OSS Watch team blog

This guest post was written by Sam Jordison, who also wrote the live blog at Open Source Junction 2.

Following on from the platform built at Open Source Junction 1, this two day conference at Wolfson College Oxford developed the theme of industry-academia collaboration on open source mobile technologies. The focus this time was on  context-aware mobile technology.

So what is context-aware mobile technology? Over the course of the two days, there arose a number of interesting definitions taking the notion of context beyond the simple idea of location based services and into the lives of end-users, taking into account such things as their emotional state, habits, patterns of social interaction and the way they use their time.

In a talk entitled ‘Context Aware Applications: Industry Landscape And Commercial Opportunities’, Nick Allott, the founder of Nquiring Minds Ltd, said context was ‘all about probabilities’ and relationships, good examples being Amazon’s suggestions that ‘if you like X then you’ll like Y’ or the idea that if your friend installs a security system, you are more likely to trust it. Julian Harty, ‘tester at large at eBay’, in his talk ‘Smartphones In Context’, asked delegates to think in terms of interaction with the outside world. ‘Do you know how many sensors your smartphone has?’ he asked. The answer he said was almost certainly likely to be more than 10; including a light sensor, a sound sensor, a compass, rotation detection, accelerometers, GPS. All of these sensors work with the context in which they are placed – and can be used to create new contexts. Elsewhere, the idea of context was touched on in talks also encompassing business and academic integration, licensing, best practice in running open source projects, financial issues, dealing with huge amounts of data  – and ensuring that  data sources are reliable.

Context, it seems, is a broad issue – but the over-riding theme was the importance of engagement; whether that be with other programmers and contributors on open source projects through mailing lists, or between developers and the wider public. Such engagements have the power to change the world – and the way we see it. More particularly, the aim of the conference was to foster engagement between industry and academic people, to help them understand each other’s interests in context-aware mobile technologies.

One of the main ways this latter aim was encouraged was in show-casing a diverse range of projects and ideas from both industry and academic speakers through a wide range of presentations. Indeed, the very first presentation from Gabriel Hanganu tackled the idea head on. He acknowledged that there are perceived barriers between the worlds of academia and business, especially relating to the different drivers in each sphere. Profit and practical production motivate business. Ideas, research and journal production push academics. But Gabriel pointed out that academics are not as slack when it comes to entrepreneurial thinking as is often supposed -  while business can really profit from academic thinking. There is plenty of common ground – and when it comes to software development, the practices and procedures relating to open source can help bridge gaps. (More on that later.)

Roland Harwood, co-founder of 100% Open built on this idea of the usefulness of partnerships with his talk about open innovation collaboration. He highlighted a large number of examples of successful collaborations, and showed how even traditionally less open companies like Lego have benefited by enabling a broad community to use their code, and unleash their creativity. Mindstorms, thanks to its fan built ideas, has now become Lego’s best selling product.

As well as generating nifty new revenue streams, new technology can also provide a fascinating new insight into the way the world works. Steven Gray from the UCL Centre for Advanced Spatial Analysis gave a fascinating run through his work on a number of game-changing projects, the most famous of which is Tweet-O-Meter, a program that uses geo-location data from Twitter to show when and where people are sending tweets. The data can be aggregated around maps to show interesting things. So, for instance, in London they can trace where there are roadworks and traffic jams because people are tweeting about them. They have also spotted that people tweet a lot on railway lines and at Heathrow airport as they take off and land and that parks are particularly free of activity during winter. They can see cities waking up and sleeping. CNN used the meter during the Japanese tsunami to show how people reacted to the news (since lots of people in Tokyo used their mobiles as landlines were down). They could also see the news spread to San Francisco.

Another clear demonstration of how mobile technology can be used to trace events in the physical world came in the form of the Nature Locator, described by Mike Jones  from The Institute of Learning and Research Technology at the University of Bristol. This emerged from a JISC-funded project that facilitated development of an Android and iOS apps to support citizen science. The app allows members of the public to submit photographic and geo-location evidence of leaf damage by a non-UK-native moth to the Conker Tree Science project – and has generated masses of important and useful scientific data.

