TransferSummit Report

A report from the OSS Watch Transfer Summit, held at Keble College, Oxford, 24-25 June 1010, by Sam Jordison.

OSS has come a long way in recent years. Even the most secretive and commercially sensitive organisations like mobile phone companies and NASA have started to understand its benefits. It is now widely used in the commercial and higher education sector and it touches the lives of millions if not billions around the world. As a number of delegates pointed out during the conference, the old cliche that 'you use Linux 10 or 15 times a day and don't realise it' could just as easily be transposed to a wide number of OSS projects.

At TransferSummit, this growing acceptance of OSS – and the principle of open innovation – was most clearly demonstrated by an absence. There was hardly any debate about whether OSS is a good way forward. That was pretty much taken as understood. Yes, there were involved discussions about its benefits and how best to make it work. Yes, most speakers and delegates were keen to explain how and why it has helped them. Yes, there was also informed debate about the limits of open innovation and acknowledgment that it might not be the best fit for all people in all circumstances. But this debate was mainly about how to make it work best, not whether it should be used at all.

That's not to say that there aren't still arguments to be had and people to be convinced. It's also probably true that there would be little point in going into too much basic detail at a conference with such a well-informed set of delegates. But the fact that so many high-achieving business-people, developers and academic facilitators are on roughly the same page when it comes to the benefits of OSS, could perhaps be taken as a demonstration of how far OSS has come in recent years. As Steven Pemberton said in his keynote speech, 'we are through the first stage' in getting open innovation technology accepted and now the main task is to make it better.

That said, one of the many nuggets of background passed on about open innovation during the course of the two days is that it isn't really such a recent concept. In his talk about the Codeplex Foundation (not to be confused with codeplex.com), Stephen Walli provided a useful reminder of how this kind of open innovation has worked in the past. 'I remember that in the late 70s and early 80s there was DECUS where you got a tape packed with software that everyone was able to share,' he said. 'Even in the 1950s (and running up to the 1980s) there were IBM SHARE conferences. In the 1940s there was the Institute for Advanced Study at Princeton University.' He also noted that at IAS, the world's first stored-program computer was running 24 hours a day and all sorts of people were able to use it and shared software. 'The bottom line is that we have shared software since we have written software,' he said.

The economic benefits of open innovation: value for money

Given the economic climate and budget cuts faced by projects across the education, public and commercial sectors, it was natural that a great deal of thought should be given to the cost benefits of open innovation and how to generate growth and save money across the board. The most clear economic benefit of open innovation to emerge over the course of the conference was the simple fact that it works. It delivers results - and does so in a cost-effective way.

One of the points that was often made during the two days was that although open source doesn't necessarily mean 'free', it does generally mean cheaper, and often more effective. It enables access to reservoirs of talent that couldn't otherwise be tapped and allows a kind of economy of scale to build up. Steven Pemberton gave the famous example of the free-at-source Wikipedia compared to the old Encyclopaedia Britannica, and how 'little things' can join together 'to make a big thing better and better'.

Mark Taylor (CEO of the Sirius Corporation), speaking about FOSS business models, emphasised that OSS brings costs down. He noted that just as the cost of hardware like PCs and laptops has plummeted, been commodotised and become a marginal expense in production, so open source is making software ever cheaper. 'It brings the cost towards zero,' he said. In a talk where there was much to-and-fro between the floor and active engagement about OSS business models, Mark was also asked about how OSS business models compare to the Cloud. His reply was interesting: 'I personally think, long term, very little. Economics will inexorably push all Cloud providers to base their offerings entirely upon open source. That's to say, the Cloud will be driven by open source technologies.'

Mark was also asked to engage in a bit of futurology about whether there will be multi-billion dollar open source companies in the future - and whether the fact that there aren't brings the model's economic validity into question. Again, his answer was striking: 'Do we need multi-billion dollar IT giant companies? Do we want another Too Big To Fail situation? Efficient markets depend on multiple producers. And it's a mark of OSS' success that there are many small companies.' Finally, he was asked to address the question of whether open source companies destroy economic value. 'No!' he insisted. 'Nonsense! In any organisation, technology infrastructure is a cost, and the economic imperative is to bring costs down. To the monopolistic or oligopolistic proprietary technology provider, open source is destructive of their artificially inflated pricing, but to the rest of the world open source directly reduces cost and improves the bottom line. To claim it 'destroys economic value' is literally nonsense, as reducing input costs is considered to create economic value in all valid economic models!'

