 Hello, everyone, and good afternoon. My name is John Paltry, and on behalf of the Berkman Center and also the Harvard Law School Library and also the Harvard Book Store, I'm delighted to welcome you to the Harvard Law School and thrilled that you've come out to hear us talk about a new book that we have been very excited to share with the world here today called Interop. I'm thrilled to be here also with my friend and co-author, Worst Gaster, with whom I've written this book. And I'm going to do a very brief overview of the core argument of the book, Worst is then going to take us through a series of examples and perhaps show a few little videos along the way. And then hopefully we'll have a conversation with all of you. The event, just so you know, is being recorded, but not webcast live. So should you care to ask a question or throw tomatoes or do something else that will get recorded, you are being recorded for posterity. We'd love to hear you say your name and so forth and do that, but since one of our topics is privacy, we wanted to disclose this in advance. We highly encourage tweeting. We are using pound interop, so the hash sign interop, as a way to curate the conversation a little bit. And there will be a reception just after this in the section right here, and we very much hope that you will stick around for the reception afterward. So I'm just going to give a really brief intro to the topic and then turn it over to Urs. The topic of interop is something that seems on its face to be very dull and very techy and very geeky. And to be honest, we started with this very dull and very geeky topic. And we're asking questions as part of a research project about whether or not higher levels of interoperability led to greater innovation. That was our core initial question, which sounds really, really academic, and it is, and actually extraordinarily important from the perspective of the development of complex systems in the world, but a relatively narrow version of the kind of question that might result in a book of this sort. And what happened over the course of working on this book and with a wonderful team of people over five or six years was we got more and more interested in how interoperability actually helped to explain a whole lot of different things. And we got quite a ways, outside of our ordinary comfort zone of information technologies, into a variety of fields. And then sought to do something more ambitious than simply to explain this relationship between interop and IT and innovation, but rather to develop a theory of interoperability and how it related to a whole bunch of other questions associated with an increasingly globalized world and one where information and technology is increasingly mediating our lives in lots of ways. So that's the overall trajectory. It was going from a very narrow technical question about technology and data to something that we think relates between it and how we interact and how institutions work together. And it's our sense that it also helps to explain a number of phenomena that we see in the headlines every day. And that's in a way what we're seeking to do is to make this topic of interoperability be something that is useful and pragmatic, as well as being highly theoretical and constructive in an academic sense. You have no doubt been following the fact that Facebook IPO'd and I think the way in which the story has been spun is largely about the spiral after the initial offering based on some shenanigans associated with how the IPO was handled. And of course, that's a very important story. But we see sort of a story behind the story in a way that has to do with interoperability. One of the debates I think about Facebook has been why is it plausibly worth $100 billion when it goes out initially or $80 billion or whatever it is falling toward when on the basis of sort of a basic economic calculation about its revenues and advertising, it's very hard to justify that amount. I think one of the real reasons why it is so valuable and may well be worth something on the order of tens of billions of dollars and why companies like Friendster and MySpace and others that had similar technologies and similar general approaches have rounded to zero or become worth next to nothing. And maybe even why LinkedIn is worth more like $10 billion and Facebook is like a lot more, is its approach to interoperability. The extent to which there are of course wonderful other things about Facebook, but one of the things that they've done is from the start built themselves into the technology, into the systems broadly and into our lives in ways that make interconnection something that may well be enduring and that be very hard in some respects to pull Facebook out of the technology and out of our experience. So the way in which they have done this, I think has wonderful benefits for the company in terms of being hard to replace and hard to extract from our lives and from the technology. It also comes with some problems, some problems well known around privacy and security in particular. I think this is a really good example when we say the promise and perils of interoperability of how that all gets rolled up into one example and we'll play out some of these stories a little bit further. The second one has to do with the extreme spread of viruses that happen across any number of different complex systems in this globalized world. This is just one example where the flame virus is inspected under the United Systems in Iran. It's no longer the case that a malware outbreak is something that is possibly localized. This is something where there are huge costs potentially to our race to interconnect the world wide web. And it's again, they're wonderful things about interoperability, but they're also extremely high systemic costs. We have to figure out how do we build the firewalls or the breakwalls in between different systems and don't let contagion spread consistently as we build, as we rush to build these extremely interconnected systems. These are problems that we simply couldn't have anticipated a few dozen years ago, or if we could have, we didn't. A third example just from this week is the one most furthest slum from our core IT example, but which I think helps explain the spread of this idea beyond the IT zone and into other fields. It has to do with the failure of the banking system in Greece, but its effect on lots of other systems. And one of the ways in which we got into interoperability was of course to this IT lens, but then we started going historically and saying, what are examples where across different systems we have tried to make things interconnect. And two of the examples that came up really quickly were transportation, where we started with different training systems in different countries, and then started to figure out how do we have the railgages be at the same distance so we could run trains across countries. In the United States there's a wonderful story of the golden spike being driven into the middle of the ground in the 1860s which joined what they called the