everybody wanting their browser firewalled off nicely.
@ear
>This, however, presumes that Android finds no way to game effectively against that advantage, a bet I wouldn’t care to make.
DoubleTwist is a (third-party) step in that direction. [www.doubletwist.com]
Re: open source graphics.
Forget it. If you want to do 3D graphics, you know where to find Windows.
State-of-the-art OpenGL is always going to be where Direct3D was two revs ago, and the open source drivers are:
a) going to support substantially less than the full latest OGL spec due to the time delays of reverse engineering
b) possibly illegal, because talking to 3D hardware is an IP minefield (setting cleanroom aside for a moment, the patents on S3TC haven’t run out yet)
> You know, you were almost sounding plausible until you said that.
You know, you just sound like an old, cranky man when you say that.
ESR says: Zero-content insult. You are close to the ban limit.
> my huge collection of oggs still won’t play on an iPod.
Yes, they will: http://www.rockbox.org/wiki/WhyRockbox
I know about Rockbox. Too bad it only works on old ‘n’ busted iPods.
It’s barely possible that someone will be able to make a jailbreak hack that will let iOS play OGG files. This would be difficult. The iOS family is perhaps an order of magnitude more complex and more paranoid than the old iPods that you could install iPodLinux or Rockbox on. Though I really want to see a full third-party OS for an iTouch, just to play with.
Dude — iDroid. Only works on old ‘n’ busted iPod Touches though :\
>>Even hackers consider the Mac the workstation platform of choice.
>You know, you were almost sounding plausible until you said that.
That’s essentially correct. *Lots* of open source developers primarily use Macs, including the Ruby on Rails core team last I heard. Mac OS X does a great job of giving you the low-level Unix environment, combined with well-designed GUI apps and a lack of futzing to get sleep mode or multiple monitors working. For me, the Mac is a great development platform for Android and web apps.
Of course, if Apple continues on their current trajectory they’ll have Macs locked down like iPhones in 5 years or so, and the hackers will flee.
>*Lots* of open source developers primarily use Macs,
I keep hearing rumors of this, but as long as I continue having trouble finding Mac OS porters for my projects I’m not going to believe it. I have an easier time finding people running *BSD, and I know that minority is tiny. I think the supposed popularity of Macs in open-source-land is an urban myth fueled by a handful of high-profile developers and Apple marketing, and will change that evaluation only on actual evidence.
Note, read Zune where I said Vuze upthread, brain fart as I’d just fired up Vuze.
@Jeff Read: Yep, if you’re part of the tiny fraction of the market who actually uses Ogg, iOS ain’t for you. Unfortunately you aren’t even vaguely indicative of the market overall as Ogg never acquired any traction in the real world, even the atrocity known as WMA remains more popular. Niches are something Android does well. But niches inherently don’t win markets.
@esr: When Android actually has a major legal content source on the scale of iTMS, Zune Marketplace (or as I’ve recently found out Netflix) Apple may need to start worrying. However the absolute idiocy of the Major Studios/Labels in general is a strength for Apple and MS here (note, both are aware of how dumb their content providers are and Apple at least has pushed towards less DRM. Apple’s business model here is to make moving away annoying, not impossible). What’s going to happen here is a market split rather than competition inside a single market and when it comes to Media, Android’s not even on the radar yet.
@Jezus Lizzard: Symbian is an evolutionary dead end now. Think Windows 9x 7-8 years ago. Huge market share but no future. Nokia’s best solution is going to be moving to a different OS and providing emulation for Symbian apps. For various reasons I expect that to be either Android or Meego.
@esr
>I keep hearing rumors of this, but as long as I continue having trouble finding Mac OS porters for my projects I’m not going to believe it.
Without knowing specifically what you’re trying to port, is it not possible that your projects simply don’t interest the OSS mac developers? Or possibly that your project compiles just fine on a mac and no real porting is necessary?
>Without knowing specifically what you’re trying to port, is it not possible that your projects simply don’t interest the OSS mac developers? Or possibly that your project compiles just fine on a mac and no real porting is necessary?
The former possibility seems excluded by the fact that I maintain a lot of general utility software that is boring but necessary for developers on any Unix-like system.
The latter seems excluded by the frequency and scale of patches I get on the rare occasions that one of my projects does in fact find a Mac OS porter.
In my experience, you always reach a point where the interface builder can’t cope anymore and you have to build/tweak the interface by hand. But thats a by-the-by and kind of tangential to my point.
My point was “I’ve worked with carbon and i found it to be a dog.” The follow on was because I knew some bright spark was going to point out that the “apple approved” method is ObjC and Cocoa and i have no desire to work with an obtuse language that is only relevant inside the walled garden.
