Leadership, Strategy and Qt
agile-workers.com
agile-workers.com
Edit: To be clear, I my intent wasn't to question Mirko, but to point out that the title here was added on HN; it wasn't the title he used and I didn't read his post as saying the entire Qt team was fired.
4 1/2 years later, Nokia is pulling the trigger on it.
My question is this, if Trolltech wasn't part of Nokia's strategy anymore, why not just spin it out?
And to compare to an equivalent business, Novell tried to shutter down Mono, thankfully the people behind it forged a new company out of it, called Xamarin.
I keep seeing cases of companies being acquired, only to then be semi-destroyed or shut down later on. Sometimes you get the impression the the acquirers only want to buy others just to shut them down and stop them becoming a threat.
For example, look at what Google did to Jaiku, or Slide, or Dodgeball.
Why can they not follow the example of HP and Agilent? Agilent was spun out of HP: http://en.wikipedia.org/wiki/Agilent_Technologies
I'm coming to the view that when you start up a business, if you care about what you've built more than you care about money, and what you've built is profitable, then you shouldn't sell to a big company, because chances are they will destroy it.
They also turned it into a proper open source project. There was no public repository history under Trolltech. Nor did Qt under Trolltech have a working contribution process; now tens of thousands of contributions have flown into Qt from the outside. There's even review rights and write access for non-employees, and maintainers that are non-employees.
Qt today has a much better chance of surviving Nokia than it had a chance of surviving Trolltech back in the day.
In dollar terms, Trolltech in 2007 was doing $30m Rev and making $8m profit. Obviously that's 5 years ago and I don't know what the current fundamentals are, but it would be interesting to know if that business within Nokia was doing more revenue or less, and what it's costs were.
If I were a former TrollTech employee, I would suggest taking a leaf out of Xamarin's book: http://gigaom.com/2011/12/12/xamarin-mono/
That's assuming a newly spun-out neo-Trolltech could be profitable again using the same business model as before, which isn't the case. Trolltech's business model was to sell Qt licenses to folks making closed-source software with it, who couldn't or wouldn't deal with the GPL-and-similar licensing of the open source version. This business model mostly died with the move to LGPL: There hardly any money left to be made selling Qt licenses now.
That leaves Qt-related consulting/development work, which companies like KDAB, basysKom, FrogLogic and others are doing.
Because they think it will bring in less than what it costs them to sell it?
You cannot just put a "on sale" sticker on part of a company and sell it. It will take an effort to find a buyer, and one (likely a larger one) to define what the sale will contain (patents? Copyrights? What source code? Which employment contracts? Pension rights? Etc.
First of, the Qt Project, that is the sum of all contributors within Nokia and without and so also including KDE, is currently busy putting together the next major generation of the technology, Qt 5. But our ongoing 4.x release series is still based on Qt 4, which Nokia's developers have not been targeting for a while now. Instead, commercial support of Qt 4 has passed to Digia some time ago, and they have been putting out new point releases since. So in the short term, this will not affect upcoming releases of KDE's platform, workspaces and applications.
In the longer term, the loss of the Nokia-employed workforce would obviously hurt the Qt Project considerably. Hopefully, however, this loss would not actually be quite that complete: It's unclear at this point whether Nokia might try to sell Qt, which could preserve the workforce entirely in the best case scenario. Alternatively, many of the developers who spent much or all of their professional careers working and loving Qt would likely band together and go on in some way, or find new homes at any of the Qt-focussed software development consultancies out there, which might also band together with us and others in the community to form a new home for the project in Nokia's absence.
For the cynics and skeptics in the audience, it also always pays to remind everyone that for the worst case, the community does have a poison pill in place: http://www.kde.org/community/whatiskde/kdefreeqtfoundation.p...
for the development of Free Software and
in particular for the development of KDE software.http://www.kde.org/community/whatiskde/kdefreeqtfoundation.p...
I think they will manage somehow...
Cheers, Mirko.
I only used Qt briefly, but it was the only time I actually had fun programming in C++.
http://webcache.googleusercontent.com/search?q=cache:QAwzH5I...
Will they beat Android/iOS? Maybe. Are they better to develop on? In my (and many others) opinion, yes.
But don't take my word for it, find out yourself.
As for major new development, well, what is lost is a matter of opinion. Already with Qt5 the focus is no longer on native widgets (it's vaguely shocking how the project's path while at Nokia has got so distorted and diverged from the story that made Qt a success).
The Mono team did not recreate what they had in Novell.