two great oceans of the world in that wonderful colonial sense of the era, but where we were able to have not just trains, but ideas and people and goods flow across the country once we had driven in that spike. Another example of this is currency. The idea that we don't have one currency around the world and yet we can have global trade. It's not all the same, but we have ways in which we can trade money across different borders. And I think that's something that has been a great globalization story. It has allowed for growth and the creation of jobs and so forth, but I think it also creates problems so we have to figure out how to solve as we race to interconnect more and more. And I think the Greece example is a really good example of how it can go bad. The notion that bad debts in one country can be such a big problem, not just for the Germans or the Swiss or the French in the example of Europe, but also for all of us, for the Americans and otherwise. What do the Greek defaults have to do with our system in the United States? The fact that we are so interdependent and interconnected I think is a big part of that story. So it goes from the very precise and classic IT example to the issues I think of the global economy broadly. One of the things that we seek to do also in this book is to situate the argument about interop in both very specific definitional terms and then in a broader frame. So in the world of technology, when we talk about interoperability, we almost always think about it in a really straightforward way. We adopt a simple definition. It's based on one that is usually kicked around through the NIST framework. But it's basically the idea of an ability to transfer and render useful data across systems. But ultimately I think where we come out in this book is that's an insufficient definition for what we think interoperability is. And more broadly, we see interop as the art and science of working together. This will become clear I think over the course of these examples. But we see interoperability playing out not just in terms of technology and data going across systems. Those are important layers of the story but also at human and institutional levels where the examples like the Greek crisis seem to me to make more sense. I'd also like before turning over to Urs, to make the case that the projects that we're working on here is one that is more than just a book ultimately. We've thought about this as an inquiry into this really interesting global phenomena of interoperability. And we've had the great pleasure of working with lots of friends and colleagues on a series of case studies over a number of years. And we've thought about this as something that is an extension of the book. And is a way in which if you want to go much deeper into these examples, we've put a series of case studies across the web. So there are ones that we did several years ago on this core question of innovation and interoperability. But with the help of many wonderful people in the room, we have done case studies across a very broad group of topics. They're openly available on the web. They're ones on cloud computing. You see Matt Becker's up there who's a Harvard Law School student. The one on electronic data interchange, one on intermodal connectors, one on the smart grid, I see the smart grid author Paul over there. So the way we've seen this project has been to do a whole lot of case studies. Adam Holland worked on many of these as well, all on the web. But then in essence get rolled up and distilled into this one argument. But we hope that the book project in a way will be not just a single distillation in a printed one, but rather an extended argument that is out across the web. And we hope as it gets reviewed and otherwise it will also be in a broader conversation about interoc. So with that I want to turn it over to us but by way of thanking this great team of Paul and others who have worked with us on the project. Thank you, John. So as John just mentioned, we've done a number of case studies. Fun ones, fun ones, complex ones, straightforward ones from daily life. Cell phones, we'll talk more about that. To highly complex systems. And what we'd like to do over the next 20 to 30 minutes is to explore with you, discuss with you five such examples or stories and share with you some general observations. These five examples are placeholders from many other cases we've studied. But they're illustrative for some bits and pieces of the theory of interoperability that we actually tried to develop in our book. So let us start with the first one which is actually a very big vision, a big interoperability problem or challenge. And that is, how can we improve our lives in cities? How can we make our cities smarter using technology? I thought we start with a video for a change. So watch this video which is actually from IDC. Have you decided where you'd like to study when you graduate? Don't know. What about New York? It's like got an awesome club and music scene. New York? Well, I'm sure it has, but... And it's got a great education system, great research facilities, schools that use cloud computing. I know, I know, but how safe is it? I mean, we've got some great schools here, too. These days New York's safe as. It's not shooting up like on TV. Really? Yeah, like the cops, ambulance and fire service are all interconnected and share information. Well, I guess, but all that traffic congestion and pollution won't be good for your asthma. OK, what about Dublin or Stockholm, then? They've integrated their traffic systems, so there's one ticket for trains, buses, ferries and toll roads. Oh, handy. People get alerts if there are delays, so they can choose the best way to get where they want to be. In Stockholm, they've cut down greenhouse gases by, like, 40% or something. But the cold and your asthma? Well, what about Spain? Spain's nice and warm. And there, they've interconnected health services, so I won't have to run around getting my medical history to different doctors, specialists and pharmacists. They'll have it all on one single electronic health record. Well, that's smart. But what about the water and utilities? Yeah. You don't want to live in a place unless it has a guaranteed energy supply and good quality water. Oh, you mean like in Malta? Malta. Yeah, they're like building a smart grid that links the power and water systems. Oh, it'll detect the leaks and give more control to the people who buy water and electricity? Well, I do stop here because they go around the globe and that many more countries, right? But this example is illustrated, this big vision that just illustrates how much we depend when we think about our future as a society, here in the example of living in cities, on sharing of information, on getting interconnectedness right, which is the story of interoperability we are interested in. There are a number of examples that we can, or observations we can distill from this particular narrative. The first one is many of the things, cases we've studied, whether it's in transportation, as Tom