Oh and by “the interface builder” i mean any interface builder. Whether it be the one in visual basic, visual studio, netbeans, intellij or xcode.
>I keep hearing rumors of this,
These guys work almost entirely on the Mac: http://llvm.org/
> as long as I continue having trouble finding Mac OS porters for my projects I’m not going to believe it.
As it turns out, the OSS developers on the Mac have better things to do than work on your code. They’re writing their own.
>is it not possible that your projects simply don’t interest the OSS mac developers?
Got it on the first guess.
Yeah….stupid blog host….I’ll just leave some oblique derisory remarks accompanied by an ethereal sneer. That’ll show him. He thinks his profitable speaking career and published status means something…HA…what a mustachioed loser…fuck ESR….when history looks back they’ll see the devastating retorts of some guy that posting something on some blog sometime, and their laughter will obscure all memory of that guy that anchored the OSS movement and clearly articulated its virtues on netwide TV.
Fuck you ESR….the future belongs to the unachievers!
>profitable speaking career
Would that it had been profitable. I’ve never charged more than expenses, because it’s not about making me rich, it’s about changing the world for the better. Occasionally they’ll think of stumping up an honorarium, but most of my material gains from speaking gigs have been mugs and T-shirts. I may have the world’s largest collection of Linux user group commemorative mugs.
If you follow that sweeping generalisation through to its obvious conclusion and you come to a very retarded place.
You’re suggesting that Mac OSS developers believe they are better off writing their own version of [hypothetical project] rather than pitching in on [hypothetical project] like every other platform? Really?
Thats a weird variant of NIH syndrome.
You’re suggesting that Mac OSS developers believe they are better off writing their own version of [hypothetical project] rather than pitching in on [hypothetical project] like every other platform? Really?
No, I’m saying right out, that working on LLVM is more important than ESR’s utilities. When I say they’re writing their own code, I don’t mean that they’re doing their own versions of things that already exist.
>No, I’m saying right out, that working on LLVM is more important than ESR’s utilities.
I’m inclined to agree with that, at least in potential. It’s high time GCC got some real competition, and it looks to me like the LLVM people are doing a bang-up job at it – sound architecture, well-chosen goals, and (to the limited extent I can judge from the outside) high-quality implementation. I wish them every success.
Still, I maintain (or have maintained in the past) a lot of unglamorous but essential plumbing – things that Mac developers need as much as anyone running a Linux or *BSD. Fetchmail is a good past example, gpsd a good present one. I do not see a level of Mac-user participation in those projects that I think is compatible with the belief that lots of open-source programmers are using Macs.
That absence can’t be explained by snorting “Hmmmph. Mac people have their own projects to work on!” because that’s also true of the Linux people and the BSD people. Linux and BSD people, working on their own important software or not, do show up; Mac OS X people mostly don’t. That there are a lot fewer of them seems like the least hypothesis to explain this.
I know a few hackers who run Linux on the Mac, because it’s a better X86 laptop. So don’t assume that just because a hacker has a Mac, he’s a Mac hacker.
…I may have the world’s largest collection of Linux user group commemorative mugs…
That is simultaneously tragic and awesome…yet how much treasure have you amassed in the simple accumulation of wonderful experiences aruond the globe? I’m betting you’re a rich man in that regard. ;)
The ‘haterboys’ that piss their inconsequential effluent in your general direction are mostly soiling themselves.
I think you’ve earned your slot in the big boy league.
No…I’m not a brown noser….that’s a straight-up slap on the back for a job well done.
>That is simultaneously tragic and awesome…yet how much treasure have you amassed in the simple accumulation of wonderful experiences aruond the globe? I’m betting you’re a rich man in that regard. ;)
Indeed I am. I don’t regret the course I chose. I can look around me and see the results every day. I changed the world. I changed the world. Not many people even get the chance to do that, and still fewer ever collect as much positive certainty as I have about the results. That is treasure better than gold.
>The ‘haterboys’ that piss their inconsequential effluent in your general direction are mostly soiling themselves.
Ssssh! They’re much more entertaining when they don’t know that.
>No…I’m not a brown noser….that’s a straight-up slap on the back for a job well done.
One of the essential psychological survival skills for anyone as famous as I’ve been is the ability to tell that difference. So, yes, I got it.
@esr
I dont’ mean to imply that they have their own more important things to work on, merely that the project in question doesn’t interest them, or doesn’t meet the demands their customers are looking for.
And honest question because I’m not very familiar with either, how are Fetchmail and gpsd good examples of “essential plumbing … that mac developers need”?
>And honest question because I’m not very familiar with either, how are Fetchmail and gpsd good examples of “essential plumbing … that mac developers need”?
Go read up on them. I think it will become clear pretty quickly.