Community

Steven Pemberton's question of how to make open innovation better received a great deal of attention throughout the two days. Andrew Katz, a partner at Moorcrofts LLP (in a talk entitled: Open licensing: cars, cartography, content and cola) gave a concise list of what he sees as the important elements in a successful OSS project:

  1. A speedy development cycle
  2. Low friction/barrier to entry
  3. Low upfront investment
  4. Low cost of entry
  5. Minimal commitment required
  6. Easy participant communication

Many other speakers, meanwhile, focused on the importance of building active communities. Bertrand Delacretaz of Day Software explained in his talk that open source will always bring in 'brighter people than your company could afford', a point that was reinforced by Phil Andrews of Red Hat when explaining the economics of OSS development. He said that if SourceForge had had to pay the 50,000 people who had worked 10 hours a week on Research & Development in their community, the annual wages bill would have come to £4.5 billion.

Phil also pointed out that the number of coding errors per 1,000 LOC generated within an open source community is much smaller than those produced by a closed company. Paul Walk of UKOLN summed up this idea well when he said: 'A well-connected community of developers is greater than the sum of its parts.'

It was repeatedly stressed that the people doing that work should themselves be valued accordingly. Paul Walk also emphasised the simple importance of giving recognition to the work that developers do and helping them better explain their ideas and projects to the world away from the keyboard. Gianugo Rabellino, former CEO of Sourcesense, told us that if we recognise communities as places 'where individuals come together to reap a reward … we come to understand why collaboration works so well - and why people are producing open source software'.

Scott Wilson, assistant director of CETIS, also made the point that 'inclusion and openness depend on collective responsibility'. Open source works when people feel they have a stake in it.

Or, as Mark Johnson, the web developer at Taunton's College put it, when there is 'a thrill of recognition of one's achievements'. Delegates were frequently reminded of the importance of letting people know that they appreciate that they are doing a great job.

In his keynote speech on day two, Roland Harwood also gave powerful arguments about the importance of community - and how effective exchanging ideas with different companies can be - a particularly striking example being how the McLaren Formula 1 team were able to help improve processes in A&E at Great Ormond Street Hospital. To illustrate his point, he quoted the proverb: 'If you want to go fast, go alone, if you want to go far, go together.'

Forking and the importance of keeping it in the family

A number of talks touching on the mobile phone sector provided concrete examples about the benefits of community - and the importance of staying within developed communities. Matthias Stuermer, project manager at Liip AG, brought up the example of Nokia, who have discovered that revealing their source code worked to increase 'their knowledge, their reputation and their presence' and gave them an effective soapdish in which to test out their ideas.

Meanwhile, in his talk about the economics of innovation in mobile technologies, Andrew Savory, open source manager for the Limo Foundation, gave another striking example of the advantages open innovation can bring. He pointed to the smartphone market, where consumers are demanding ever more features for ever less money, meaning that we have not reached a point where value creation is less than the amount that companies are having to invest. So they've now started to look more seriously at open source software and have discovered that it brings not only reduced costs to the acquisition of software, but also reduced costs of access to innovation and - crucially - reduced costs of software ownership, since there is a greatly reduced maintenance burden for true OSS. HTC has leapfrogged the competition thanks to its use of open innovation.

Clearly, there is going to have to be a big cultural shift among companies who are generally secretive about their innovations, and who are unused to the meritocracy that exists within OSS development, but evidence that OSS is the way forward is beginning to stack up. Andrew also highlighted a paper - the 'Mobile Open Source Economic Analysis' white paper - showing that it's even cheaper for companies to merge early and contribute early to OSS development streams, rather than 'forking' off and trying to keep their own innovations with regard to the software to themselves for as long as possible.

Similarly, Matthew Langham from Indiginox (in a talk entitled Open source in commercial institutions: surviving the 10-year itch) discussed Vodafone's early attempts to use open source software on the quiet. He said that they tried to fix problems themselves - and so ran into ever-deeper trouble. So Matthew persuaded them to open up and embrace the community and soon everything became much 'easier' for the company. The companies are starting to discover that - as Simon Phipps, a director of the Open Source Initiative, put it - that giving code back is the best way to avoid becoming 'a slave to the remaking'.

Usability

One of the hottest topics at the conference related to usability. In his keynote, Steven Pemberton highlighted it as 'a big challenge to open source' and an area in which progress was 'vital'. He stressed that 'being usable doesn't mean simple. A violin is usable but not simple.'

Other useful definitions were also provided during the course of the conference. Paul Fremantle, CTO atWSO2, in a talk entitled 'How to build innovation using the open source model', said: 'Usability for me is not just about ease for the user - but about how well companies can consume the product.'

Less easily answered was the question of how to ensure that OSS could be made 'usable'. When questioned from the floor, Steven Pemberton said that we need 'a paradigm' for open source usability. He said that there are two options available at the moment:

  1. Copy existing programs (which is not open source-like)
  2. Do it with money (which is not scalable)

'We need to find a third option,' he said, 'to find a way to get programmers to "scratch other people's itches", using the terminology of the Cathedral and the Bazaar paper. I don't know how to achieve that.'