mentioned, whether it's a smart grid, the future of efficient energy supply, whether it's air traffic control, many other areas. It turns out that interop is, in a way, a key success factor for making these future infrastructures that are so important for our economies, for our societies and cultures. But interop is really the key piece, so we need to understand interoperability very well when we think about the solution to some of the most pressing problems we face, really from the health care crisis to climate change, wherever you look, this question of how much interconnectedness, how much flow of information across systems, that's the driving question in many instances. A second observation from the story, and that's also kind of a key characteristic of interoperability is interop is not the black and white thing. It's not something that is either there or not. The flow of information across systems maps along the spectrum. We know that from daily experience. I assume many of you are traveling quite a bit, and sometimes we need adapters for our power cables to carry with us, so this is kind of somewhere, it's a hack to create interoperability, but it's not the perfect interoperability where you could have one plug and travel around the road and plug it into the plug there. And of course, not only this degree, different degrees of interop, not only is experiencing in simple systems like an adapter or a dongle for the computer here, but also complex systems take the alliance of airlines. They, I'm sure many of you again, travel example of experience difficulties. You say, book a flight, make a flight reservation with Swiss air or Swiss, right? And then you fly United, which is in the same alliance, but actually there is a translation problem quite often that the seat you reserved within the Swiss system didn't perfectly translate into the United system, and you end up with a window seat instead of a aisle seat. So if you look behind the scenes, these reservation systems, they are hugely complex systems, and so you see again a certain degree of interoperability, actually, you're on the right flight, but then you don't have to ride seat, so it's not the point. You should be glad you didn't get a middle seat, which is what I get here. You're such a serial flyer. You get messed up into the aisle. What also, what is a kind of common thread across, oh, that's cool, what's a common thread across the book, certainly looking at all these different cases and trying to synthesize them is that interop is a really fun challenge. The smart city example is a great bomb. If you think about making the city, the city's transportation system more efficient, we really need to think hard about different elements that need to play together, the workflows, the organization models, and so forth. So the first challenge, and we will get back to this question is really, how do we determine if interop is a spectrum? Where exactly do we want to be on this spectrum and have systems work together? And then the second challenge is how to get there as a design and planning process. It turns out once you mess up with interoperability, it's actually very hard to create it, take air traffic control as another example. It's very hard once you have a given technology in place and a high degree of complexity actually to move to the next version of a system using new technologies such as GPS, turns out actually in air traffic, we don't use GPS that much as we actually use it, otherwise all the time for navigational purposes. So there are real issues where it's easier to plan for interoperability rather than to create it once you have a system in place. What also has become clear from this introduction video, looking at smart cities, just very briefly of course, is that there are many benefits to highly interconnected systems when the video showed healthcare and the efficiency you can introduce there when you share information such as health records across institutions. The example also showed firefighters and other first responders working together how this may enhance safety in a city. There was an argument about ecological benefits, more efficient water supply, more efficient utilities, again as one of the effects of increased levels of interoperability. And that's a finding across many of the case studies you looked into that indeed, interop increases systems efficiency that it also increases user autonomy and choice straightforward if you can take whatever device you want or cell phone you want and access any app store you want to have more choices to consume or if things are interoperable. Turns out of course they're not always interoperable. And John mentioned that at the beginning also one of the core findings certainly is that more interop is good for innovation and we'll return to that. Just very briefly, John already made this point with the headlines that he mentioned at the very beginning. The story told in the book and in our research is not only a story about the technological layer and the data layer of interoperability. Of course those matter a lot. Go back to the smarter city's example and first responders, emergency services to map those on the different layers. Of course the police forces and the firefighters need to be able to connect at the technological and data layer. Their radios need to communicate with each other to be more efficient and provide increased safety to the citizens in a city. But then there are also higher level interoperability questions that are equally important. Human interoperability turns out that for many years police force and firefighters and ambulances use different language, different coded language to signal emergencies and there was no interoperability because code 333 meant a man down in the police jargon and something different like a barbecue that got out of control in the other case. And so this coded language problem, the lack of interoperability is just an illustration of the human layer of interoperability and we see that human layer of interoperability as a challenge across many of the apparently technical case studies we study. And then on top of that is the institutional layer of interoperability where organizations come into play where also laws come into play where actually the laws of Europe may not be interoperable with the laws of the US if you consider the example of privacy protection you have very different laws in Europe than here which makes it very hard for instance for international companies such as Google and Microsoft and many others to operate on a global scale. To stick with the rescue first responders example in 9-11 the institutional layer of interoperability or the lack thereof was a matter of life and death even because the police helicopters say actually did see that the towers were unstable and you had all the layers of inch up down there they could communicate with the firefighters and so forth but the firefighters in the tower basically said well that's not our chain of command we're not following the orders from the police helicopter because we're firefighters