SomeGuy wants us to think Mac developers don’t need gpsd. Joke’s on him, because that’s about the only one of my infrastructure projects for which I actually do have a long-term reliable Mac port guy. Whatever this Core Location Framework thing is, it’s clearly not meeting his needs.
…I know a few hackers who run Linux on the Mac, because it’s a better X86 laptop. So don’t assume that just because a hacker has a Mac, he’s a Mac hacker….
Shhhh…don’t tell anyone….but I’ve been thinking much the same thing for a while now….I just resent paying the anal-rapage price for the kind of Mac that would suffice.
This is why Mac and iOS developers don’t care about gpsd:
http://developer.apple.com/library/ios/#documentation/CoreLocation/Reference/CoreLocation_Framework/_index.html
@Some Guy
Which works for iOS and related devices. gpsd seems to be a general purpose library for any GPS device on USB.
Still not quite seeing it (also it’s getting late here, so my patience with documentation is waning). I can see that gpsd would be extremely useful to *some* mac developers, but how many mac developers do you think are making location aware applications for OS X? For an iPhone, sure, but as Some Guy points out, there’s a Framework for that (obviously not OSS, but then again none of the iPhone Frameworks are), but for the desktop? Like I said, I think this is a matter of interest. I would imagine that if you took a survey of the people that used a GPS attached to their computer for activities, that you would find BSD / Unix folks by far outstrip the windows and mac users, not only in relative numbers to their user base, but in absolute numbers as well.
>but how many mac developers do you think are making location aware applications for OS X?
The ones that have laptops? It’s the same set of application cases as Linux laptops, really.
Ok sure i can get behind that. LLVM is fantastic.
If thats what you meant then maybe i misinterpreted you. When you said “the OSS Developers on the Mac” in response to Eric not believing that “*Lots* of open source developers primarily use Macs” due to lack of interest in Mac support, I assumed you meant OSS Developers that are on Mac in general, certainly not the subset of mac hackers working on LLVM.
@ear
> Indeed I am. I don’t regret the course I chose. I can look around me and see the results every day. I changed the world. I changed the world. Not many people even get the chance to do that, and still fewer ever collect as much positive certainty as I have about the results. That is treasure better than gold.
For mine, I think that’s absolutely right. I disagree with a good proportion of your politics and a reasonable fraction of your writing about social issues and I think you are dangerously wrong about climate science … but I remain in awe of your achievement. Even as a mere user I am deeply grateful for it.
@Adam Maas:
“Unless something happens that provides a significant content portal for music/TV/Film for Android which can compete with iTMS (and to a lesser extent Vuze) Android simply can’t compete here. And no Netflix isn’t the answer, it’s the default low-end for everybody.”
I would bet on Amazon stepping into that Android niche. They already sell media, and media to Linux.
@esr
>The ones that have laptops? It’s the same set of application cases as Linux laptops, really.
Of course it’s the same set of applications, my question is, if you’re wanting to spend your time writing or porting an application for mac users, it’s likely going to be an application you think that users will use. And ultimately, if you polled all the mac users out there, you’d probably find less of them use GPS attached programs on their laptop than of UNIX users. If I had to guess at the cause of this, it’s the types of users. Apple has a lot of consumer users and artist / creative users, neither of which have a particular need for GPS attached computing, and the needs they do have likely are these days complimented or satisfied by a smart phone (android or iOS). On the other hand, UNIX has a lot of academic and research users, where there are far more chances for there to be a use or need for GPS attached computing.
Just because both computers can run the same types of applications, doesn’t mean that the demand for those applications is the same, and therefore doesn’t mean the developer interest in working on those applications is the same.
>Apple has a lot of consumer users and artist / creative users, neither of which have a particular need for GPS attached computing
That changes the subject. For purposes of scoping the percentage of open-source developers who use Mac OS X machines, the people you’re talking about are completely irrelevant.
> SomeGuy wants us to think Mac developers don’t need gpsd. Joke’s on him, because that’s about the only one of my infrastructure projects for which I actually do have a long-term reliable Mac port guy. Whatever this Core Location Framework thing is, it’s clearly not meeting his needs.
Point-in-fact, on MacOS X, CoreLocation works *without* a GPS. (Try *that* with gpsd!)
(It will use a GPS if one is detected, too.)
http://cocoawithlove.com/2009/09/whereismymac-snow-leopard-corelocation.html
http://developer.apple.com/library/mac/#documentation/CoreLocation/Reference/CoreLocation_Framework/_index.html
> Ssssh! They’re much more entertaining when they don’t know that.
If Eric’s ego wasn’t so strong, he would have caved under the onslaught of the haters by now.
Fortunately, he has yet to meet an ego as strong as his in this field of battle. The haters remain out-gunned.
>If Eric’s ego wasn’t so strong, he would have caved under the onslaught of the haters by now.
