Blog Post

Symbian: A Lesson on the Wrong Way to Use Open Source

The Register’s Andrew Orlowski recently offered a hopeful eulogy for Symbian, the still-dominant but fading mobile operating system that Nokia (s nok) took open source in 2008. Nokia hoped to revive Symbian’s importance, which once dominated more than 50 percent of the mobile market, by reinvigorating its developer base in light of a rush of Linux-based operating platforms like Android (s goog) and LiMo. It hoped in vain.

For years, companies have looked to open source to salvage dying products, and each time these efforts have failed. Often dismally.

After all, if a product can’t make the grade as a proprietary product, it will almost certainly fare worse as an open-source product. Great open-source projects are founded on great code and robust community, two things largely lacking from failed proprietary products.

With respect to Symbian, its problems go far beyond its source code, but a lack of source code is the foundation for several of them. A year after announcing that its code would be available under an open-source license, the Symbian Foundation hadn’t actually released the code. Small wonder, then, that Symbian also couldn’t match the app store momentum of Apple (s aapl), Google, or even its one-time sugar daddy, Nokia. No code, no developers.

No developers, no relevance, as its dying market share demonstrates:

Symbian: It used to be a contender

An open-source strategy generally favors challengers in a market, not incumbents. This is the genius of Google’s Android assault on Apple’s (s aapl) uber-closed iOS, and it is equally the reason that a swath of open-source upstarts — from Alfresco to BonitaSoft to Cloudera – has mounted successful challenges to the incumbents in their respective markets.

Open source, then, is a great way to spark or accelerate momentum. It’s a terrible way to reverse a product’s decline. If anything, it does the opposite by calling attention to the lack of interest in a project — freely visible through SourceForge or JIRA tickets or forum activity — and thereby compounding the indifference to the product.

Was open-source Symbian dead the moment it was announced? Perhaps. It certainly needed to demonstrate significant developer interest in the platform immediately upon announcing the code would be open-sourced, and then continuously thereafter. It did neither.  As a result, one by one Symbian’s key handset licensees like Samsung have dropped it for Android or other open-source alternatives. And it started way too late: by the time Symbian announced its code would be open-sourced, developer survey data already showed Android and OpenMoko interest running rampant. In the past two years, developer interest for Symbian has all but evaporated.

While interest remains relatively strong for Nokia’s Qt developer tools, it can’t compensate for fading interest in the overall Symbian platform.

Had Symbian gone open source when still strong with developers, and had the Foundation done a better job of engaging developers, it might have had a chance to survive as more than a dusty Wikipedia reference.

Open source isn’t a one-time announcement, coupled with a code drop. It’s exceptionally hard, ongoing work that requires equal parts evangelism, programming, and customer success stories to keep developers believing that their work matters. It’s especially difficult, as Drupal founder Dries Buytaert intimates, when a commercial entity gets involved, because it can frighten away community.

These principles aren’t exclusive to open source, of course. Consider Skype, for example. Skype appears to be pulling a Microsoft (s msft): turning inward, seeking to build out its own ecosystem rather than fostering a robust, third-party developer ecosystem. Microsoft is famous for its developer outreach on the enterprise side, but is equally famous for a failed go-it-alone approach in consumer technologies. Its dying brand, as CNN reports, reflects this failed strategy.

Both companies need to engage and encourage a third-party developer ecosystem. Open source is a critical way to accomplish this.

But neither they nor anyone else can hope to use open source as a purely palliative remedy for what ails them. Open source can be used to inspire and complement successful products. It can accelerate momentum. What it can’t do is resurrect dying technology products.

Disclosure: I work for Canonical, a Linux vendor. I am also a former Alfresco employee.

Related content on GigaOm Pro (subscription required):

16 Responses to “Symbian: A Lesson on the Wrong Way to Use Open Source”

  1. Alessandro Pedarra

    I think, like others, that Symbian (3) is actually good and will be better in the future.

    Nokia has done a lot of great things recently and their future is looking really bright after some unlucky smartphones (like the N97).

    Today Nokia is pushing really hard with the new (and open source!) platform (Symbian ^3), the new phones (especially the new C series), and the QT framework.

    I will happily buy one of their new smartphones this holiday season.

  2. kushi purac

    “After all, if a product can’t make the grade as a proprietary product, it will almost certainly fare worse as an open-source product.”

    Netscape–> Mozilla Firefox

  3. Developers in large companies often have an established way of working. What’s clear with Nokia is that that way is not or was not uniformly good. The complaints about consistency in the UI etc are reminiscent of people working in silos.

    If people can’t work together internally (my conjecture) then they aren’t really going to do it externally either. In other words there is more to open source than dumping the code.