so we have also high level to a signal interoperability problems. So that's just one example here. A second one that I want to touch upon are open platforms. This highlights the point that John mentioned at the beginning that interop is really good most of the time especially in the digital network environment for innovation and open platforms such as Facebook are great interop stories. John already made this point so I won't go into that in great detail. Will mention though that when Facebook opened its application programming interface made available the possibility for other services and application builders to communicate with Facebook we have seen a spike in innovation so 4,000 applications within a couple of months were created by non-Facebook entrepreneurs that then plugged into Facebook as apps I'm sure many of you use apps and get the idea here so you see how a decision made by Facebook to become more interoperable led to innovation at least horizontally. Another example or story to tell is Twitter where we see a similar effect on opening up interfaces making more interoperable one particular service how that leads to more innovation and much of the story behind Twitter is actually also an interop story again. There are also less commercial example and more human examples of interop in this open platform space I just want to highlight one that's the Ushahidi Ha'iti platform in the aftermath and this is just another illustration of the power of interoperability across the different layers I just mentioned. In the aftermath of the Ha'iti earthquake a bunch of students and volunteers actually created using a platform that was developed by a HLS graduate by the way created the most powerful and most important crisis coordination mapping tool for the recovery efforts and relief efforts in Ha'iti where actually volunteers that helped in Ha'iti first responders could use whether it's Twitter, whether it's SMS whether it's laptops or even phone calls to map on a literal map where is help most needed how does the damage look like in given areas of the city what kind of support, what kind of devices do we need for our relief efforts so the power here of course as you see immediately is a couple of volunteers build a platform, use a platform that is highly interoperable to do really good and actually become a very important driver in a crisis in a moment of crisis like this earthquake. It's worth pointing out of course again you see how you need to have lower level interrupt Twitter connects the data set from Twitter connects to the map here but then over time there is certainly also the human layer of interoperability where first responders start to coordinate work together going back to what John said the arts and science of working together and now also at the policy layer for instance the UN got really interested in these mapping techniques as an efficient way to coordinate relief efforts and you have conferences hosted by the UN so at the top institutional level thinking hard about how much interconnectivity do we need to have in this decentralized world where we have many relief workers there are less serious examples safe2p.org right if you look for a restroom it's the same idea that applies here so looking at these examples and many more case studies from a more theoretical perspective we see this cycle that more innovation increases competition which then in turn leads to more innovation there are numbers of theories that support that in addition to case studies I have to say one of the struggles that John and I had writing this book is there is not much empirical evidence out there that supports this claim that the interrupt is good for innovation there's plenty of anecdotal evidence in case studies but also and that's comforting of course support through theory I think Jonathan Citrain is here I've seen him before oh yeah hello welcome his theory of generativity von Hippels Professor von Hippels concept of user driven innovation small step innovations that these are some of the theoretical approaches to explain and support the claim we're making and we can go into that in greater detail during the discussion so it's important to note for the economists in the room or people watching us later on that it's not always the case that more interrupt is good for innovation there are instances where if companies have the hope to create a new product of service that isn't interoperable with other services they actually get the entire market share the entire market instead of just a slice of the market and therefore have something that is close to monopoly profits which may increase of course the incentive to innovate because you have the hope to get huge revenues right so there are caveats I think since we're at the old school we keep it short on this one for the moment there is also another argument why interrupt is good for innovation and that's not only because it enables these great cases as we just touched upon but also it helps to spread and helps with the adoption of innovation the transition to high definition TV in the US is a great story where actually only one's interoperability between analog TV sets and digital programming was enabled people like us switched from analog to digital TV and it's also a nice story because we'll get back to that in a few minutes that's also an example where the government had to intervene and actually send out coupons to households so that we could buy converter sets to make this transition successful which has taken much longer than the government predicted to start with example three only briefly and I hope we can get more into that during the discussion credit cards of course credit cards a highly interoperable system whatever you know you're shopping opportunity is you swipe your card whether it's a visa master card or annex or whatnot and you know whatever your bank is whether it's Bank of America city or whatnot this transaction works ultimately and the store or the seller gets the money so it's a highly interoperable system super convenient if you look at the consumer data convenience is also one of the main reasons why people actually use credit cards but of course as we all all know this high level of interoperability comes with certain costs and risks we all read the stories about identity theft about privacy problems and security problems data breaches almost every day headlines in New York times and elsewhere recently Sony that lost like 75 million consumer records due to a hacker attack so you see how more interoperability actually increases vulnerabilities. Now our argument is that actually that's true there are costs to higher levels of interoperability especially in terms of privacy just from noted and security but that these are again costs that we need to address or prevent ultimately in the design process when we think about the creation of the next generation of technology. So the reason only briefly why I just love this illustration so mean thank you for creating this slide the reason why more interoperability creates certain vulnerabilities