I think you may be overestimating me, and I’m certain you’re overestimating them.
:-)
I’ll just leave this here:
http://sixarm.com/about/phone-review-nexus-s-ugly-buggy-slow.html
Eh. A lot of reviewers liked it.
On the other subject du jour — fetchmail has, of course, also been ported to the Mac. Note that using that as an example is poor anyhow, since Eric stopped maintaining it years ago and presumably we’re more interested in the current situation than the situation back then.
While it wouldn’t surprise me if that case does exist somewhere, to my knowledge the general case is slightly different. The best open source developers spend time writing and porting applications that THEY use. This use of “they” also extends to things like business cases (i.e. writing and porting things that will make their jobs easier).
Carbon is a dog. It’s the legacy Mac OS API, more or less. It’s the reason why the Mac application space once had a high barrier to entry, which plummeted to the floor with the proliferation of Mac OS X and Cocoa.
Seriously, try Cocoa. It’s completely different and awesome. This isn’t fanboyism; the open source GNUstep implementations of the relevant technologies do a competent enough job of bringing the awesomeness to Linux (why the Linux community persists on being stuck in the GNOME ghetto, I’ll never know).
As for Objective-C, gcc compiles it just fine without the Mac libraries. I use it for a project of my own because the alternative — C++ — is even more obtuse. Look at the hoops the Qt devs jumped through in order to bring to C++ Objective-C’s salient feature: dynamic message-passed method dispatch.
esr Says:
>I think you may be overestimating me, and I’m certain you’re overestimating them.
People don’t hate me as much as they hate Eric, but they hate me plenty enough. My experience, which I suspect parallels yours, is that I feed off the hate. It doesn’t make me want to back off my position, it makes me want to advocate it more strongly. Hatred is an interesting thing. A positive response to hatred puts more daylight between you and the hater, and makes your position more attractive to sensible rational people. You’ll never convince the hater, but the undecided middle generally prefers hate free arguments.
However, it does get wearing sometimes. That is when Eric’s major flaw becomes a problem. Jack Daniels is a great friend at such times.
@ esr
>That changes the subject. For purposes of scoping the percentage of open-source developers who use Mac OS X
>machines, the people you’re talking about are completely irrelevant.
So too is the number of people willing to port your specific software. You believe the a lack of OSS developers using OS X accounts for the lack of people willing to port your specific software compared to a smaller base (BSD). I believe that it’s possible that a lack of interest is your specific software accounts for this discrepancy. Until one of us locates either a survey of OSS developers and their OS choice or a survey of what percentage of developers on each platform care about gpsd then we’re both just guessing based on anecdotes and assumptions.
That isn’t to say that your overall assumption that there are less OSS mac developers on the whole is wrong, in fact I would put odds on it being right (especially if merely developing and releasing your source is not enough to qualify as an OSS developer) but that the process by which you have reached that conclusion is flawed
@ JonB
That may well be the case. My question is how many OSS developers using macs use gpsd? If they don’t use it, then it’s a simple lack of interest, not a lack of developers. Eric seems to be arguing that because the number of use cases is the same across the platforms that the (relative) number of people and developers using it should be the same, and that’s faulty reasoning.
>So too is the number of people willing to port your specific software.
I think not. I’m sampling from the same population we’re trying to estimate; you’re arguing that the size of a different population is relevant.
The only assumption my conjecture requires is that my software is equivalently interesting to open-source developers across Mac, Linux, and *BSD. That assumption may be false, but arguing against it by talking about the sizes of the platform populations of people who are not open-source developers is fallacious.
@ Jeff
>Objective-C’s salient feature: dynamic message-passed method dispatch.
I have to say, the syntax of ObjC’s message passing has got to be one of my favorite non-feature-feature’s of developing mac apps. When reading through code it is so much nicer and easier to read:
[object attachPerson: personObject withName:”myPerson” displayColor:”blue”]
rather than the most common forms in dot-syntax languages of:
object.attachPerson(personObject, “myPerson”, “blue”)
or
object.attachPersonWithNameAndDisplayColor(personObject, “myPerson”, “blue”
Carbon is a dog
You won’t find many Mac developers who would dispute that anymore. It needed to be done back in ’98, but any carbon code written later than about six or seven years ago is good money after bad.
SomeGuy wants us to think Mac developers don’t need gpsd,
I know that I don’t. Maybe some other Mac developer does, but I have yet to meet him.
Regarding OSS developers on Mac OS X, this just came by on twitter:
http://www.sharms.org/blog/2010/12/on-why-open-source-developers-run-mac-os-x/
@ esr
Unless we’re assuming that all or most OSS developers only develop things that interest them and never what people who aren’t developers might use or what they previously used before they became a developer, I don’t see why the users of the platform you’re developing for would not matter as far as developer interest in a particular project is concerned.