1. Mono in Novell put a lot of work on building Mono as a general platform. Xamarin is focused on where the money is: Mono for Android and iOS.
2. Mono in Novell was developed to form part of the SUSE Linux desktop. Xamarin does not do Linux, and in fact the devs themselves have apparently mostly personally switched to OS X.
3. Practically all Mono work at Novell was open source. Most of the work at Xamarin is closed source, the business model is to sell software for iOS and Android.
There's nothing wrong with either model, but just pointing out, they didn't just recreate the same thing outside of Novell. The two situations are hugely different, so it isn't surprising they are not that similar.
Will the pace of progress slow down? Sure, you're going to be losing a lot of developers all in one go, who made the majority of the commits. Is it game over? Well, no, it will end up community lead by the very smart cookies in the community.
It's hugely sad Nokia has gotten rid of it, but if they hadn't it wouldn't have made any sense. They don't use it and they'll struggle to find uses for it now unless they ported it to Windows Phone 8, but that's a whole other level of interesting fantasy.
Google should buy Qt. if Android was based on Qt/c++, not java, then Android would be the perfect platform!
Shuttleworth announced ending Canonical support for KDE (Kubuntu carries on as a full flavour). The one developer that Canonical paid has since I believe taken a job elsewhere.
http://arstechnica.com/business/2012/02/canonical-ending-sup...
Unity2d is being phased out in future versions of Ubuntu
http://askubuntu.com/questions/134346/why-is-unity-2d-being-...
Canonical is a very small company. Rounding error compared to Nokia.
No, it wouldn't. It's significantly easier for bad developers to make mistakes in C++ than in Java (one of Java's few advantages). Manual memory management alone would account for a significant increase in crashy apps. And while afaik, iOS has manual memory management, it also has Apple's App Store review process to at least ensure apps don't crash from a segfault every second launch, or similar issues.
Plus, I'd rather figure out a segfault bug than improve a laggy java program.
Thanks and goodbye, but thanks!
Consider the alternative; continuing internal struggles between the Windows and Symbian/Qt teams, funding all sorts of frivolous projects while strapped for cash. With Windows Nokia has a chance of surviving (although it doesn't look all that brigh right now), but without focus and determination death is certain.
1) Implying that Elop is not competent or responsible because he fired the team rather than "talking to them" is ridiculous. It is the only sensible option. Qt developers are not .Net developers and retraining so many at one time is impossible.
2) Finding developers for Nokia's strategy won't be a problem. There are lots of .Net developers around. And I am sure more than a few would be interested in the mobile space.
3) I fail to see what possible benefit Qt brings to Nokia's future strategy. They are unquestionably a Microsoft shop now. It is irrelevant how great Qt is as a technology.
4) The blame for Nokia's current woes does not lie with Elop and his current strategy. It lies with the previous one. And those employees need to start taking some responsibility for letting Nokia lose control of the industry. And it is not just CEO level it is product manager and engineer level.
Just FWIW, Windows Phone 8 adds a native development option, so .Net is no longer the only game in that town.
(It follows that it may be possible to put Qt there, too, but that's not really what I was going for.)
> 3) I fail to see what possible benefit Qt brings to Nokia's future strategy. They are unquestionably a Microsoft shop now. It is irrelevant how great Qt is as a technology.
That's true, but to add the historical context: Until this round of layoffs, Nokia had several hundred people at their Ulm office work on a new Linux/Qt-based mobile operating system. It is now, with that project shut down, that they have no internal use case for the technology anymore.
Many of the workers at Ulm were genuinely surprised at this development.
It's common sense that Nokia was not going to have a use for Qt or anyone involved with it once they moved to a MS strategy. I would honestly love to know what is being communicated internally for Nokia employees to be surprised at this.
Not because everyone will make the transition but because what separates a great QT/.net/java/c++ developer from an average one has vary little to do with experience. So, while such transitions are a great time to remove deadwood, and you want to find some experts in the new stack laying off the old team is expensive and rarely worth it.
The radical perspective is that they have a "MS strategy" and that that "MS strategy" will do them any good.
You seem to believe something which, judging by Nokia's results and MS recent (10 year) history, it seems ludicrous if not suicidal.
Plus, if that "Surface" thing goes well, what's to stop MS trying to build their own phones? From a mobile leader that it was in the past, Nokia will become --at best-- a insignificant MS home making division, and at worst an ex-partner with no recourse.
Of course the more probable outcome is that Windows Phone 8,9,10,x will continue it's Zune-like irrelevance, and Nokia with enter a RIM like free fall...