Some speakers from the floor claimed that this has already been achieved, highlighting Firefox as a 'triumph of usability' and one created by open source developers. Also pointing the way forward, Andrew Fennell, a practice leader at Strategic Growth and Innovation, PERA, in his talk about 'Open innovation involving business and academic teams', brought up the striking example of Nintendo and their development of the Wii. 'They realised that they had to create extended usability,' he said. And so 'they focussed away from the traditional value concepts and looked at people far away from traditional audience'.

Instead of asking the people who already owned consoles what they wanted (generally bigger, faster consoles), Nintendo looked to very different audiences of 7-70 year-olds and realised that they had very different desires. They gave them what they wanted (briefly, an easy-to use - and above all social - gaming tool) and the rest is console history: 'Nintendo created a new value curve.' He suggested that those in OSS could think along similar lines. 'If you ask the same people the same questions you get the same answers,' he said. Ask new people something new and a different kind of solution may become more apparent.

Licences and other challenges

One of the other important challenges for open source developers that the conference highlighted was what to do about licences. Simon Phipps even asked delegates to re-imagine OSS so that 'in 2010, licenses are done. No more open source licenses are needed.'

Others thought that licences are vital to protect both developers and users - although which one works best was generally said to be contingent upon circumstances. Martin Michlmayr, in his talk 'The state of open source licensing and how to improve it'. Mark Taylor, CEO of the Sirius Corporation, in 'FOSS business models', ably demonstrated the tangled wood of licences and legal complexity faced by anyone hoping to launch an open source project - and provided a good route through.

Simon Phipps also worried that although the benefits of open source are clear for business, once businesses (whom he likened to reptiles, since they have no ethics) become involved with OSS, the benefits to developers aren't always clear. He said he fears that unscrupulous business practice can often destroy what makes open source so effective. He warned that the nature of open source is being undermined by companies taking the benefits of others' innovation without giving anything up of their own - such companies declare themselves to be ‘open core’, which he sees as a euphemism for traditional closed software. The Open Source Initiative was set up to combat such misuse.

Linking business with academia and getting OSS out there

Aside from the issues noted above, the single biggest nut to crack was the question of how to link developers with businesses, how to transfer the skills and technology developed by OSS programmers to all sectors and how to release the value and innovations that universities generate. 'Instead of having these things locked away in an ivory tower, we want to recognise that there are potential partners that can use them out there,' said Ross Gardler in his speech welcoming delegates to the conference.

Gill Rysiecki gave excellent practical examples of how the Technology Strategy Board initiative, Knowledge Transfer Partnerships, can help academic developers reach out to the business community and vica versa, while a number of speakers dwelt on how to develop the right kind of mindset to get innovation out into the wider world.

To a large degree, making this leap depends upon the kind of forward thinking that Andrew Fennell highlighted in relation to the Wii. Ross Gardler also brought up the example of the Sheldonian theatre in Oxford, noting that in giving the young Christopher Wren the opportunity to design such a high-tech building (by 17th-century standards), the University demonstrated a healthy degree of openness and willingness to transfer technology.

Roland Harwood also provided a further example of McLaren's ability to transfer their skills in the predictive software they have helped develop for airport traffic controllers based on the system they use in F1. A 'clever conversation' from a jam session became a potentially multi-billion-dollar spinner, thanks to the willingness of their programmers to pass on their skills, and work as a team, under the radar, to drive an idea to reality. He also provided six tips for starting - and finishing - a new project:

  1. Start new conversations
  2. Ask more engaging questions
  3. Address unmet needs
  4. Make it easy to contribute
  5. Share risk and reward
  6. Build mutual value

'There is still an idea of the mad, lone genius, but actually far more is achieved in a group,' he noted. 'Innovation is about shifting from a mode of invention and creativity to one more akin to being a detective; to finding opportunities or competition that appear outside of your core domain of expertise.'

Conclusions

To a large degree, TransferSummit highlighted the success of open innovation. The presence of so many powerful speakers and the sheer breadth of experience on display speaks volumes about how far things have come. Naturally there are still issues and challenges that the conference set about addressing. There will be many further debates to be had about licensing, monetisation and ensuring that those in OSS communities feel sufficiently rewarded.

Nevertheless, as Roland Harwood said in his keynote: 'The genie is out of the bottle. We now have to get our mindsets right.' Given a strong, multi-skilled community, there is very little to which an open source project cannot turn its hand.

Further reading

OSSWatchWiki: TransferSummitReport (last edited 2013-04-15 13:56:24 by localhost)

Creative Commons License
The content of this wiki is licensed under the Creative Commons Attribution-ShareAlike 2.0 England & Wales Licence.

OSS Watch is funded by the Joint Information Systems Committee (JISC) and is situated within the Research Technologies Service (RTS) of the University of Oxford.