 Thank you. I'd like to start off with a little confession. I'm completely unqualified to give this talk. I work as a technical writer in Google's Zurich office, and I'm the executive VP at the Apache Software Foundation. I have absolutely no experience in crisis management or disaster relief. But what I do have is a passion for open source software and a passion for saving the world. When the CFP opened for this mini-con, I proposed the talk because I wanted to share those passions with these attendees. Many of the people in this room are people I've followed and admired for many years, people who've encouraged me and people who've inspired me. And these people have made me believe that I can make a difference in the world. So I was looking forward to meeting those people in person and telling them a little bit about the things I care about. I expected to be standing in front of a room with a few of my friends speaking about something I'm not really an expert in, and that was going to be fine. What I didn't expect was to be sitting in my apartment in Zurich two weeks ago with my talk written in my bags pretty much packed, looking at Google Maps imagery and terrain maps for this building and trying to work out how far the water could rise before my adopted Australian grandparents were getting their feet wet. Happily, I got a phone call late that night from my mother to say that they'd evacuated to Fiona's house and all was well. By the time I made it to Melbourne, two of the lads had made it back to the unit in kayaks at low tide. The water had made it up to the front door and stopped. So they were very lucky. It was several days before it occurred to me that I was now planning to speak about open-source crisis management and disaster relief software at an open-source conference in Queensland. It was, I will give you a disaster that from an information strategy perspective was very successfully managed, but nonetheless. I looked up the Queensland flood map again. I clicked through the ABC emergency site. Did a quick search. Ushahidi, awesome. Okay, this is software I know about. This is software that I'm prepared to talk about. So I was delighted to see that it was software I knew, open-source software, and it was already, more or less, in my slides. But I'll admit I was and remain a little bit daunted about doing this talk right here and now. So for those of you who came to Linus Confey to hear experts speaking on topics they know inside out, I suggest you check out Brianna's talk or come back in half an hour for Arianna's talk. If you'd like to stay and hear what I know, please know that we have plenty of time and I'd love your participation. If you have something to add about the projects I'm talking about, particularly if you know anything about the specific deployments here in Queensland, I'd love you to pipe up when we get to that. And finally, if the last thing you want to hear about this week is disaster management, I totally understand, and I just ask you to be patient with me for a few minutes. Open-source software is saving the world in more ways than you might know about, and I will be talking about more than just crisis relief, I promise. So with all of that out of the way, and now that you know who I'm not, let me give you a little bit of background on who I am and how I ended up standing up here in front of you today. I suppose first of all we should start with my family. My mother is a midwife. She works in the community where I grew up and her job is all about bringing life into the world. My father, for his part, runs one of the largest and busiest emergency departments in Ireland, serving a population with major challenges, poor, elderly, immigrants, a lot of people with a lot of social problems. And his job is about saving lives, quite literally on a daily basis. So that was kind of a high bar to match. But it did mean that as I was growing up, I was always pretty clear on what I wanted to do. It was only one choice, really. I wanted a job that would let me save lives. Sadly, I'm actually really squeamish. So Metsu was not going to work out. So I gave up on that idea and I did a degree in computational linguistics, which is all about language and computers and communication and how all of these things connect together. And somewhere along the way I started matching up the technology and the communication and I got involved in documenting the Apache web server. And over the years, I've come to realize that there are many ways of saving the world. Technology, in general, and open source software, in particular, offers so many possibilities. So since I have you all here, I'm going to take some time to talk to you about just a few of the open source projects that I'm particularly fond of that are dedicated to saving lives and making the world a better place. How many of you remember where you were on St. Stephen's Day in 2004? It's a while ago now, but for a quarter of a million people, that was, sadly, their last day. About a quarter of a million people lost their lives after an earthquake off Sumatra triggered a tsunami that just devastated coastal communities across the Indian Ocean, 14 different countries. And perhaps worse than the immediate death toll were the about two million people missing, injured, displaced, no longer had anywhere to live. And we're talking, by and large, about the poorest of the poor here. Per capita, by far the worst hit country was Sri Lanka, which is a truly beautiful island nation, famous for its lush tropical landscapes and its delicious teas. Not Queensland. In a crisis situation, hours and days mean lives saved and lost. And duplicating effort often means that some critical areas get missed out. No matter how many or few resources you have, whether it's developed world, relatively rich country, or whether it is the poorest of the poor, coordinating your resources is vital to maximizing the impact of whatever it is you're doing. And the Sri Lankans quickly realized that the problems they faced would be better tackled using software than an old fashioned pens and sticky notes approach. So they contacted the US Federal Emergency Management Agency, FEMA, assuming that having dealt with emergencies and so on for many decades, that they would have something that would help solve these common