From MeeGo to Tizen: the making of another software bubble

[Just a short 1.5 years from MeeGo's birth, Intel dumps it to shift focus to a new platform, Tizen, in partnership with Samsung. Guest author Dave Neary discusses the underpinnings of Tizen and why both MeeGo and Tizen are software bubbles].

Eight months after Nokia embarrassed Intel by withdrawing support for the MeeGo project, Intel has followed suit. On 27th September, Intel and Samsung announced the birth of a new mobile platform called Tizen. After only 19 months, MeeGo has been left parentless, and appears to be on life support. Tizen is, in fact, a successor of the Samsung Linux Platform, a reference platform of the LiMo operator consortium, with some components taken from the MeeGo stack.

Given that LiMo and MeeGo have both failed to set the mobile computing world alight, and Android has a four year head start, can we expect better things from their offspring? What has changed with this announcement? Is this Intel’s last chance to have a stake in a credible smartphone platform? And what  should Samsung, Intel and the Linux Foundation do to give their new platform a fighting chance at success?

The Birth of Tizen

Last year, when reviewing the progress which MeeGo had made in its first few months, we reserved judgement on the project, on the grounds that it was “too early to be able to tell how the final product will compare to iOS or Android”, but we noted that there had been some growing pains between Nokia and Intel.

Those growing pains stretched to breaking point earlier this year, when Nokia finally gave up on MeeGo and turned to Windows Phone to revitalise its smartphone products. Intel was left looking for a heavyweight consumer device partner to come in and lend credibility to their claim that MeeGo was no longer a one-man show. Rumours that LG would be joining the project failed to materialise. Finally, Intel ran out of patience, and partnered with Samsung on a new platform, Tizen, to be based on SLP (Samsung Linux Platform), a platform which Samsung have previously provided to the LiMo Foundation to be used as a reference plaform for its members.

While the move has obviously been in the planning for months, Samsung were perhaps encouraged to partner with Intel on the back of the news that Google has acquired Motorola Mobility in August – a view supported by their recent settlement of an Android-related patent dispute with Microsoft. In addition, as LiMo members, most notably Vodafone also ran out of patience, SLP was left as a platform without a home.

MeeGo on Life Support

How does MeeGo fit into the big picture now? High profile participants like GENIVI, China Mobile, Asus and Acer have committed to shipping MeeGo devices. Will they be based on the unreleased MeeGo 1.3, or the previous 1.2 release? Or will these companies move en mass to Tizen?

Given the lack of reaction from partners like GENIVI, we suspect that the Tizen announcement caught these vendors unawares. Jerermiah Foster, a community manager working for one GENIVI member, informed me that his company would reuse MeeGo 1.2 in the short term, and while Tizen looked interesting, there were no current plans to move development to the platform. He also confirmed that he found out about Tizen through the project announcement, and not before.

In spite of vendors withdrawing their support, part of the community is banding together to salvage their work. After Nokia pulled out of MeeGo, community developers working on the MeeGo Handset UX banded together to continue work (with several Nokia engineers) in the MeeGo Handset Community Edition, aiming to provide MeeGo support for the Nokia N900, N950 and N9 devices. In spite of the Intel announcement of Tizen, these developers have vowed to continue the development of MeeGo on ARM, and released the MeeGo Handset Community Edition 1.3 at the end of September. The current plan proposed by these developers is to create a lightweight core distribution based on Qt, under the brandname “Mer” (“MeeGo Rebooted”), on which vendors can build custom user interfaces. The MeeGo Handset Community Edition will be the first consumer of Mer’s core operating system.

The MeeGo community mailing lists are full of developers wondering where they stand now. The announcement suggests that no software will be released from the Tizen project for another 6 months. According to Joel Clark, MeeGo IVI Program Manager, the MeeGo 1.3 release has been shelved, and only incremental updates to the previous 1.2 release can be expected until then. While the MeeGo community certainly has some enthusiastic community supporters, it is unlikely that any major vendors will adopt the community-supported Mer.

Ironically, the move away from MeeGo comes at a time of potential wins for the project. Nokia’s MeeGo-based N9 is finally shipping, and getting rave reviews. And continued demand for netbooks has fueled the launch of several MeeGo based netbook and tablet products, including the Asus EeePC X101, the Acer Iconia M500 and other devices from Samsung, Lenovo and Fujitsu.

Perhaps Intel ran out of patience just as the project was about to take off.