The two biggest strategic blunders you can commit: One, staging a ground invasion of Russia in winter. Two, partnering with Microsoft.
wrong. see http://communities-dominate.blogs.com/brands/2011/12/calcula...
Nokia is not in trouble because of Elop. It is in trouble because of two trends. One is the rise of the middle class in emerging markets who don't want cheap, nasty feature phones. And two is the move from both Apple and Android into lower and lower price points.
And the fact is that Meego had nothing going for it. It had no apps and no related ecosystems like Google, Microsoft and Apple have. And as we are seeing with RIM developers simply aren't interested in building 4 versions of their apps. So this idea that Meego was going to save the day is pure delusion.
What a shame...
Aside from timing and suicidal memos, given the obvious incompetency of Elop, I doubt WP was the right decision at all. Of course I do not have all the information that Elop had back then, but I wonder did they consider other options, such as forking android (maybe in cooperation with amazon), seriously. I wonder what role his previous employment played here.
Sorry but only a few geeks care about QT, and other developers care about the ecosystem and go where the money is. I wrote some QT code about 10 years ago and it was good, but is it so great compared to iOS/ObjC, Android/Java and WinRT/C# that developers would just flock to it blindly? I seriously doubt it.
I care about QNX too, it's such a cool RTOS.
But what can it do for RIM that can't be done by Darwin, Linux kernel or the NT kernel? As I said, us geeks by our very nature get excited by the things that the users don't, unless they can see a difference in usage.
Also, the fact that it's POSIX compliant helps immensely when porting existing code.
Moreover, RIM is doubling-down on Qt. It's unclear what this move by Nokia does, but I think you'll see a lot of Qt developers being hired by RIM. Whether they'll pick up the project and run with it is another story.
(Not trying to be snarky, but I would indeed be concerned about RIM following the same path as Nokia, especially considering the massive job cuts they've recently announced.)
RIM may be on the decline right now with BBOS, but you simply need to look into the great work they're doing to seed the developer community with the support they need to create great apps for BlackBerry 10. I know the typical thing to spout is that RIM is dead, long live Android/iOS, but if you actually looked into it yourself, you would see that RIM is doing some really great things with BlackBerry 10. They're reaching out to FOSS communities, seeding actual hardware, and making their tools easier to use than ever.
But I know that many, maybe yourself included, will call me a fanboi, downvote me, and won't bother looking into it themselves. They'll continue spouting this nonsense about RIM's death, never changing their opinion, and taking Apple/Google's word for it. That's fine, continue being uninformed.
The rest of us developers with an open mind will happily pick up the slack on the sinking ship that is RIM. We might be wrong, and you might be right. But what happens when we're right? Where will you be?
That's great, but the health of a company and it's products depends on a bit more than their developer relations team. The engineers at HP/Palm were also doing great things with webOS before it got axed entirely.
> But I know that many, maybe yourself included, will call me a fanboi, downvote me, and won't bother looking into it themselves. They'll continue spouting this nonsense about RIM's death, never changing their opinion, and taking Apple/Google's word for it. That's fine, continue being uninformed.
I have no personal stake in the matter nor do I particularly care whether RIM survives or dies. I was only making the observation that if I was a Qt developer that just got laid off from Nokia, I would not want to take up a job at RIM for fear of the same situation repeating itself in 12-18 months. I mean, in the best case scenario it's a gamble that BB10 succeeds in the marketplace at all. If I had a family to support I would be looking to find a less risky proposition.
> The rest of us developers with an open mind will happily pick up the slack on the sinking ship that is RIM. We might be wrong, and you might be right. But what happens when we're right? Where will you be?
The same place I always was? RIM's success or failure is entirely irrelevant to me.
I did not mean applications like yours and never implied it was a toy.
That being said, we already knew Nokia was going to dump it, they already sold off the commercial support and licensing division and made a shallow claim "we're going to continue contributing to Qt," which almost always should be read by adding "until..." at the end.
I'm happy to see Qt back in the hands of someone who cares about it. Personally, I'm not much interested in whether or not Nokia plays in the sandbox.
But Android? On devices with a completely different form factor, input schemes, and resource requirements? I don't think that is feasible, particularly from a user experience point of view.
And having ported a fairly successful iOS app to Symbian^3 (using mostly Qt Quick/QML), I would seriously contest the claim of 1) Qt working like a charm on Symbian and 2) how well Symbian works in general.