problems. Facilitating the search for missing persons, coordinating aid and volunteer management, looking after victim tracking through refugee and displaced persons camps and so on. No, no, the only software FEMA had was for cutting checks to those who needed compensation. Now, one thing you wouldn't necessarily know about Sri Lanka if you stuck to reading the tourist guides is that they actually have a very strong tradition of free and open-source software. So it was no surprise, really, that when technology was needed to help manage this massive relief effort, the Sri Lankans turned straight to open-source. And so, under the auspices of the Lanka Software Foundation, open-source developers, primarily in Sri Lanka, but also around the world, worked literally day and night, 24 hours, to produce the first release of what's now known as the Sahana Project, which was almost immediately adopted by the government and went into production within about a week. So the Sahana Project is all about enabling a cohesive, coherent disaster response between multiple groups and agencies and bringing everyone together to help those who need it most. Of course, to refer to Sahana as a piece of software is kind of an oversimplification. In reality, Sahana now embraces three distinct projects. Lion is the easy one to explain. It's the localization work that ensures that everything else can quickly and easily be pieced together and thrown out there wherever it's needed because you usually don't have a whole heap of warning. Agastiy is the original PHP-based emergency management software and Eden is a broader system based on a Python port of Agastiy, but also encompassing some more general humanitarian assistance and disaster response functionality. And I'll talk a little bit about the difference between those things in a moment. But what really strikes me about this is can you imagine a commercial company that can do the same, that could divert literally thousands of hours of engineering work at a moment's notice to develop a modular, reusable piece of software that was ready to be deployed a week later? I mean, I'm all for agile, but that's pretty good. But all this was undertaken by volunteers and the resulting software has been used literally across the globe from Haiti to Peru, translated into more than a dozen languages, ready and waiting for wherever it's needed next. And for me, that's really just the power of open source. Going back to humanitarian assistance, emergency management and disaster response, they do sound similar and a lot of the software requirements are pretty much the same for all of these. So please excuse me if I use these terms interchangeably from a software point of view, sometimes they are. But more generally, humanitarian assistance typically describes the sort of longer term relief work that's required in various places around the globe, usually in chronic settings where the challenges faced just have to be addressed with continued ongoing work. There's no band aid that's gonna sort out these problems. The work is usually well understood. The requirements are clear, even if they're a little complex and the pace is slower. It's necessarily slower because it's so much more sustained than during a crisis. Emergency management on the other end of the scale is immediate and brief. Speed is of the essence. Communication is often nigh impossible. Even in the developed world where we have a lot, we have N plus two, backups to backups to backups. It's very easy for phone networks to go out or simply to get saturated by the number of people trying to stay in contact. And it's the most basic things that are missing. The human nature is that we want to do something at this stage. People want to go there to help, to do whatever is needed. But the problem is the coordinating those people is absolutely critical. It's very easy for those rescuers to become an additional burden on a system that simply can't manage with what it's already got. And so there's a combination of telling people to just calm down and slow down a little bit but also coordinating what you have and what you need. Taking Haiti's earthquake last year as an example, the emergency management required there was focused on finding survivors, digging out people who were still alive, were talking still in the first few days, stabilizing the casualties, working to reestablish the basic infrastructure. And when I say basic infrastructure, we're talking like running water. Sometimes shelter, but really the truly, these are the things you cannot live without. Sanitation, food and water supplies. And this phase rarely lasts more than a few weeks after a natural disaster. Although it can be more prolonged when you've got war or conflict going on. So somewhere in between these, and if we go back to Haiti still going on more than a year later, is disaster response. It's usually more chaotic and immediate than humanitarian assistance. You don't have that sense of, okay, here's what we need to do, here's how we're going to do it. Here are our ongoing donors and so on. And it's often broader based than the initial emergency management. Sanitation can remain an issue. Basic infrastructure, medical care, these things need to all be brought back online. And what's needed at this stage is still an organized, coordinated response. But by now, usually it's easier to manage volunteers coming in and hopefully a willing pair of hands is more of an asset than a liability. Emergency management and disaster response were really the topics of the hour in Kenya as 2007 turned into 2008. Some of you may have heard earlier today about openness in the public sector from opening up government data to open democracy. But when the Kenyan presidential elections declared victory for the incumbent against a backdrop of manipulation, accusations of intimidation, and in fact the opposition leaders claims that he had won. An open source project sprung up and helped to save lives as those government processes were breaking down. In the month of followed the elections, as many as a thousand people lost their lives as riots and violence broke out. And it could easily have been many more if it weren't for the work of Ori O'Colo, a Kenyan blogger with a background in law and activism, who suggested and drove through