Monday, June 8, 2009

Symbian: Evolving toward open

It's fascinating to watch the evolution as Symbian remakes itself from a traditional OS company into an open-source foundation. They've made enormous organizational changes (most of the management team is new), but the biggest change of all seems to be in mindset. A nonprofit foundation has a very different set of motivations and priorities than an OS corporation does. I get the feeling that the Symbian folks are still figuring out what that means. It's an interesting case study, but also a good example for companies looking to work with open source.

Symbian recently held a dinner with developers and bloggers in Silicon Valley, and I got to see some of those differences in action.

The first difference was the dinner itself. About six months ago, Symbian and Nokia held a conference and blogger dinner in San Francisco (link). It was interesting but pretty standard -- a day of presentations, followed by dinner at a large, long table at which Symbian and Nokia employees talked to us about what they're doing and how excited they are. The emphasis was on them informing us.

The recent dinner was structured very differently. The attendees were mostly developers rather than bloggers, and we were seated at smaller, circular tables that made conversation easier. They talked about their plans at the start, but most of the evening was devoted to asking our opinions, and they had a note-taker at each table. This had the effect of not just collecting feedback from us, but forcing us to notice that they were listening. That's important to any company, but it is critical to a nonprofit foundation that relies on others to do its OS programming. And it's essential for a company like Symbian, which has been ignored by most Silicon Valley developers.

So that's the first lesson about open source. The task of marketing is no longer to convince people how smart you are, it's to convince people how wonderful you are to work with. Instead of you as a performer and developers as the audience, the situation is flipped -- the developers are the center of attention and you're their most ardent fan.

It's an interesting contrast to Apple's relationship with developers, isn't it? It'll be fun to see how this evolves over time.

Here are my notes on the subjects Symbian discussed with us, along with some comments from me:


It takes time

Symbian said its goal is to have a lot of developers on the platform and making money, but that can't be achieved in three months. "In three years time," is what I wrote in my notes. That is simultaneously very honest and a little scary. It's honest because a foundation with its limited resources, working through phone companies with 24 month release cycles, simply can't make anything happen quickly. It's scary because competitors like Apple and RIM have so much momentum, and can act quickly. Still, in the current overused catchphrase of sports broadcasting, is what it is. An open-source company, based on trust, simply cannot afford to risk that trust by hyping or overpromising.

Speaking of Apple and RIM, Symbian made clear that it considers its adversary to be single-company ecosystems like Apple, RIM, and Microsoft. I didn't think to ask if Nokia's Ovi fits in that category, but that probably wouldn't have been a polite question anyway. Symbian also took some swipes at Google, citing the "lock in" deals they have supposedly made with some operators.

You get the feeling that Symbian is intensely annoyed by Google. It's one thing for a mobile phone newcomer like Apple to create a successful device; it's quite another for an Internet company to step into the OS business and take away Motorola as a Symbian licensee. I think one of Symbian's arguments against Android is going to be that Symbian is more properly and thoroughly open.

The question is whether anyone cares about that. Although the details of open source governance are intensely important to the community of free software advocates, I think that for most developers and handset companies the only "open" that they care about translates as, "open to me making a lot of money without someone else getting in the way." Thus the success of the Apple Store, even though Apple is one of the most proprietary companies in computing. Symbian's measure of success with developers will be whether it can help them get rich -- and I think the company knows that.


Licensees and devices

One step in helping developers make money is to get more devices with Symbian OS on them. Symbian said phones are coming from Chinese network equipment conglomerates Huawei and ZTE. They also said non- phone devices are in the works.

Licensees will be especially important if Nokia, as rumored, creates a line of phones based on its Maemo Linux platform. Lately some industry people I trust have talked about those phones as a sure thing rather than speculation, and analyst Richard Windsor is predicting big challenges for Symbian as a result:

"It seems that the clock is ticking for Symbian as technological limitations could lead to it being replaced in some high-end devices.... I suspect that the reality is that Symbian is not good enough for some of the functionality Nokia has planned over the medium term leaving Nokia with no choice but to move on."
Source: Richard Windsor, Industry Specialist, Nomura Securities

David Wood at Symbian responded that people should view Maemo as just Nokia's insurance in case something goes wrong with Symbian (link). But the point remains that Nokia is Symbian's main backer today. That is a strength, but also a big vulnerability. If Symbian wants developers to invest in it, I think it needs to demonstrate the ability to attract a more diverse set of strong supporters.


App Store envy

Another way to help developers is to, well, help them directly. Symbian said it's planning something tentatively called "Symbian Arena," in which it will select 100 Symbian applications to be featured in the application stores on Symbian phones. Symbian will promote the applications and perform other functions equivalent to a book publisher, including possibly giving the app author an advance on royalties.

The first five applications will be chosen by July, and featured on at least three Symbian smartphones (the Nokia N97, and phones from Samsung and Sony Ericsson).