Rather, that to provide the best user experience on each different type of device (i.e. to follow platform conventions and paradigms and take advantage of unique device capabilities where present), it's best to use platform specific tools and design your application to be consistent with the expected user experience for that particular device.
Go back and read his blog in 2008-2009 where he conclusively proved every time he wrote that Nokia phones were much better than the iPhone and Android, and that no one really wanted a touchscreen device anyway.
I believe the author (Tomi Ahonen) gives a window into how most businesses product management and marketing work for good or for bad.
Companies start with a technology base for a product and hire pretty smart people to look at the market and identify the features which will sell more units or the segments they need to target. The decisions are usually very fact driven because they have to justify in terms of a business case, and in general I think Ahonen is factually very solid.
If you looked at the figures and facts from 2008-09 I don't think you'll find much wrong and public position of Nokia (outside of the US tech community) was pretty strong.
However, the fundamental problem about relying solely on this approach is that it is a. very reliant on execution of your engineering org. b. takes no account of the ability of engineering to scale c. remains susceptible to disruptive big-bang technologies.
This is why product development organizations need a coherent personality rather than be treated as factories and prescient leadership.
What I disagree with is his mindset: it is all Elop's fault, and there is a vast Microsoft conspiracy to destroy Nokia.
I don't agree with Elop's decision to go to Windows Phone. But I do agree with Elop's conclusion that the Nokia platform(s) were "burning" and that they should be abandoned.
I think he should have gone with Android, but I understand the reasons why he went with WinPhone.
Meanwhile, Tomi Ahonen still claims that Meego was the ultimate solution to all of Nokia's problems. Ahonen has never accepted that Meego was 2 years too late for it to be a viable player, and he's never accepted that Nokia isn't a big enough player to build a sustainable ecosystem around an operating system that was so late.
If we are talking about full time employees - they should have been given the option to work with .NET if this is what they wanted.
Also - employees have to take responsibility for Nokia's previous strategy? Weren't you just talking about those employees like an easily replaceable cogs? Get rid of a few "C++/Qt"s and get a few new "C#/.NET" ones.
Regular employees are responsible for managing their own career not the company.
About Elop - his statement about the burning platform reminds me about the Léo Apoteker public announcement about dumping the PC division of HP. Even if we assume both decisions about new strategic direction were correct - the public announcements were not very helpful to say the least ...
-- 1) Implying that Elop is not competent or responsible because he fired the team rather than "talking to them" is ridiculous. It is the only sensible option. Qt developers are not .Net developers and retraining so many at one time is impossible. 2) Finding developers for Nokia's strategy won't be a problem. There are lots of .Net developers around. And I am sure more than a few would be interested in the mobile space.
Can you not see how the clean slate approach where you leave behind all of the Linux/Qt/Symbian baggage could be appealing ?
It's not dissimilar from what happen with Apple's failed Copland project where a new team came in (NeXT), took control of a new OS/apps strategy and ultimately saved the company.
The question is - how is this clean break and leaving "the baggage" behind going work?
I can't stop thinking about this imaginary company:
They have Java/Linux based system that is not selling very well so they hire new CEO. The new CEO announces to the company's customers that the Java/Linux is "burning platform". He fires all 100 Java developers and hires 100 C#/.NET developers ...
Somehow I think this is not going to end up well ...
A competent CEO (or CTO) will try to sell the new platform to developers by pointing out the advantages. He would hire maybe 25 experienced .NET developers to replace the 25 developers that left because they believe more in Java/Linux in the long term. He may have to let 10 developers go because they just don't want to learn anything new ... You end up with a team that still has the business knowledge, is committed to moving to the new platform and has people with experience about it.
Meanwhile you still make money selling the Java application until you are ready with the .NET one.
Isn't this 101 common business sense?
-- Can you not see how the clean slate approach where you leave behind all of the Linux/Qt/Symbian baggage could be appealing ?
Since RIM is doubling-down on Qt for BlackBerry 10, it allows them to recruit experienced developers to help them. Maybe they can leverage it to pull off something similar to Apple with NeXT.
Here are some job postings:
http://webcache.googleusercontent.com/search?q=cache:QAwzH5I...
So you don't get rid of them all.
I saw a team go from LISP to C++ in the space of a few months. Some people got disgusted and left, some people got disgusted and stayed (and we'd wished they'd left), but a surprising number prospered, hunkered down and shipped.
As long as you're in roughly the same domain (say, "phones") I imagine it's less expensive to keep most of a team around, deal with attrition, and hire in new folks to fill the cracks.