developing a website that would track eyewitness reports of violence and of areas that were in need of aid. Within days, developers from Africa, from the US and further afield, had set up a site that would let the public submit reports of riots, lootings, violence, intimidation, either by SMS or online. Online seems like a very easy and natural way of doing things for us. But you'll find particularly in Africa, it's much more likely that people will have access to a cell phone than to any kind of a computer. Volunteers would then attempt to verify the reports that came in, either with local media, with non-governmental organizations that were on the ground in those areas. And the incidents, both reported and verified, could be displayed on a map. Shortly thereafter, the site was expanded, enabling people not only to report ongoing violence, but also to request aid as it was needed. Those requests could of course be verified, mapped and tracked. And the site was eventually even extended to be able to accept donations from abroad. Kenya, of course, having a pretty large diaspora. So Ushahidi was the name given to the software that powered all of this. It's a Swahili word that means testimony or witness. And as with Sahana, although this software was developed to meet a specific local need, as open source software took on a life of its own. Just weeks after it had been rolled out for the Kenyan troubles, the same software was used to create a site tracking anti-immigrant violence in South Africa. It's since been used to track everything from stocks of vital drugs in pharmacies across East Africa to blocked roads after snowstorms in North America. From deciding where to send volunteers to help with rebuilding after wildfires in Russia, to election monitoring in Mexico and India. Ushahidi has really, truly gone global. And if that website looks strangely familiar, that's because it's been used here too. It was literally hours after the flash floods toured through Toowoomba before ABC launched the Queensland flood crisis map. I've mentioned time spans a few times here. Days after the tsunami hit Sri Lanka, Sahana was up and running. Days after the Kenyan election crisis, the original version of Ushahidi was launched. And here truly it was a matter of hours. It was less than a day that a new instance of that existing ecosystem was deployed. In emergency management, as you can imagine, time is of the essence. And using open source software facilitates that in a number of ways. For a start, bypassing management approval is a whole lot easier when you don't have to pay thousands of dollars to license or buy your software. But way more importantly, when you use open source software, you have a community of talented, interested volunteers who can help you and who provide follow the sun coverage as people on one continent go to sleep, people on the next continent are waking up. And of course, the sheer number of people willing to contribute means that no matter when a crisis hits, Stephens' day, no problem. There'll be somebody available to get things up and running. There's no company, party, or national holiday that can take it out. But time is not the only factor in getting things up and running. Accurate information is also vitally important. And to me, that's one of the key strengths of Ushahili. In the early days, ABC saw many reports about road closures, flooded streets, evacuation centers, sandbagging locations, so on. But as time moved on and the disaster moved into more of a recovery phase, the map has started to collect reports of volunteers needed, pets lost and found, locations for bottled water drops, bulk bins, and public health information from disease risks to vaccinations. And each of these reports can, if necessary, be verified. Obviously, local authorities don't need to confirm who has space for a few extra pets when a shelter needs to be evacuated. But knowing which school building still has place for evacuees and which health centers are open and providing vaccines builds public confidence and helps keep that infrastructure ticking over. And so you can see pretty easily on the reports which have been verified. And once you click through, you can see who they've been verified by and then which are simply unverified. Since it was first developed, Ushahili has also grown. It's no longer merely a reporting and mapping website, but it contains a whole suite of tools that can be used to gather data from almost any source. And in fact, part of the secret to how quickly ABC had this up and running lies in this. Because it turns out before the rain started falling on Queensland, they had already been piloting it. For Feral Month, the ABC were asking Australians to report any feral animals they saw. And they described this as a citizen science experiment as part of an ongoing investigation into crowdsourcing and social media and how they contribute to emergency management, major event reportage and information gathering. So they already had a deployment up and running. It was simply a matter of just customizing it a little bit and making it specific to the floods. But some of the tools that they were piloting and testing out there are SMS Sync, Sweeper and Swift River. SMS Sync allows you to turn any Android phone into a cheap SMS gateway. So rather than having the time and expense of trying to set up a short code or a commercial gateway, you simply install this application on your Android, anybody can text you, and a script passes these messages to the Ushahiti deployment. You can also set up filters and so on so that it only passes the messages you're interested in. Sweeper, which is the next part of the tool chain, uses natural language processing and Yahoo's Placemaker API to curate the tweets and map them as they come in, which frees up volunteers who otherwise might be spending time and effort on doing this, trying to wade through just the vast quantity of data that's generated. And finally, the Swift River API also helps to get through all of that data by deduplicating the stream, trying to make sure that noise is kept to a minimum and that signal is boosted. In both the Feral Month experiment and the Queensland Floods deployment, open source software made massive amounts of crowdsourced data completely awesome and useful. But