arguably is because you have more connection points you have more people organizations and so forth that tap into naturally into a data flow and can misuse of course the data that is exchanged in these highly interoperable networks if you would only have a connection between here and there you would have a lower risk accumulated risk than you have so many different points of connection. Example four that's actually my favorite one cell phone charges many of you know of course or may have many charges in your drawer I have like a gazillion of them right from my previous cell phones like one for Nokia one for my iPhone another one for my old BlackBerry so I have a real collection at home I think many how many of you many of you as well yeah many of you good and their interoperability is a real question right because it would be super convenient if you could use one particular charger for all the devices and actually it's hard to figure out why has no one come up with this solution because it again turns out there are huge benefits to that now the first stage of evolution was to make cell phone charges smarter here are just a few examples one is if you have a gas mask through breathing you can charge your cell phone and another one is the aircon where you have through wind you will charge your cell phone and the craziest one is if your cat is playing with this then it charges as well so this is kind of more on the funny side of things but now let's see what actually has happened in Europe to address this problem of diversity and non-interoperability of cell phone charges until now if you changed mobile phones regularly this often meant owning multiple incompatible phone charges these different charges are a pain for users and produce waste which harms the environment today a single common mobile phone charger which can be used to charge different mobile phones made by different manufacturers has been made possible based on the micro-USB standard the new charger is compatible with data-enabled mobile phones and soon these kinds of mobile phones will be predominant on the EU market the new charger exists thanks to the EU powers of persuasion it managed to convince manufacturers who make up 90% of the European mobile phone market to agree on a common charger model a common charger will make everyone's lives easier so again a great story I think and since we talked about the EU at the beginning and the crisis in Greece and the Euro I thought they want to include at least one positive story even if it's only about cell phone charges but why is this case study of interest leaving aside that we are all affected as users of cell phones I think one point that is really interesting about it is how long it can take to achieve higher levels of interoperability in this case the unified charger to agree on this standard that of course has to do in many situations with the fact that you have many actors in the cell phone charge example many different producers of cell phones and they for a number of reasons of course economic reasons first and foremost have no interest that you can also use and charge your cell phone with the competitors charger and therefore try to lock in consumers into their product line and make money by selling chargers and cell phones of course another reason we see in many systems is especially complex systems health care is a great illustration there is that it's hard to achieve interoperability due to the complexity of a system I save that for the discussion I hope we'll have some more time to look into that legacy problems, legacy problems I already mentioned that in the air traffic control context once you use one technology it's very hard to change high switching costs keyboards, great story, QWERTY standard all the keyboard layouts we use totally inefficient we know there are more efficient layouts but the switching costs are too high so we stick with something that is actually not the best available technology not the most efficient one due to such problems the cell phone charger example is also nice because it indicates different ways how you can and who can work towards higher levels of interop in this video it is the European Commission with some pride saying well are you know convincing power actually led to this higher level of interop in fact of course there was nothing else than regulation by threat, right the European Union, the Commission threatened the producers of cell phones and chargers if you don't find a solution to that problem we will mandate the standard and that's an interesting kind of tool in the toolbox there are many more more in the book we explore a broad range of instruments that we can use some of there are more on the private sector side others are more on the regulatory side approaches through legislators and regulators, state actors mandating standards, something we don't recommend generally we don't think governments do a good job in figuring out what the best standard is for a technological problem but on the other hand governments are actually doing a really good job in convening different stakeholders again here the cell phone charger example is illustrative and then you have a lot of tools in the toolbox on the private sector side to increase interop we will not go through all of those of course the baseline here is there is no one single way to interop and the challenge is really to figure out what's the best mix of tools to solve or address a particular interop challenge be it in healthcare, be it in transportation or any of the other areas we just mentioned and that's not a simple task and this goes back to this notion of interop as a design challenge I think with that I turn it over to you for the last example, Charles Wors, thank you and as you can see we get very excited about these many examples and I think that's one of the reasons that we spent several years working on this book was we kept getting excited about new examples and this is a warning which is if you start thinking too much about interop you see interop problems everywhere and it actually gets a little scary but one of the places that I see lots of interop problems is where I've spent much of my last four years which is working in libraries and I'm so delighted to see many library friends here too and I thought I would use this as our closing example before we get to some questions I think libraries actually are a wonderful view into the problem of interoperability and in some fairly troubling ways actually we cover two topics in particular depth in the book one is electronic health records which we're not getting into it's a wonderful thorny problem why we don't have a higher level of interoperability within electronic health records despite all this political will to do it the other area where we go in depth and this is because of our geeky interest in preservation of knowledge is in libraries and I think there are two things that really stand out of these case studies the first is to think about the preservation of knowledge over the long term as an interoperability problem and it