    The dropping of S^4 etc sounds like a brilliant move to me. I bet that it removes a *great* deal of complexity. The idea of producing continuous improvements rather than epic releases is also much more like the way Open Source works. It’s all much less rigid.

    With some hardware for developers to play on (e.g. a beagleboard and the new S^3 phones) and some better collaboration from a re-organised Nokia, I think that open source Symbian has a long term future. It’s actually essential that Nokia’s Symbian outfit learns to work in a way that’s compatible with OSS.

    With Meego, for example, there is no other option.

  4. soon the chart will say Qt instead of symbian. a post on that would have been far more insightful and informative. no brownie points for being bearish on symbian.

  5. Just a thought but maybe Nokia is quite happy with the situation it finds Symbian in now. Since going open source all of the partners of Symbian who actually owned stakes in it have given them up to form the open source Symbian Foundation. Now that it seems SE and Samsung are not interested in further development and Motorola went away ages ago Nokia is now free to focus the code on improvements it needs which should if anything help it rapidly improve.

    Another point is that before open sourcing Symbian all licensees had to pay a license cost per device sold. As nokia sells by far the most Symbian devices it’s definitely saving it self a ton of money not paying a license anymore; savings it can invest in things like QT.

  6. Two counterexamples to your thesis: Netscape, re-emerging as Firefox; and Blender, re-emerging as Blender. Each required a massive re-write. Each is now a well-known and robust success.

    So there is a right way to go about it. It does require actually releasing the code.

  7. They released the code. They were handing it out on flash drives at MWC in February. You can download the current version here:

    Also, while I have issues with the way Symbian has handled… well… everything since the foundation was formed, it still has an active developer community.

    And Microsoft’s consumer brand is dying? It’s nice that CNN had that headline, but even the article admits that Xbox is far from dead. And the article doesn’t even mention Windows, which is doing very well. Microsoft’s consumer brand is doind quite a bit better than, say, Canonical’s (I’m an Ubuntu user on my netbook, and Win7 user on my dekstop).

    There is still a ton of work to do on Symbian. A ton. But the gradual move by Nokia from the high end to mass market has been effective, and the road map is still strong. If Symbian doesn’t die in the next few months – which, if Nokia CEO Elop goes another way, it will – it is still a huge presence in the market.

    That said, I can’t call the decision to go open source a success. I can’t help but wonder if Nokia having full control over the last couple of years would have led to more timely releases, which would have helped. A lot.

  8. John G. Doe

    The lack of developers interest for core platform development is not an indicator of platforms viability. Nokia has a lot of R&D personnel working on the OSS components in house. Fujitsu likely has more as well. So have various big operators which are customizing the OS for their needs. Is it relevant that there is no interest from small group of hackers? How much of Android (your pet Linux with stolen Java crap on top) is developed outside Google? Zero or 0?

    Don’t confuse the platform development with the application development, the only segment which in current context has relevance. Consumers define interest in the platform by the number of iFart apps available and not by the number of committed bugs in Bugzilla. After all, few bug reports may mean few bugs … (wouldn’t that be sweet?).

    As for Samsung or Sony-Ericsson ever being “key” supporters of Symbian … dude, whatever you are smoking I want to make sure I’m not getting close to it.

    Now, how about you stop collecting random links to try to make them support a point, and rather get you a** behind Symbian as an OSS project, and try to be supportive about it. Why not start championing it for a change? After all, it may not be the OSS you favour, but it is the leader of the market space and that would clearly be a good position for an OSS movement to start from. How often do you get a chance like that?

    There are enough idiots writing today about Symbian dying, and tomorrow about it not having died today. These idiots have been writing the same thing for years now. The news is old and is not news. It is cheap dirty ignorant redundant gossip.

    So, what do do you say? Ready for some OSS work? Start by creating yourself a account and post here your firs bug report. Or better yet, post a bug fix. Let’s see what you can actually do, rather than talk about… Or is it your “church” affiliation more important than the greater good? Is your OSS-ness so shallow?

  9. Whilst I agree with the general points, I’m not sure two historical data points and two _predictions_ count as proof of declining market share. As mark points out, market share can go down as volumes go up. That is the point og MeeGo.

  10. Uh… Matt.

    When you say ‘dying’ market share you might notice that projected unit shipments go from 80 million in 2010 to 264 million in 2014. This is called ‘expanding the market’.

    Nokia just realised what the PC world realised a long time ago – no-one’s really interested in true open source software (no, not Google either). Incidentally Nokia have publicly stated that they will continue to use Symbian (their own approved version obviously) on C, E and X series handsets (5 and 6 numbers and above) however the N Series will use MeeGo – an open source Linux variant.

    So, you see, Symbian still thrives and still has a major sponsor. Which kind of makes your article as pointless as your articles over the last few years that ‘this is Linux’s year’. That never really happened either, did it?