where did that data come from? By and large, it came from the social networks and the social technologies that we're already using. Technology facilitates and exposes the connections that we already have and it enables some amazing possibilities simply by connecting people quickly and efficiently. Twitter in particular came into its own during the floods, but Facebook and even YouTube were credited with helping to keep information flowing. All of these technologies were successfully used not only by individuals to connect those who needed help with others who could provide it, but also by official bodies, most notably the Queensland Police, to respond to the public, to bust the myths that were circulating and to give people the information that they needed to stay safe and well. As Kim Charlton, Director of Media Affairs for the Queensland Police, said of Twitter and Facebook, rather than us having to create a website and build an audience, we're going to where the people already are. And when you're dealing with a crisis situation, that's really the last word, going to where the people already are. Because somebody who's evacuated, maybe to a friend, maybe to a shelter, they don't have all of their stuff at home, they're not looking online particularly, they might be getting the tweets coming in on their phone or whatever, but that's really all they've got. And so it's a matter of getting the information to them rather than asking them to come to the information. Now, I promised I wouldn't talk about disasters all day long, but what all of the software we've talked about so far has in common is the assumption that the users have access to computers and the internet. Even with Ushahili in Kenya, where you can text in your reports, the assumption to see the maps is that you're looking at a computer. So there's a disconnect between those using the software as well and those who are being helped by it. The people using the software, the crisis management agencies, maybe it's ABC, maybe it's governmental, whoever it is. I'd like to talk now about some open source software and in fact, open source software and hardware that's saving the world, starting with people who don't have those things that many of us in the room would consider pretty much essential. Globally, more than 776 million adults do not have even the most basic literacy skills. That's more than 36 times the population of Australia. If it were a country, it would be the third largest in the world after India and China and it would be more than twice the size of the USA. As it is, it's one in every five adults globally. And of course, it's not as simple as looking at the two people to the left of you and the two people to the right of you. Poor literacy is a problem of opportunity, not of intelligence. And like so many crises, literacy problems disproportionately affect the poor of this world. Litteracy can impair people's ability to obtain or transmit vital information about their health, their livelihood, tools they use, the challenges they face. Even in the developed world, literacy is truly a matter of life or death. For example, not being able to read your prescription bottle or your hospital forms, truly can kill. A study at the Feinberg School of Medicine, a university hospital based in Chicago, showed that older people with inadequate literacy skills had a 50% higher mortality rate than their literate counterparts. In fact, after smoking, low literacy was the number one predictor of mortality. In the developing world, not being able to read seed packets or fertilizer instructions limits crop yields, costs lives. Not being able to read ballot papers or election manifestos, limits full participation in society. Not being able to read how much you're owed and how much you've been paid, opens doors to serious corruption and further exploitation of those who are already the poorest of the poor. Literacy Bridge is an organization that seeks to combat these disadvantages. By giving communities in the developing world the tools that they need to grow literacy on their own and, more importantly, to obtain and transmit information in immediately accessible, environmentally and culturally appropriate and locally relevant ways. So they're trying to build literacy but they're also spreading this information while they're on that path. The Literacy Bridge team have built a device called the Talking Book. It's designed specifically for people who cannot read and who live without electricity. It's the world's most affordable and durable audio device and it has some amazing features I'll talk about in just a moment. But in the words of the Literacy Bridge team, it's like a small touchpad computer that enables local experts to spread information to the most inaccessible communities on earth. Before the Talking Book came along, NGOs in places like Ghana would teach a local volunteer about topics from agriculture to education, politics to health. Those experts would get in trucks on motorbikes or go out on foot visiting villages and rural communities. They would often quite literally stand in the shade of a tree and shout out what they knew. And if you weren't there, because you had to look after the livestock, you had to mind the children, you had gone to get water, you weren't well, whatever it was, that was lost. If you were there, you heard, you went off, you did your work and you sort of weren't sure what had gone on or you missed a bit or you got confused, forgot some of it, that information was just lost. You don't get a second chance. And even in optimal situation, if you're there, you hear it, you get it, it's all good, there's kind of only so much information. A guy standing in the back of a pickup or under a tree can transmit while you're sitting waiting for him. Now, field partners can record information directly onto a Talking Book. They distribute these devices to willing participants and they give a short demonstration on how to use them. The audio menus are easy to navigate by touch and users can repeat the information they're interested in as often as they need to. They can play it back for family and friends or pass the book on to someone else. In fact, in pilots, each Talking Book was shared naturally between many families. Anyone can record new audio onto the Talking Book