takes the form of interop today in the context of e-lending which is the present crisis in books but it also has to do with interop over time which has to do with the reformatting of materials so just as a way into the topic you probably have somewhere in your home a series of little disks, computer disks that are three and a half inches or five and a quarter inches or perhaps those USB sticks that you now have but ultimately won't be that useful and if you think about it they're extraordinarily, extraordinarily unhelpful to have those old disks you do not have any way to play them unless you've kept successive versions of your computer there's a wonderful computer history museum on the West Coast where you could go and punch them in and actually render the data but the printed version is much more useful than those outdated technologies that shift over time and this is a real problem for libraries and cultural heritage institutions and others who are thinking seriously about how we preserve knowledge over time so let me talk about today and then the long term so the today problem is this very weird perverse situation in which libraries find themselves which is if you end up trying to provide digital copies of books you often can provide much less than if you provide physical copies of books to people even though increasingly people want these digital books so how does this work five of the six major publishers today in the United States will not let libraries get digital copies for the purpose of lending that's one of the problems another of the problems is that even if you could not everybody has the same kinds of readers there's not a standard, a good open standard format for books that everybody agrees to when they publish it's not that they don't exist it's that they haven't been adopted in the market and libraries have yet to be able to exert enough control to ensure that that is so it would make a lot of sense to have a new book provided on an open standard that we've all agreed upon and have libraries have access to that and lend it out and of course publishers, I love our publisher and others and MIT Press, our publisher for other books we want them to get paid and we want authors to get paid and we could absolutely find ways to ensure that that happens but today we have a system where much of the time you can't actually get the book in the first place from the publisher to be able to lend it as a library and secondarily even if you could people wouldn't necessarily have the right devices on which to play it this is crazy right why is this still a better technology than an e-book that just doesn't make all that much sense that's the kind of present day version of the internet problem but it gets much more interesting and much more complex over time which is increasingly libraries are spending more and more of their budgets on the electronic copies of knowledge rather than on the physical copies we are being told users want data users want electronic things so cancel that print in our library here at the Harvard Law School despite the fact that our budget goes up every year in the library we have to cancel because we can't keep up with the rising prices and because people want it in digital formats and digital turns out oddly enough to be more expensive so if what libraries end up having is digital copies of things but not so many of the physical copies and then the equivalent of going from the three and a half inch disk to the five and a quarter inch disk or to the little USB stick happens which has happened relatively quickly over the last several years we will have gotten access to a fair amount of information that will not have been preserved effectively over time now obviously there are lots of ways that one could think about solving this but turns out that even computer scientists think this is a relatively hard problem which is how do you ensure that we can continuously update materials now we make a fairly complex suggestion in a chapter in this book that there ought to be a deposit of a copy a digital copy with a particular registry and should be kept in an open access type of format following an open standard and therefore we can always be sure to update it and that's the sort of interoperability over time but today we have no such system when they're created as digital files they're held in these various hands in various formats and if we all switch from one to the next or switch on a various basis we will have done less well than books in this format that just seems to me completely crazy and a really interesting and hard interoperability I think there's a lot that we can do knowing what we know about interoperability knowing what we know about open access and open standards and so forth but we're not doing it today and we are rushing very, very quickly into this new world of knowledge without having solved this particular problem so in a sense I think the argument that we are making here is that when it comes to climate change which is the case of the smart grid or electronic health records which is the instance of trying to improve the healthcare system through highly interoperable systems or preservation of knowledge interop matters both as a theory and as a practice and that from the IT realm or from systems like transportation and currency and so forth we know how to make it work but we also need to do it by design up front and to have a sense of how interoperable we want it to be what's optimal and how we cut off some of the problems like privacy and security that are often attendant with high degrees of interoperability and maybe yourself turn it over to you for some hard questions no I get to close it out all right very good very funny he leaves me with the hard questions yeah very nice because I'm leaving that's true thank you not going far up the street so we'll leave you with a few questions we have not actually answered all of the questions in this book or otherwise there are more books to be written on interop in our view how much interoperability and for what purpose I think is a key one that one has to answer in the in the case of any of these tricky complex questions how to get there in many respects we think that there are a variety of different techniques that can be used and we need the subtle and nuanced assessment of costs and benefits in these cases and we could push on whether some things tend to be better approaches are open standards generally a better approach than proprietary standards processes yes and why but ultimately we adopt a framework that said there are many different ways to get there the role of government I think is one of the most controversial aspects of the argument we do not come from the purely libertarian strand of technologists that said the government has no rule we think that there are times and places where the government can help on the other hand there are not always roles for the government at every stage of the process and it's not always the case the governments are the most effective and trying to figure that out is something that