And if we are assuming that, perhaps that is a reason why OSS advocates find it so hard to sell open source to the masses.
>I don’t see why the users of the platform you’re developing for would not matter as far as developer interest in a particular project is concerned.
Dude, they’re users. Not developers. Users aren’t going to show up wanting to tinker with the plumbing joints. You’re not thinking about the kind of stuff I write – not games, usually not GUIs, mostly development tools and service libraries and DSLs. What user is ever going to care about reposurgeon and deheader? OTOH, these are tools that ought to have uniform cross-platform appeal to developers because they’re about things that all developers do.
That’s incorrect. It also requires that developers on each platform be equally willing to ask you for help.
> Not developers. Users aren’t going to show up wanting to tinker with the plumbing joints. You’re not thinking about the kind of stuff I write – not games, usually not GUIs, mostly development tools and service libraries and DSLs. What user is ever going to care about reposurgeon and deheader?
Which is why they don’t care about gpsd (much) on MacOS X.
> OTOH, these are tools that ought to have uniform cross-platform appeal to developers because they’re about things that all developers do.
Not on MacOS X / Xcode
I would never assume all (i said as much in my previous comment), but certainly i’d assume that most OSS developers only develop things that are interesting or useful to themselves (or people they know). To me this is obvious based on the assumption that unless it’s OSS to make your job easier(which certainly implies interesting or useful to people you know), it’s probably OSS in your spare time. If i’m coding something in my spare time i’m not going to work on something i find uninteresting or have no use for.
And if we are assuming that, perhaps that is a reason why OSS advocates find it so hard to sell open source to the masses.
It’s certainly a plausible reason why GUI is an achilles heel of OSS in general.
>So before I bust a gut laughing at the hating on GPSD… can I just point out that ESR has already stated that gpsd is one of the few of his projects that has MacOS hackers porting it.
Haters. Their imperviousness to facts is such a large part of their entertainment value.
@esr
>Dude, they’re users. Not developers. Users aren’t going to show up wanting to tinker with the plumbing joints. You’re not thinking about the kind of
>stuff I write – not games, usually not GUIs, mostly development tools and service libraries and DSLs. What user is ever going to care about
>reposurgeon and deheader? OTOH, these are tools that ought to have uniform cross-platform appeal to developers because they’re about things
>that all developers do.
And gpsd is a tool who’s use cases mostly fall into academic and user software, not developer software. This is why I specifically asked you why gpsd was your good example of a tool developers needed but you couldn’t get many mac porters for. If you were thinking about reposurgeon and deheader, we were having different conversations, which is why we’re not getting anywhere. I agree that those two tools have more developer appeal.
However, reposurgeon only appeals to developers using a specific version control systems who also have a desire to do the tasks reposurgeon does. It’s also a python program, and at least at a cursory glance appears to run under OS X as is. The task domain goes back to my argument about interest in the project and the python / compatibility part are both strikes against OS X specific patches or developers stepping forward.
Deheader is language specific is it not? That means you’re excluding any developer that doesn’t use that language. Again, not equal appeal to all developers on all platforms.
@ JonB
>If i’m coding something in my spare time i’m not going to work on something i find uninteresting or have no use for.
Which is exactly what I’m arguing.
>And gpsd is a tool who’s use cases mostly fall into academic and user software, not developer software.
It’s not that simple. I get a lot of interest in gpsd from developers writing location-aware applications. Most of my bug reports are from devs who’ve noticed a problem because they’re working on client software.
>However, reposurgeon only appeals to developers using a specific version control systems who also have a desire to do the tasks reposurgeon does
The second qualification is trivial. Of course “who also have a desire to do the tasks [it] does”; how could it be otherwise for any program? And the first is near-trivial considering the range of VCSes it supports. You’re verbally thrashing around without actually saying much; are you actually thinking?
>Deheader is language specific is it not?
Yes. Next I suppose you’re going to tell me nobody writes for Mac in C, C++, or Objective-C?
deheader runs without porting on the Mac, seeing as how it’s written in Python. I was a bit surprised, since the python that ships with the Mac is a minor version behind the release branch and a major version behind the dev branch, but it doesn’t need porting — thanks for minimizing dependencies! In any case, we probably don’t need to argue about deheader. At this point I’m really not sure what Eric’s examples are. His experience maintaining fetchmail is years out of date. gpsd has someone porting it. reposurgeon runs as is. deheader runs as is.
> gpsd has someone porting it.
I get the feeling that many people use GPSD not because they want multiplexed access to a GPS (or APRS) unit, bur rather because they don’t
want to have to deal with interacting with the hardware themselves. This, mixed-in with GPSD dating from a time when GPS units were actually really bad.