to expand on existing ideas, share new information, give feedback to the organization providing the devices or simply to tell their story. And most impressively, two Talking Books can be connected to each other to share audio between them so that new information, updates, local translations can be added and passed on without any further intervention. And a lot of a thousand of these costs on the order of five or six thousand US dollars. So they're basically a five dollar iPod. Never mind your hundred dollars for an OLPC. The people that these are going to could never afford that. Five dollars, just about doable. For use in the classroom or for learners, you can adjust the speed of the audio. And audio links within the files can prompt the user to listen to definitions, answer questions and so on and so forth. That's pretty good for five dollars. And of course, it's all open, all powered by a locally available D-cell battery and freely available open source software. I mentioned a pilot program. Oh, hang on, this clicker has gone mad. Sorry, excuse me. Let's go back. So I mentioned a pilot program a few minutes ago in which, thank you. No, okay, let's just give it one second. Let's see if we can. Talking books, 21 talking books were sent out to a village of a hundred people. And, okay, we'll leave the slides for now. Where non-users saw crop yields drop by 5% from the previous year, farmers who had access to a talking book saw crop yields increase by 48%. That's a return on investment of more than 300% in the first year. And the talking books can be used again and again. And of course, they can be used to provide much more than agricultural information. More than 91% of those who had access to talking books applied both new health practices and new agricultural practices in their households. And local women regularly reported listening to the books with their children. Sorry. Can someone just come up and tell that to go away? Which is a key ingredient in bringing change to the developing world where it's been repeatedly shown that educating women proves more effective than any other education provision in bringing about change. As you can see, if I can get my slides up, many farmers chose to, okay, just exit in the corner. Many farmers chose to apply the agricultural advice only on a portion of their land. When you're living in Ghana, when you're a poor subsistence farmer in Ghana, you simply can't afford to take good advice on the risk that it will all go horribly wrong. Yeah, just done, yeah, or exit. I would really like to show you this image if I can, because I think it really speaks to the whole thing. No, okay, oh well. What I can do, perhaps, just drag this over. That's showing you the difference between part of the field planted using the traditional methods and part of the field planted using the methods advocated by the talking book. I think it's pretty awesome. I think I'm really awesome, sorry. No, that's quite all right. Do you know what the differences was with extra fertilizer or whatever? Oh yes, so the differences that they were advocating were essentially in how you were planting, and what they were advising doing was to plant the corn or the maize in a, to make a little hill, essentially, just a small little hill, and to dig a little ditch around that so that when you got rain, whatever rain you did get, essentially went on, came down this hill, there was less erosion, and then it was caught in this little ditch, which basically provides you with the most basic, the simplest irrigation system in the world. So it was literally, it was just in how you planted, pretty much. It's really, really simple information, but it's just trying to get that information out. So I've told you about some of the projects that I love that are dedicated to saving lives, but open source software is saving the world in a much wider variety of, and much more subtle ways as well. So let me just return to the mundane side of things for a few minutes. By a series of happy accidents, while I was in university, I started getting involved with open source software, and worked on contributing documentation to the Apache web server while I was there. I'm sure most of you've heard of that web server. It started to evolve out of NCSA's web server in the mid 90s, after one of the principal developers left there. People who relied on that software to run their sites got together and started sharing the work that they had been doing to keep things up and running. Those people eventually became the Apache Group and later the Apache Software Foundation. But that's kind of enough history. There's another aspect of the Apache web server that we don't always think about. It's basically saving the world by proxy. As we speak, charities around the world are informing the public, they're soliciting donations, they're getting volunteers, and they're carrying out their missions online. And many, many organizations from Amnesty International to Médecins Sans Frontiers, Save the Children to UNICEF, are all able to do this because that Apache Group wrote a web server, published it in a free and open way. These organizations have accessed hundreds of thousands of hours of engineering time for free. It doesn't cost them a cent. And what's more, if the software doesn't do what they want it to do, it's pretty easy to get someone to tweak it, adapt it, add onto it so that it does do what they need. And if they need something to keep their website in order once it's up and running, how about Drupal? If they want to keep track of donors and volunteers, Civi CRM. There's basically nothing you cannot do with open source software once you start looking. It's blissfully simple, really. Open source is a common good. Each of us receives far more from it than we put in. No one can take away from any of us the contributions that we've made, the work that we've created, and yet we can all take from what others have done. And we can use that, we can build on it, we can adapt it, we can do whatever we like with it. And that's the wonder of open source software. It's the joy of open source software. It's the strength of open source software. And it is how open source software from crisis management tools to audio players to web servers can and will and must save the world. Thank you.