we create a framework for in the book and then last of course this notion that even if you get to interoperability at one moment in time how do you manage it over time and I think this actually is in some ways the hardest one air traffic control is a perfect example from my perspective which is we figured out a way as of a few decades ago to get to a pretty good version of air traffic control but as new technologies have come along we haven't introduced them because we've gotten locked in at a certain level and I think we can see that across many different complex systems and the very last one as a hard question that we leave up there for all the academics in the room are methodological challenges I see my friend Esther Hargitay there with whom I taught a class on methodologies this is actually a really hard question to try to figure out how do you study this and understand reliably the answers we adopted a case study approach but to the extent that they're better empirical or other ways to study it that over time may well help come up with better solutions to the interoperability problems with that, thank you so much for the chance. So by my clock we have about 10 minutes until drinks and so forth outside so if anyone has thoughts, questions, challenges Doc Searles, we have a mic because we are recording if you don't mind. Doc Searles, are there any where you say you know what, give up, it's not gonna happen? What an awesome question. Yeah. I'm a cockeyed optimist. Every glass is completely half full no matter how much is in the water so I don't think we found that we didn't come to one where we thought it's just too hard. I felt that the electronic. So electronic health records was the one in the US which I think we thought ultimately was hardest so if you're the Danes and you have a few million people and you've got the Swiss for that matter, very good. Although I think the Danes might be a little further ahead than you on this, I don't know, maybe, maybe. And relatively, let's say you sort of contain political system and so forth. You can get to a relatively high degree of interop but as you know in the US and the UK you've had Prime Minister Blair at the time and you've had multiple US presidents, President Bush and President Obama say we're gonna get to electronic health records in the US by certain dates, 2014, I think most recently and despite the fact that there's political agreement that this is a good thing to do, oh by the way, there are privacy concerns but we just simply haven't gotten there. So I don't believe that there's an unsolvable interop problem but that one is I think the one that is the hardest. Hi, Mark Tomizawa. So I also saw at Berkman the gentleman who did the work on the consultation problem which was how does government conduct consultations with experts from across the United States? If you put that idea together with the idea that we're all addressable via our devices, what would it take to create an expert network of embedded people who just constantly report in what they're seeing? Because in an emerging dynamic system, what you need to do is put the sensors in the right places. Well what if the sensors are mines? That's a great question. To a certain extent this is happening when we think about the internet of things which also include human sensors. So there indeed is this trend towards connecting everything which of course comes also with risks and the special case of experts, I would argue we have lots of expert networks that have emerged on whether it's in medicine, whether it's in other areas of science, whether it's experts addressing governance problems. So I totally see that happening, of course enabled by the technology that we described a little bit in this presentation that fosters from the bottom up these higher levels and enables these higher levels of human interoperability. So that's the power of the technology that it lowers the costs for making these human connections possible ultimately and I think is also the most powerful argument why I'm interested and enthusiastic about the digital environment despite all the downsides of course that we acknowledge as well. I think we're in the experimentation stage and that's perhaps a good thing. This is a great one in particular just that one additional note is it's not clear to me that we want that mode of governance, that we could handle that level of interconnection. I think lots of people might think, oh that's great, we get lots of consultation but I don't know if direct democracy enabled it that way is something that we could in fact be able to manage. I'm sorry, I was three questions. Okay but maybe could we just try one per just because there are a lot of hands? Do you mind doing one question? Okay. Your favorite one and we'll talk to you after about the other two. We can also vote them up or down. Or you can ask all three and we'll answer one. Do you think a human's dreams can be controlled by someone else? I don't think we've addressed that question in this book first of all. So can human dreams be addressed, controlled by somebody else? Nor do I have a particular insight into that so I'm sorry. What you talked about today included the social system, not only technology system is right. Yes I think that's completely right that what we're looking at is there are social systems as well as technological. Social system include every citizens is right. I think they certainly could, I think that's the governance example we just heard actually. Oh, that's the question. If it is me, I had this kind of experience would you like to tell me what's your response? What's your response? Maybe we'll talk to you about it after but I'm afraid I don't have an insight into this particular problem, I'm sorry. Not a question we know enough to answer, thank you. Yes please. I'm an anthropologist by training and curious if you thought about the limits that you would like to see emerge on interop. I'm thinking in particular of the problem in biological systems of hyper coherence, things that are too tightly wound together. For example, a single species of rice or corn that has been regarded as more efficient and in interoperability terms often regarded as a real advance can turn out with one blight of rust to have been a big mistake. And working towards interoperability in a sense was building into a whole system self suicide in some sense. So there are a lot of people doing thinking in biological systems for example, how to build in questions of self limit in order to avoid the hyper coherence and then collapse. And it would seem to me to be a great realm in which you guys could help us think about the limits we would like to see imposed on interoperability, not just the privacy questions, but the fact that we live as one species in a multi-species, millions of species system. And we are doing much too much to make it efficient for ourselves and hyper coherent and therefore suicidal. It's a brilliant question. It's a wonderful point. And we try to address it in the book some what in a chapter on diversity, which is inspired