The Symbian Foundation will share the lessons learned by itself and its contributor community, during the first months on its journey towards open software development. We will explore challenges and reflections on community building, open source leadership, collaboration, development and incubation processes as experienced in this ambitious open source endeavour.
Membership model – similar to the Eclipse Foundation The funding members pay our bills and in return get a board seat
170 members so far And growing Funding members shown in the book Device Vendors Operators OEMs
STORY OF THE CREATION OF THE FOUNDATION STORY STARTS before the foundation was created: middle of 2008
ONLY PLANNING WAS POSSIBLE – LEGAL REQUIREMENT. NOTHING COULD BE EXECUTED UNTIL Symbian – Nokia sale was completed I was on one of those teams We talked to a lot of OPEN SOURCE ORGANISATIONS to find out what needed to be done, e.g. Mozilla, Apache, Eclipse
Talk through some of the detail!
The process concluded with a number of VERY LARGE documents All BASED ON THEORY ... BUT: Cannot anticipate everything up-front AND this was a blessing
DATA: TOTAL: 16% of packages have an instance of an IP issue today : 94 individual cases (58 are of them have been addressed since beginning of OCT) 38 still to be resolveUse-cases: Commercial vs. Incompatible open source license RVCT: compiler run-time. Binary R&D. We put a built library into all the kits (can develop, but can’t modify) Adobe & Flash: wrapper/adaptation is open source – library is delivered as binary Generic: codecs = plug-in idea, eventually we are working towards getting full OSS variants Giving us full source: mention company ARICENT of MM codecs – all adaptation work Monotype – fonts MTP ?TapRoot? 3GPP TSY FAT32 Column “How present”One of the challenges we are facing are: Open source business models – some of our members & members of eco-system still establishing their OSS strategies. What to do with their assets? What is core/differentiator and what is an enabler. Do not know whether they would give a differentiator away?Evolving people’s BIZ model!
Resolution mechanisms: Contribution = companies contribute source code to the foundation R&D license = companies provide binaries that can be shipped in development kits Non-core item (NOT NEEDED FOR A DEVICE): a hole that can be filled by commercial 3rd party IP. E.g. QUICK OFFICE SUPPORT
[15 mins]FILLING IP HOLES PRESENTED AN INTERESTING CHALLENGE ....
That IP challenge is part of a bugger challengeChanging an eco-system from a proprietary world to an open source world
UIQ – Examples: Prof Services – IXONOS, Accenture, & others – creating competition Operators – Orange Opportunities for companies who have not been part of the eco-system before: e.g. SUN
All people who knew the technology worked in a proprietary environment beforePERSONAL:More than I do now?Which people in my team do I need to get on my side to be successful? How do I convince them to help me?Never done this beforeSF STAFF:4) Showing to package owners that this can be done – guiding them through – anticipating problems
2) OSS organisation & leader
At this point I wanted to THANK COMPANIES THAT HAVE ALREADY CONTRIBUTED – I wont talk you through the slidesAT THE BEGINNING: I ACTIVELY TRACKED SMALL CONTRIBUTIONS – BUT I HAVE GIVEN UP. I CANT KEEP ON TOP OF THEMINVOLVED IN BIG CONTRIBUTIONS BIG THINGS: like introduction of SMP, a bluetooth v3 stack, a new printer framework
A lot is about building enough momentum: THUS the image of the DOMINOs
Note: Hard to find data as to how long it takes large open source projects to get into “gear”
One of the challenges we have been and are facing are: Open source business models / strategies Our eco-system and members were/are still establishing their OSS strategies What to do with their assets? What is core/differentiator and what is an enabler? Do not know whether they would give a differentiator away?Evolving people’s BIZ model! ... THIS IS A MOUNTAIN WHICH WILL RECURR!!!
COUNTER INTUITIVE: Talk with other OSS orgs and they say “a big OSS organisation removes the need for the community to participate” WE DON’T CODE
[30 mins]SOLVING THE COMMON GOODS PROBLEM = REMOVING BARRIERS TO CONTRIBUTION
Target: at the latest June next year
The COUNCILS were initially CREATED to GOVERN THE PLATFORMBUT: they are becoming ....