1) Yes, common wisdom. 2) Experienced devs can learn anything if they want easily, what happens if they don't want to?. 3) Yes, a little 4) No is not.
It is more complex than than. Employees are people, and people are complex.
First there is a philosophical issue. The best c,c++, Qt programmers love Unix, and that's one of the reasons they work there. If you want to force them into Windows and .NET those people will take the door. It is not really difficult for them to find another job that needs c, c++ if they are great.
I know it sounds ridiculous from some mindsets who believe everything in life is money and he who pays is the master that could slave their serfs(suit's mentality) but geeks tend to be the more idealistic people I know.
Once the best programmers(best programmers could automate things and be more than 10x efficient than standard ones) are gone the entire system collapses.
Except c/c++/Qt programmers -- they're pretty simple to predict, apparently.
The best devs I know tend to accept challenges to learn new systems and do well whatever they're working with, even if there's a particular language/platform they love more than others.
For instance say I am looking for a bit of text.
In *nix I have vi in which I can type "esc/foo." Now, before I can blink the I am at the first match. Hitting "/" goes to the next one just as fast. If I don't know what file to look at, grep will tell me quickly.
Visual Studio has a search function, but it's not regular expressions by default, will not work on all of the design surfaces, and requires going through several dialogue windows. Typically I have to put my hand on the mouse to get any results.
For some tasks platforms and tools can make a big speed difference.
However it is true people are complex and hard to predict.
Perhaps some c/c++/QT programmers when given a, a shiny new windows based development environment, may enjoy the extra time they spend searching through source code. Who knows?
The issues were: (from the gravepine)
Old Symbian team kept beating the Symbian drum and heading towards the cliff
At the same time, Linux team developing Maemo, viewed by the rest of the company as "the linux kids and their toys"
Linux team trying to assemble a jigsaw puzzle of several Linux services (and since most of them are not focused on mobile, and some are just crap but it's the only solution) it takes a lot of time
Not to mention "ok, let's merge with Intel and rewrite a crapload of stuff" then "ok, let's buy Qt and rewrite a boatload of stuff with Qt" (not sure how those are ordered, still)
Meanwhile Android just built the bare minimum, discarding lots of Linux infrastructure and came ahead
So when Stephen Elop got there, sorry, boat had already sank
I don't believe being (yet another) Android OEM would have done them any good either. Microsoft was the way to go.
And don't say Microsoft wouldn't allow it. Windows Phone was irrelevant and Nokia was Microsoft's few chances to change that. And more than that, Microsoft needed access to Nokia's patents (to keep up the pressure on Android, among other things). Nokia was in the driver's seat and could have written their terms. Instead, they let Microsoft effectively acquire them for peanuts. If Nokia had only kept their options open and jumped on what worked, today's mess was entirely preventable.
I agree, but hopefully/probably Nokia got something in the deal making, for shutting down a Linux based competing system.
The switch to Windows Phone was tactical in that sense. The strategic aspect of the switch was that Microsoft offered a long term partnership, allowed them to liquidate maps, and could provide a clear technology roadmap for their product line.
Nokia's software was good and competent, but they simply don't have access to the mountains of UX data needed to develop the next generation of OS's to compete on the touch screen.
This statement doesn't seem correct to me. Nokia has a vast amount of research on mobile UX, even from before Apple joined the play, probably as no one other company. As a (stellar) example, check out Jan Chipchase's work (http://janchipchase.com/) with a ton of in-loco research on mobile phone usage, and many research papers presented - with a particular focus on the non-developed world.
In my opinion Nokia suffered from big-company chronic slowness: the current cash-cow was still providing the income, hence all future investment was unfocused, under-funded.
PS: A nice question is still how much/if any of the investment of Nokia in the undeveloped world will still be able to bring some returns. Windows Phone completely eschews that...
OK, but the reason tethering is hard is that carriers want to charge extra for it, not because of any decision of Nokia (or any other phone manufacturer)
Have you looked at Android OEM's financials lately? Samsung is the only one making any profit, HTC is barely ekeing out one, LG, Sony, Motorola etc. are running in losses.
And bringing out an Android phone would cut off the $1B/yr lifeline platform support from Microsoft, which Nokia knew it would need in the painful transition. Nokia did approach Google for a similar deal, but Google told them to take a hike. Also, they're licensing maps etc. to Microsoft for a huge amount too.
Also, people seem to be ignoring that Nokia's signature cash cow feature phone business is collapsing because of ultracheap Chinese clones and white label phones being dumped into the market. Nokia trying to start their own ecosystem with QT would've been foolish in the face of entrenched competition.