The useful practical applications for such technology were clearly demonstrated by Serge Pawlowicz from the Centre of Geospatial Science in the University of Nottingham and his talk on a Particpatory HealthGIS that uses geospatial data and public participation surveys to help in all kinds of health research, for instance, tracing the sources of viruses. ‘It works!’ he said.

Ben Butchart, an experienced software engineer from EDINA (a JISC-funded national academic data centre providing online data services to academic institutions) continued this theme, explaining that the uses of HTML5 caching technology has opened up all kinds of new possibilities for developing useful technologies for geo-dependent projects (such as geology projects) even in isolated locations where signals are unreliable. At the other end of the spectrum, Tim Fernando gave an overview of how useful context-aware technology can be to residents of Oxford, and the success of Mobile Oxford, a campus-wide mobile service providing information on everything from bus times to library book availability to gigs.

Such projects can also be extended to provide useful information to and further engage with the wider public (end-users) by inviting them to contribute to location-based media platforms. So said Mick Lockwood from Salford University who demonstrated this in his talk about Maxamundo. He explained that using OpenStreetMap and a range of open source software, and getting user contributions helped him become ‘able to fulfil a dream’ even if he was just an ‘amateur hobbyist’. He’s now built up a detailed overview of Manchester and its attractions (a great many of which appear to be pubs) and the Maxamundo platform even becomes the subject of a sociological project. This latter project was run by Yuwei Lin, also from Salford University, who gave a fascinating account of the way it worked against open source development methodologies after Mick had finished speaking, explaining that Maxamundo has started to fulfil many of the functions of social networking sites, contextualise city lives, and re-order personal stories on a map.

The conference even explored the imaginary plane as well as showing so much about mapping the real world. Philipp Breuss-Schneeweis described how Wikitude, an augmented reality app that overlays virtual images and information over the real world (as revealed in, for instance, smartphone camera viewers), has the potential to alter the way we interact with, well, everything. Current uses include navigation devices for driving and 3D gaming, but the potential is limited only by the imagination of the huge community that can be harnessed to create new ‘worlds’ to overlay the context provided by the phone.

Elsewhere, one of the most impressive examples of context-aware engagement came from Samuel Carlisle and his colleague Matt Gaffen and their talk about Sukey. Sukey, we were told, exists to keep demonstrators safe and mobile during protests. It was created in 2010 in the aftermath of the occupation of UCL and most particularly in response to the police tactics used during that winter’s student protests, whereby protestors were ‘kettled’ for long periods of time and had their freedom of movement restricted. (The name comes from the nursery song, ‘Polly put the kettle on, Sukey take it off again’.) Sam said that they started putting out maps of protests that started giving them upwards of 60,000 hits in just a few hours – which inspired them to produce a proper app.

Using information crowd-sourced from the app, alongside information in Twitter streams, Sukey attempts to put out the most relevant and useful material surrounding demonstrations. So, for instance, they show where the police are forming kettles, where streets are blocked and also provide compass directions designed to help people on the ground get out of trouble. It’s run by volunteers in their spare time and it’s not for profit, so the use of community engagement was vital in making it work… And work it did; providing helpful information to thousands of protestors, and even acting as a conduit for advice and information from the police themselves. The application also received widespread media coverage and helped inform the wider ongoing debate about police tactics. A clear demonstration of just how quickly and effectively free-at-point-of-use software can change our society.