Moreover, marine units that talk NMEA-0183 are going away in-favor of NMEA-2000, which is CAN-based, and consumer GPS is being cannibalized by cell phones with integrated GPS. While gpsd remains unavailable on Android, there is a different protocol for a similar use that does, (and it uses GPSD’s port) http://code.google.com/p/gpstether/
Meanwhile, GPSD is changing (http://gpsd.berlios.de/protocol-transition.html) in ways that break existing applications.
Who can tell how much of GPSD’s userbase (or developer base) will follow in the midst of these swirl of change?
This is not an attack, but I think it does illustrate that GPSD faces some challenge in the near future.
>This, mixed-in with GPSD dating from a time when GPS units were actually really bad.
That time has not ended. The presence of multiple incompatible vendor protocols, in itself, assures that.
@esr
>It’s not that simple. I get a lot of interest in gpsd from developers writing location-aware applications. Most of my
>bug reports are from devs who’ve noticed a problem because they’re working on client software.
Which does absolutely nothing to disprove my claim that gpsd is applicable primarily to user and academic applications, and therefore developer interest from any given platform is a function of the users of that platform.
>Yes. Next I suppose you’re going to tell me nobody writes for Mac in C, C++, or Objective-C?
Not at all, but I would bet more of the mac also program in languages like Java / Python to keep the level of porting they might need to do to a minimum than those that program for unix systems, thus reducing the interest in the deheader application. And if deheader also runs as is, per Bryant, then of course there’s no interest in porting as there is no need.
Like Bryant said, you’ve chosen some rather poor examples to make your point, which was why I asked for you to clarify why you felt the applications you chose were good representative samples. Instead you said to go read up on the apps and draw our conclusions from there. Well, the conclusions drawn suggest that none of the provided apps are good ways to measure porting action and extrapolate developer availability from. If you still think so, you’re going to need to come up with a more concrete example. And again, I’m not suggesting that the number of OSS developers on mac is low, but I am stating how you have reached that conclusion is faulty.
>Like Bryant said, you’ve chosen some rather poor examples to make your point, which was why I asked for you to clarify why you felt the applications you chose were good representative samples.
This is ridiculous. You both seem determined to ignore the nature of the population I’m actually sampling in order to reach preconceived conclusions about different ones. Given that, I doubt any attempts I made to ‘clarify’ matters would be more than a waste of my time.
> This is ridiculous. You both seem determined to ignore the nature of the population I’m actually sampling in order to reach preconceived conclusions about different ones. Given that, I doubt any attempts I made to ‘clarify’ matters would be more than a waste of my time.
I don’t have a dog in this fight, but I suspect the most persuasive arguments (on both sides) would involve equations and numbers, even back of the envelope equations driven by Sophicated Wild Ass Guess numbers. And without such actual back of the envelope calculations people are just going to go with their gut and not be persuaded. For example, a certain percentage of artistic types are going to be developers. And a certain percentage of those developers are going to be attracted to open source. They are more likely to work on projects with artistic goals, and not gpsd and fetchmail. However, some of them will be interested in both, since artistic types do use GPS and email. But without equations and numbers we are just going with our hunches and human beings are not good at statistical hunches.
I see no reason to trust ers’s gut over tmoney’s gut. It isn’t like the smartphone stuff where esr is looking at actual numbers and doing actual math. There I do have reason to trust esr!
Yours,
Tom
I do. Just another one of those SWAG’s, but I’ll bet that esr knows a lot more hackers on a personal, first-name basis than tmoney simply for the fact that esr has been involved in the hacker community much, much longer than tmoney.
I’ll bet not many people commenting on this blog can say definitively that they were there when RMS first started speculating about what would become the FSF and the GPL.
Call it an appeal to authority all you want, but experience counts.
>I’ll bet not many people commenting on this blog can say definitively that they were there when RMS first started speculating about what would become the FSF and the GPL.
To be scrupulously accurate, I’m not sure I was around when RMS first started speculating. From the way he talked, I think the idea had been forming for a few months prior – but I was probably among the first two dozen or so people he discussed it with. And I was the person who, when RMS said he didn’t yet know what the FSF’s flagship product ought to be, said “An Emacs implementation, of course. Duh!”
A few years later (I think in 1989-1990, but am not sure) my wife became an early FSF contributor. Cathy was a few years into practicing law then, and RMS asked her to perform a legal review on a draft of what became GPLv2. It is likely that some of her wordsmithing and technical corrections are still in there.
@esr
Illustrating the importance of being around at the right time and the right place. Then making sure you will not regret for the rest of your life what you did do or left undone.
This can be extremely frustrating, like that Beatle that left just before the group had their success. It seems to me you have few such regrets.