The most interesting aspect of the program is that Symbian said its goal is to take no cut at all from app revenue for its services. Obviously that means the program can't scale to thousands of applications -- Symbian can't afford it. They said they'd like to evolve it into a much broader program in which they would provide publishing services for thousands of apps at cost. My guess is they could push the revenue cut down to well under 10% in that case, compared to the 30% Apple takes today.

It isn't clear to me if Symbian will produce the applications store itself, or work through others, or both. If it works through other stores, those stores might take a revenue cut of their own. But still, from a developer point of view it's nice to see an OS vendor trying to lower the cost of business for creating apps.

It's been interesting to see how many of the Palm Pre reviews this week have said that the iPhone application base is the main reason to prefer an iPhone over a Pre. I'm not sure how much purchase influence apps actually have -- at Palm, we had ten times the applications of Pocket PC, but they didn't seem to do anything for our sales. (On the other hand, Palm never had the wisdom and courage to advertise its apps base the way Apple has.)

--"Compared to the iPhone, the real missing pieces are those thousands of applications available on the App Store." Wired
--"Developer courting still seems like an area where Palm needs work. They've got a great OS to work with, but they have yet to really extend a hand to a wide selection of developers or help explain how working in webOS will be beneficial to their business. The platform is nothing without the support of creative and active partners." Engadget
--"The Pre's biggest disadvantage is its app store, the App Catalog. At launch, it has only about a dozen apps, compared with over 40,000 for the iPhone, and thousands each for the G1 and the modern BlackBerry models....It is thoughtfully designed, works well and could give the iPhone and BlackBerry strong competition -- but only if it fixes its app store and can attract third-party developers." Walt Mossberg

Anyway, if applications are the new competitive frontier between smart phones, mobile OS vendors should be competing to see who can do the most to improve life for developers. This is another area where Symbian's motives, as a foundation, differ from a traditional OS company. If you're trying to make money from an OS, harvesting some revenue from developers make sense. But as a nonprofit foundation, draining the revenue streams from your competitors is one of your best competitive weapons. Symbian has little reason to try to make a profit from developers, and a lot of reasons not to.


Driving Web standards

That idea came up again when we talked about web applications for mobile. As I've said before, I think the most valuable thing that could happen for mobile developers would be the creation of a universal runtime layer for mobile web apps -- software that would let them write an app once, host it online, and run it unmodified on any mobile OS. No commercial OS companies want to support that because it would commoditize their businesses and drain their revenues. But if Symbian's primary weapon is to remove revenue from other OS companies, a universal Web runtime might be the best way to do it. I asked them about this, and they said they're planning to use web standards in the OS "like Pre," and said they're interested in supporting universal web runtimes.

I'm intensely interested in seeing how the runtime situation develops. I think Symbian and Google are the only major mobile players with an interest in making it work, and Google so far hasn't been an effective leader in that space. I think Symbian might be able to pull it off, and become a major player in the rise of the metaplatform. But it'll take an active effort by them, such as choosing a runtime, building it into every copy Symbian OS, and making it available for other platforms. Passive endorsement of something is not enough to make a difference.


Other tidbits

Symbian said it's going to "radically simplify" the Symbian Signed app certification program, which may be very welcome news to developers, depending on the details. Many developers today complain bitterly about the cost and inconvenience of the signing program, and unless it's fixed it'll outweigh any of the benefits from Symbian Arena.

The QT software layer that Nokia bought as part of its Trolltech acquisition will be built into Symbian OS in the second half of 2010. I had been wondering if it would be an option or a standard part of the OS; apparently it'll be a standard.

Symbian plans to bring its developer conference to San Francisco in 2010, after which it will rotate to various locations around the world. This is part of an effort to increase Symbian's visibility in the US market. The company is creating a large office here, including two members of its exec staff. That makes sense for recruiting web developers, but it will be hard for the company to have a big impact in the US unless it gets a licensee who can market effectively here. In that vein, it must have been frustrating for everyone involved when Nokia announced the shipment of the N97 and it came in a distant third in coverage in the US (after the Palm Pre and the iPhone rumors).


What it all means

There are a lot of things that could kill the Symbian experiment:
--Nokia could decommit from the OS (or just waver long enough that developers lose faith).
--Symbian licensees could fail to produce interesting devices that keep pace with Apples, RIMs, and Palms of the world.
--Android could eat up all the attention of open source developers, leaving Symbian to wither technologically.
--The market might evolve faster than a foundation yoked to handset companies can adjust.

But still the Symbian foundation is worth watching. It has a different set of goals than every other mobile OS company out there, goals that potentially can align more closely with the interests of third party developers. It's still up to Symbian to deliver on that potential, but the company has an opportunity to challenge the mobile market in ways that it couldn't as a traditional company.

-----

Prof. Joel West of San Jose State was also at the Symbian meeting and posted some interesting comments about it. You can read them here.

Full disclosure: My employer, Rubicon Consulting, did a consulting project for Symbian a year ago. None of the analysis conducted in that project was used in this post. We currently have no ongoing, or planned, business relationship with Symbian.

No comments :

Post a Comment