Many of these projects are already successful and those in their infancy are already demonstrating real potential, but at Open Source Junction 2 there was no shying away from the challenges they face. Speakers made it clear that taking contributions from large numbers of people – whether they be citizen scientists or developers sharing in an open source project – is not always easy. The team behind Sukey, for instance, have to dedicate a lot of energy to combating spam and false information. The Nature Locator has had to deal with a lot of incorrect data. Checks and balances are going to play an increasingly important role, it seems. W3C fellow Dave Raggett, meanwhile, highlighted the fact that ’simplicity is hard’, explaining the need for good communication paths based on trust relationships, an easier way to manage security and logins across a number of platforms than we have at the moment and suggesting, cleverly, that it would be better if users could check the credentials of the website – as well as the website checking them out and for there to be support for pseudonymous identities that reflect the real world. No easy tasks – although we can at least take solace from the knowledge that he and others are working on them.

It was also emphasised that open innovation isn’t a panacea in every case. It can offer a fantastic way to cut costs, save time and maximise the potential of developers. Ross Gardler outlined a number of open source business models that can help you make sure ‘your company doesn’t go bust because it’s spending so much maintaining software’, ensure long term sustainability for projects and ideas, and open up many opportunities for commodotisation. But it was also emphasised that there are all kinds of procedures to follow and tough decisions to make to ensure the smooth running of projects – and that they remain sustainable in the long term. Just how careful projects have to be to follow the correct processes was clearly demonstrated in talks from Sander van der Waal concerning best practice relating to open source projects (emphasising the need to have a good management structure in place, transparent updates and progress reports via mailing lists) and a look into the complications surrounding patents, licences, Intellectual Property Rights and European law from Rowan Wilson.

In spite of such challenges, another clear advantage of open source development became clear over the course of the two days – how much it can help in academic and business collaborations, just as Gabriel had hinted in his opening talk. Camille Baldock from Softwire explained to me that while there is plenty of desire in the business world to tap into the academic knowledge base and expertise, such collaborations remain rare, thanks to perceptions about the different cultures and priorities found in the two ‘tribes’. But, as Gabriel  repeatedly stressed during the day, in software projects, developers can offer an alternative common ground in the form of the clearly defined work practices associated with open development.

Another thing that wasn’t in doubt at the end of the conference was how much participants did have to give each other. The ’speed dating’ session invited delegates to write down three things they could offer a potential partner  – and three things they wanted in return. Ross Gardler got the ball rolling. He offered: a wide network from working in real open source, guidance on sustainable models, and practical experience on making things happen. He said he wanted: real projects before they go to market, innovation companies to work with, and introductions to more useful people. There followed a fast and furious twenty minutes with a huge array of services offered and requested. Offers included, contacts and development from the London Mobile Developer community, an open invite to hackdays – including organising them, a strong network for industrial partners in the telecoms, students and resources, expertise in security, 3D Printing knowledge and (let’s not forget!) another conference in September – TransferSummit, co-organised by OSS Watch. Wants included, ideas and products to sell, the promotion of open source, partners, funding opportunities, opportunities for future consortia.

These connections were further aided by the format of the event, with numerous other such interactive sessions built in. As well as the ‘offers and wants’ session, delegates were invited to ’self pitch’, giving a 45 second overview of their mobile app’s interests and invite feedback. They also took part in a ‘3 minute joint venture’  session in which they were invited to pair up with someone they hadn’t yet talked to and spend three minutes coming up with an idea for a joint venture. (Highlights included a mashup of eBay and location-based sales, using targeted marketing on people walking down the street, and AID – Am I Dying – a mix of a virtual patient project in St George’s with an intelligent clothing company… ). There was also plenty of space for informal conversations over breaks, drinks and dinner saw many start to form those connections.

In short, there were many successful interactions building from the platform established at the first Open Source Junction. Even the fact that the attempt to put geo-location technology into practice on a country walk resulted in all delegates getting soaked in a rain storm and two chased by a bull didn’t dampen spirits. People were exchanging details, forming links throughout the two days and I’m told many are discussing opportunities in more detail now. There was the feeling that a community was forming. Let’s hope it continues to grow and prosper, and watch out for the next OSJ3 planned for November.