Tizen = SLP, with a pinch of MeeGo

Technically, Tizen is a successor of the Samsung Linux Platform, a reference platform of the LiMo operator consortium, with some components taken from the MeeGo stack. The project governance and infrastructure, however, will look a lot like MeeGo. According to Imad Sousou, the director of Intel’s Open Source Technology Center, and head of the MeeGo project: “in the new project, a lot of things will be the same as they were in the MeeGo project”.

We also know is that the primary APIs for 3rd party developers are targeting HTML5 and WAC environments. WAC stands for Wholesale Applications Community, a set of APIs for building and delivering rich HTML5 applications, based on APIs from JIL (Joint Innovation Labs) and BONDI (a platform specified by the now-defunct Open Mobile Terminal Platform, OMTP). The Enlightenment Foundation Libraries (EFL), are also set to be a key part of the platform. We can infer two things from this: Qt will be taking a back seat in Tizen, if it is part of the platform at all, and it appears that SLP will be the basis of the Tizen platform.

One thing which has not changed from MeeGo is the wide range of participants being targeted by the project. At the moment, the target audience can best be summarised as “everyone”. Tizen is aimed at platform developers, integrators, vendors, application developers, and mobile enthusiasts. That’s a very wide range of target audiences, each with different needs and expectations. Not knowing your target customer is a surefire way to throw money down the drain.

Challenges, challenges, challenges

Tizen’s main difficulties at this point can be broken into three groups.

First, there will inevitably be teething problems between the project founders. The fact that Samsung have not yet mentioned Tizen in any press releases or announcements, and the lack of new information coming from Intel representatives since the launch announcement, suggests that there may be some communication issues to be worked out in the relationship. In fact, at this point it looks like the active partners have not yet agreed on what will and will not go into the platform. Intel and Samsung will have to work hard to overcome the cultural dissonance which is inevitable given the very different corporate DNA.

On top of this, unless something changes soon, there could be a major mismatch between the reality of working with Tizen and the public positioning of the project. The project isn’t yet open for business, and when it is, it will only be useful for a small subset of its target market. If it were a new project, they might get away with it. But with the legacy of MeeGo, Moblin and Maemo, disappointing early adopters could be a very dangerous thing to do for Intel and Samsung. Getting the project governance and community dynamics right from the start is vital to learning from the mistakes of MeeGo and Moblin.

Beyond the community, there are question-marks over Tizen’s potential to make an impact in the industry. Google’s purchase of Motorola Mobility, not just a patent portfolio play, has created a disturbance in the force around the Android universe. Samsung does not want to find itself competing with Google at the same as they are dependent on them for their smartphone platform. This creates an opportunity for Tizen which it is too immature to exploit. For third party developers, concentrating on HTML5 is great. But will there be a demand for a native API also? And if so, will Tizen be capable of providing the kind of unified developer experience you get on iOS or Android?

It will be interesting to see if Intel and Samsung manage to get substantial support from other ARM vendors. As long as Intel are seen as the main custodians of the project, that seems unlikely. It will also be interesting to see the effect which Nokia’s first Windows Phone based devices, due to be announced at the end of October, will have on the project.

The main challenge for the Tizen partners will be getting devices to market. The key constituency for the change, vendors who were committed to MeeGo before, appear to have been neglected during the announcement. Intel and Samsung need vendors to adapt the platform to sell more chips, to give breadth to the ecosystem around the project, and to give credibility in the industry that this is not a party of two.

The Long Road Ahead

To succeed and make a space for itself in the mobile ecosystem, execution will need to be flawless on Tizen. If the internal bickering which dogged MeeGo rears its head again, if the initial release of the platform does not meet vendor and community expectations in terms of functionality and quality, or if there is an 18 month wait for well integrated finished products running Tizen, then the project may not have a second chance to make good.

Tizen seems set to be another victim of misaligned incentives across several industry partners. Samsung is bringing SLP to the “standards” table simply to find a new home for it, now that LiMo is winding down. Intel is seeking another marriage of convenience, trying to tempt a major OEM to ship significant x86 chip volumes.

- Dave

[Dave Neary is a regular columnist at VisionMobile writing on how companies can work more effectively with open source community projects. Dave is the founder of Neary Consulting and  has also been an active member of the GIMP, GNOME, OpenWengo, Maemo and MeeGo communities, with over 10 years of experience in open source community issues. He can be contacted at dave (at) neary-consulting (dot) com]

•