>This can be extremely frustrating, like that Beatle that left just before the group had their success. It seems to me you have few such regrets.
Oh, I have one or two. I’ve already mentioned one here; had I known the problem was important, I could have accelerated the development of version-control systems by a good 15 years. And I not only had but broadcast the core idea of a DVCS five years before that wave hit.
Another one I’ve already mentioned here is that I had the key insight about the phenomenon of quantum decoherence and the proper resolution of the Schrodinger’s Cat paradox about eight years before that one hit the physics journals.
One I only half-regret is send-expect. Don Libes made his name by building an implementation in Tcl that could be used in a fairly large class of scripting problems. I had built something isomorphic in C years earlier and didn’t know it was important enough to make noise about. But that’s OK, Don is a nice guy and deserved his fame.
> I do. Just another one of those SWAG’s, but I’ll bet that esr knows a lot more hackers on a personal, first-name basis than tmoney simply for the fact that esr has been involved in the hacker community much, much longer than tmoney.
OK. I amend my statement. I see only one, rather good reason to trust ers’s gut over tmoney’s gut.
Yours,
Tom
Here’s a pretty good, hands-on review of the Nexus S that someone linked to from HN http://www.avc.com/a_vc/2010/12/nexus-s-review.html
Sorry for pestering you about it, ESR..
It been at least two weeks since I asked you about bitcoin.
What do you think about bitcoin?
ESR sys: There’s potential there. Still evaluating.
The send/expect paradigm predates Libes’ expect by quite a bit. It was common in scripting languages present in modem terminal software (the kind used for dialing BBSes) from as early as the late 70s or early 1980s. If I’m not mistaken, PC-Talk, Crosstalk, and Qmodem all had this.
>The send/expect paradigm predates Libes’ expect by quite a bit.
Yes, but his design (and mine) were both decoupled from serial comm (thus, usable with other sorts of programs) and much more flexible. I think we both did the same thing – looked at the piles of kluges in PC-Talk etc. and tried to isolate a clean set of domain primitives. One indication that we succeeded is that our independently-developed final designs looked very, very similar.
@kiba:
I dunno. Most people can’t keep their computers secure to start with, now you want to store and create value there?
Not to mention the possibility of outright theft by zombie networks from honest individuals:
Good luck with that and all, but the system will probably only be secure until anybody actually stores value in it…
@kiba
I find bitcoin a bit of a solution in search of a problem; or, rather, a problem that enough people think is worth solving.
>bitcoins can be sent easily through the Internet, without having to trust middlemen.
I can already easily make payments over the internet (banks, paypal, credit card); trusting middlemen doesn’t seem a problem.
>Transactions are designed to be irreversible
Aren’t most transactions pretty much like this?
>Be safe from instability caused by fractional reserve banking and central banks.
I’ve absolutely no idea why I (as a typical purchaser) need worry about this.
These spring out as serious problems:
1. No widespread adoption
2. My wallet needs to be backed up after every transaction
3. If I lose the wallet backup, I lose *all* my coins; and this is far worse than if I lose my normal wallet — in that case, I only actually lose what’s in the wallet, the vast majority of my other funds are safe elsewhere (and I have additional protections / safeguards depending on the organisations those monies are with)
I’m a bit lost as to what bitcoin offers that end-users don’t already have (and that end-users actually want), and thus see no reason end-users will take it up (other than a marginal niche of them).
@Patrick
The zombie network would rather mine blocks and process transactions rather than compromise and destroy the integrity of bitcoin. Only the government would want to do that. So botnet operators would actually secure the network instead of trying to cheat.
Beside, it only mean you can do doublespending of your own money. If you don’t have access to other people’s private key, you couldn’t spend their money. With more and more nodes coming online everyday, the network will get much more difficult to attack from a nation-state perspective.
Right. That was the cool thing. But the kids these days all take it for granted and tend to think of stuff like expect as old and crufty. But I still remember coming across expect sometime around 1990 or so and going “Wow! Cool! A real send/expect!” I also remember another implementation that didn’t use Tcl; maybe it was yours.
>I also remember another implementation that didn’t use Tcl; maybe it was yours.
Could be, but I wouldn’t be surprised to learn of more designs very similar to Libes’s and mine around that time either (mid-to-late 80s; I think I wrote mine in ’84, and IIRC his dated from 1987). The conceptual leap involved was not large.
If paypal doesn’t take care, they might find their payments over the internet thing starting to dry up. I’ve heard from 5 independant sources that resolving paper work with paypal is like pulling teeth. The most public one is what happened to the poor bastard who built Minecraft.
And trusting middlemen is still a bit of an issue. See Knuth’s rationale behind the bank of San Serriffe. You’d think a bank would hire people who know better, but apparantly not.