I don't understand the notion that QT is some magical technology that's so much leaps and bounds ahead of other GUI APIs like iOS, Android, WinRT etc that developer would automatically flock to it. QT is definitely nice though.
In short, Nokia knew that they were going to go through extremely tough times for a couple of years regardless of what they pick, and they chose a partner with deep pockets($60B cash in the bank) rather than going solo.
Nokia uses gate5 Maps and acquired gate5 AG in late 2006. According to this http://www.bing.com/community/site_blogs/b/maps/archive/2012... there is a joint venture between Nokia and Microsoft/Bing with regards to maps.
> they chose a partner with deep pockets($60B cash in the bank)
Is only half the story. They chose a partner who has a record of being a horrible partner for companies in their condition -- which is in a lot of ways similar to both HP and SGI pre-Windows (good hardware, trying to get out of the OS business, rely on Microsoft). See http://en.wikipedia.org/wiki/Richard_Belluzzo
So, this partner has a horrible relevant track record, twice. I don't see how Nokia can expect (or could have expected a couple of years ago) to fare better than SGI or HP's high end computing unit did.
> Nokia's worst case here is more then likely getting bought vs. being left for dead like SGI.
... for 10% what they were worth at the time Elop gave the "burning platform" speech.
SGI was technically bought as well. That's really a bad resolution for Nokia, and is far from being a "worst case". If MS decides to produce phones themselves, it is likely Nokia will be left to whither and die.
It might look unlikely now, but unless you were in the know, just a month ago it seemed very unlikely that MS will make their own tablets.
a) already have a good relationship with the carriers
b) have someone like Steve Jobs
Nokia at least has (a), Microsoft has neither (Verizon likely still has enough Kin that they won't be hurting for paperweights for quite a while). Microsoft can't afford to kill their only guaranteed in with the carriers.
Also, another parallel is Palm. They had PalmOS phones when they partnered with Microsoft and made the Treo WinMo line that ended up competing and reducing their own offering's relevance.
Nokia's market cap was over $38B at the end of 2010. It was just under $32B at the end of Q1 2011 (after the "burning platforms" memo). It is under $9B today. As you note, not all of that is Windows Phone, but plenty of it is. Microsoft's $1B/year is peanuts. To put it another way, how much was avoiding the "Nokia rejects Windows Phone" headline worth to Microsoft? Plenty more than $1B/year, I bet.
Second, Nokia's feature phone business is one big reason to have an Android project going. Low-end Android phones are one of the big things eating feature phones. Until recently, Nokia had been spending plenty of effort (e.g. Meltemi) trying to create something to compete with Android at the low end. They shouldn't have been re-inventing the wheel. They should have picked up Android and leaned on some of their other strengths (e.g. manufacturing, distribution, etc.) to win low-end devices.
Third, I don't think Qt is a magic elixir. Establishing a Qt ecosystem easily might not have worked. But it was a way to transition off Symbian without immediately deprecating those devices (and we saw how immediate deprecation turned out). And it was a decent story to tell developers: "Build with Qt and you can make your app available on every mobile platform you care about (except maybe Blackberry)". Of course the devil is in the details and it wouldn't be that easy, but plenty of developers want a cross-platform mobile story. It could have been the hook that got an ecosystem off the ground, so it was at least worth trying. And if it didn't work, you'd have been offering devices from each of the alternative platforms you could consider. It would be a lot easier to pick if you could see how those devices were doing in the marketplace.
Fourth, the measure of success of including Android (and I'll emphasize again that I didn't say don't do Windows Phone, too) isn't whether they would have remained profitable while doing so. It is whether they would have had better options today and would they have lost less money along the way. I think the answer is obviously yes on both counts. If nothing else, being able to point to a more successful Android line and a less successful Windows Phone line would have driven up their eventual acquisition price.
Samsung has done a very good job with marketing the Galaxy brand so far. In 2012, you know the S III is the best Samsung phone. And SII is the second best. This is simple to understand, even though Samsung makes dozens of Android models.
But if someone asks you which Sony phone should I buy: http://www.gsmarena.com/sony-phones-7.php . Good luck with that. The consumer chooses to go with the simpler options of Samsung, Apple or even HTC.
Sony in particular had(and has) some massive opportunities with deep Android+Google integration with their Playstation Network platform. But they continue to shoot themselves in the foot by throwing money at half baked phones. And dedicated gaming devices, which have taken a massive beating from smart phones.
This is only true on the phones themselves. Nokia are mostly Java based on the server side.