actually by biology in terms of biodiversity and translate that into the technological environment. And I think there is a key conceptual clarification and necessary when we talk about interoperability and this ties back to the different tools how to work towards higher interoperability. We do not suggest to make systems identical, to merge them, to unify things. Actually to the contrary, in our view interoperability is a way to preserve diversity, but also prevent fragmentation within diversity. So have enough diversity, but still enable communication among these diverse components within a system. Now it's extremely difficult to find out case by case, where is this optimum? And it's extremely difficult also to think about the speed bumps you want to build into such systems. Of course, very different in biology than when we talk about Facebook or any tech context, but very much yes, that's part of our thinking and we I think only made the first few steps towards learning more from various disciplines and thinking about the design repertoire from that perspective for sure. I think you've drilled right to the core of what we think is the long-term implication of the study and we'd welcome others taking it from here or talking more about the pieces we don't know. Well, if you're an ecologist and work on it, you might want to bring in that literature because the question they pose quite simply is that humans in ecosystems maximize for gross return, whereas ecosystems move in another direction all together. In other words, humans maximize for something and ecosystems maximize for something else and it's that intersection that will be the crucial question of where do you set limits? Wonderful. I'm just coming out of a conference with biologists totally with you, that's all the literature we haven't digested. That's good. Maybe two more and then we'll adjourn. You mentioned a couple of the different benefits of interoperability as user autonomy, economic growth and efficiency and I would say that those three things are potentially at odds with each other and so when you go blindly towards interoperability and you guys mentioned monopoly very briefly as a danger there, you end up losing out user autonomy to the economic growth and the efficiency factor. How important do you think decentralized systems are gonna be as we look at interoperability going forward? It's sort of related but a little bit more technical. To me it's absolutely a variation on the same theme so I can give the same answer again. We're not arguing in favor of one uniform system for instance that we only have one type of cell phones. That's exactly not what we're arguing for but the argument we tried to present and I think where you have often a significant overlap or congruence between the different policy goals such as user autonomy and innovation is to work towards that sweet spot of the right level of interconnectedness, right level of ability to mash up applications and use different devices for various purposes and to aim for that optimum. That's kind of the search or the quest of what you're working on. I think you're right that lots of these things are in tension with one another but I think there are highly complex systems where many of them can coexist and I actually think the web is a really good example of that. I think email systems, maybe the emerging social web are ones where you have high degrees of diversity. You have enough standardization that information can flow and you have a great deal of user choice and innovation. I think you actually can design really interesting complicated systems where those things do and that's this kind of sweet spot in the biological zone too. So it's right to see the tensions and that's why the topic is so rich I think. Maybe one last one on this side since we have not passed the mic this way. Sungmin has got. Hi, thanks so much. I wanted to push back a little bit on your cell phone charger example. I think it's actually illustrative. I definitely agree there's a big convenience boon to interoperable standardizing on micro USB but there might be not only a performance trade-off like a sort of modularity trade-off where you can't choose the optimum voltage or the optimum current if you're standardizing on five volts and 500 milliamps but also it may be an innovation deficit if you can't do things like have special magnetic plugs or waterproof plugs or designed exactly right for that particular type of device. So I'm just curious if you could comment on that both from the performance side but also from the innovation side since you specifically mentioned innovation as something that's usually helped by interoperability. Please. How are you? You used the cell phone, why don't you? But you're up on it together. No, I agree again. This is one of the potential downsides and costs of interoperability that it may lock you in, especially when the means to get to interop is choosing a particular standard, right? Then you run the risk that you freeze actually this standard for the future and that some variations of the standard or better solutions may not be adopted in due course. So one of the additional challenges to all of the ones we already identified is how can we incorporate mechanisms of learning into standard setting in particular and keep open the conversation while at the same time also of course benefiting from settling on a certain standard for a certain amount of time. A great case study in that area is of course cell phone standards where some countries have stepped in early on and regulators and declared a certain standard to be the standard, the mandatory standard which has led to huge innovation but then of course also creates some sort of lock-in and how can we overcome this problem? That's another governance and mechanism design challenge we haven't figured out where we're still in the stage of experimentation and learning. I think another way to see this is certain instances I think it's okay to go for something that is more standardized rather than more diverse and just merely interoperable and cell phone chargers, I think you're sort of an example of that where I don't think there's so many layers to use Jonathan's notion of generativity above cell phone chargers as there are in other instances where I would argue for not doing that and going for more diversity but I think that's exactly the conversation you have to have every time. So I might take this as the moment first of all to say a big thank you to all of you for coming. A second thank you is to the Berkman Center team and those who have worked with us on this project for a long time. We've devoted this book, dedicated this book to the Berkman Center team for teaching us what human interoperability is all about. And on a particular personal note, my closest best friend, Urs Gasser, I've been so, so, so glad to have the chance to work with this a second book with you. It is a vastly better book at least from my perspective than I certainly could have written alone. Very grateful to you so much. Thank you. Thank you so much.