On second thoughts “poor bastard who built Minecraft” is probably a bad way of putting it. He is a millionaire now that he’s gotten the money out of paypal. Still, I bet it was a bit of a shock.
One interesting aspect of the smartphone wars is the emergence of a new kind of Capitalism that is much more egalitarian, much less capital-heavy or capital-oriented than the former ones.
In 1920 or 1960 if J. Random Engineer had a great idea for a new radio he had to somehow find enough capital investment in order to buy or build a factory.
The HTC guys don’t own factories. They rent manufacturing capacity which is a whole world of difference. In fact they have very few assets: some smart people on their payroll, plus enough capital to finance the manufacturing and distribution of the next batch, plus the usual stuff, marketing, PR, admin.
This I find exciting. If it becomes more and more possible to be a succesful entrepreneur – make a new product and actually be succesful with it – without having to become something like the classical Capitalist ownership-class type of person or organization who owns office buildings, factories and suchlike, the possible economic, social, and political consequences of this can become very interesting. For starters, it will become more and more difficult to frame it politically as “haves” and “have-nots”. HTC as a corporation is not exactly a “have” in the same way as General Electric or Siemens is a “have”.
@Shenpen:
Yes, I agree that the barrier to entry is far lower than it once was. HTC doesn’t need chip fab; their phones consist of off-the-shelf components surface mounted to a PC board. Apple fans may recognize this pattern as it is exactly how Apple, Inc., is structured today. These companies aren’t hardware manufacturers, they’re design studios.
Even in this scenario, though, the venture capitalists don’t entirely go away unless the entrepreneur either has his own startup cash, or can manage to get a very large down payment from one of those more traditional big companies. Granted, you don’t need to build a big factory or anything like that, but you’ve got to have enough operating capital to get you through your first few production runs.
FWIW, Alvin and Heidi Toeffler predicted this a long time ago. We’re just now starting to come into the information age, what the Toeffler’s call the “Third Wave”.
>The Black-Scholes theorem applies; a portfolio of options is more valuable than an option on a portfolio.
(that observation/result is actually Merton’s, although i guess it’s valid to say it derives directly from B&S eliminating the error-term from the predecessor’s work during a stroke of inspiration in the common room when Scholes was describing his day’s trading-floor consulting discoveries (despite what people are taught now, B-S did *no* maths for their landmark paper!))
>that observation/result is actually Merton’s
You are correct; it was a thinko on my part that I did not write “Black-Scholes-Merton theorem”.
Some years ago during an email exchange with John Quarterman I realized that there’s a deep connection between the Black-Scholes-Merton theorem and the expected utility function of a network of peer nodes. Remind me to blog about this sometime.
frankly mate, i’m more than impressed you’ve even HEARD of Merton, let alone aware of his work’s implications.
>Some years ago during an email exchange with John Quarterman I realized that there’s a deep connection between the Black-Scholes-Merton theorem and the expected utility function of a network of peer nodes. Remind me to blog about this sometime.
a/ sounds very interesting. i look forward to hopefully seeing it. i’ll remind you in a few months maybe ;)
b/ OK, that’s verrry weird. i had to doublecheck myself it was Merton (20yrs (!) since i studied this), so googled, and ran across a fascinating looking paper that’s not too dissimilar and looks like it might appeal to your anthropological/economics bone.
http://www.scribd.com/doc/2529714/Modularity-in-the-Design-of-complex-engineering-systems
“This paper will make three basic points.First, we will show that modularity is a financial force that can change the structure of an industry. Then, we will explore the value and costs that are associated with constructing and exploiting a modular design. Finally we will examine the ways in which modularity shapes organizations and the risks that it poses for particular firms.”
can’t print it without facebooking so set it aside till i had time to read it onscreen, but i’m stuck at home with virus tonight so might have a crack at it now…
>ran across a fascinating looking paper that’s not too dissimilar and looks like it might appeal to your anthropological/economics bone.
ROFL. I suspected it when I read your description, and yes; you’re referencing the people who taught me how to apply Black-Scholes-Merton in a broad qualitative way. It’s central to their analysis of payoffs from engineering modularity. Baldwin and Clark subsequently expanded that paper into an excellent book called “Design Rules”. By all means read it.
just popped back in to say “definitely worth a read — and maps neatly/complementarily onto your own observations re hacker cultures/economics flourishing within fixed Design-Rules environments (in their terminology).” and had my own ROFL moment when i saw your response.
their thoughts re IBM/360 and the consequences of mgt not funding investigations of module redesign, though, are thought-provoking for the wider business-approach implications for R&D-model companies. well, assuming they haven’t already told you!
Well…TechCrunch and Mashable are both echoing Eric’s prediction for an Android victory in 2011.