Anecdote: When we set up “UI migration” working group We made a call to the entire membership whether they wanted to participate (about 30 people on it, actively working on a problem) Many companies came forward, MANY companies were surprised we asked them to participate
Example: IDEAS: want to avoid embarrassment if they are not executed – e.g. That’s why we do not share CULTURE CHANGE: not sure what can be said CULTURAL COLLISION: different stake-holders in the mobile value chain are moving at different speeds
1: Neutrality is about being confident that a change in strategy, amount of spending, etc. of one party in the community does not affect your investment2: Is about being able to influence the direction of the platform, APIs, see roadmaps, etc. – about INDIVIDUALS and COMPANIESWhere are we?1: Neutrality – Governance & Councils, BUT due to the initial contribution the code is NOT YET OWNED by a DIVERSE community2: Development – When all is moved to EPLed, development will be open to EVERYBODY (not just member companies)
The X6, 5230 and N97 mini are coming soon – expected Q4 2009. All the rest have already shipped.All are S^1 devices = S60 5.0 / Symbian OS v9.4
7<br />Looking to solve problems such as ...<br />What Infrastructure<br />is needed?<br />Bug tracker, Mailing Lists, <br />SCM system, etc.<br />How do different parts ~of the community work together?<br />Collaboration Process<br />What are the different open source roles?<br />Package Owners, Committers, etc.<br />How is the platform governed?<br />Feature & Roadmap, Architecture, UI and Release councils<br />How will the code base be structured?<br />System model and technology domains<br />How are roadmaps created?<br />AND MANY MORE …<br />
8<br /> BUT:<br /> of course there were some issues !<br />Gaps, unintended consequences, somethings did not fit<br />Helped give Symbian a GOOD START<br />
9<br />Preparing the Code3rd party IP and the initial contribution<br />Preparing the Code<br />3rd party IP and the initial contribution<br />After 10 years the codebase contained a big portion of 3rd party IP<br />Which could not be open sourced<br />40 Million Lines of C++<br />
The IP Challenge<br />40 million lines of code had to be checked for IP violations<br />Many false positives<br />Many benign cases (e.g. code copied from a book)<br />It took 6 months to identify all serious IP issues<br />16% of components had an instance of an IP issue<br />94 cases altogether<br />Affected code needed to be removed initially<br />Ultimately replace by open source friendly code<br />
Do I need to change my business model?</li></ul>This takes time<br />Many eco-system companies<br />are concluding this process<br />Early birds are thriving<br />A shock for the eco-system<br />14<br />
Apache average: 14 months</li></li></ul><li>20<br />How can you do this as FAST as possible when you have a HUGE initial contribution?<br />
Lesson: A STRONG FOUNDATION<br />21<br />A strong team of community managers and<br />technology managers<br />An open source organization with some technical capability<br />And prove to community that the model works<br />And show opportunities<br />Can find and work withvendors who may want tocontribute<br />Can initiate projects<br />Because of results the community takesinitiative<br />
23<br />Members Companies<br />Using and contributing<br />SFL Package<br />EPL Package<br />Everybody<br />Using and contributing<br />A beta period towards going open source : <br />learn, fix & de-risk<br />Two stages to true Open Source<br />
Moving to the EPL<br />So far: learning how to EPL, such that ourcommunity can movefaster<br />Build<br />16 out of 134<br />Kernel:<br />Kernel,beagleboard, qemu<br />Security<br />IDE, hostenv, etc.<br />24<br />
Lesson: Two stage process<br />25<br />The Good:<br /><ul><li> Time to staff the foundation
Time to build momentum</li></ul>The Bad:<br /><ul><li> Negative publicity
Negative sentiment from parts of the OSS community</li></li></ul><li>Lesson: LEAVE CHALLENGES<br />26<br />When the foundation started, significant parts of how it was to operate were still to defineCouncil members had some CHALLENGES to resolve<br />builds sense of<br />OWNERSHIP<br />Actually this wasa blessing!<br />Problems (which people care about)<br />Big stake in solving these<br />+<br />=<br />
27<br />Lesson: COMMUNITY REPRESENTATION<br />COUNCILS are becoming mechanisms<br />for<br />COMMUNITY REPRESENTATION<br />Councils:<br /><ul><li> F&R
Release </li></li></ul><li>Lesson: LET EVERYBODY INFLUENCE<br />28<br />SIGs: Special Interest Groups<br />Sense of<br />OWNERSHIP<br />Symbian Idea Sitehttp://ideas.symbian.org/: <br />400 ideas added in a week<br />Working groups (e.g. UI migration working group)<br />BoFs and other interactive parts of our annual conference!<br />UI Brainstorm<br />
Lesson: OPENESS<br />29<br />Open Management<br />Our community like the fact that roadmaps, backlogs, minutes, etc. are publicly available!<br />Worry to share information that can be used to deduce trade secrets<br />Culture Change & Cultural Collision<br />Many are experimenting with IDEAS – being open can lead to embarrassment<br />Be as open as you can<br />BUT <br />do not force your community to be!<br />
Lesson: INCENTIVE STRUCTURE<br /><ul><li> Can I see roadmaps, influence the direction of APIs and the platform?
Some data to close...<br />>60% of these will be based on open source operating systems<br />Symbian is projected to be the most used OS in 2012 & 2014<br />In 2012 - 2014 there will be 372 - 525 million smart phones.<br />Open Source in Mobile is here to stay !!!<br /><ul><li>Juniper Research