Conversation about open source, complexity, km and some work

The following is an imperfect transcript of a conversation I had in Oct with Matt Kovalick about some of the work I have been doing over the last couple years.  Were both fans of Charlie Rose and there is a video which I hope to get around to posting.  This was fun, though I would like to have deeper conversations that explore some of the themes we touched on.

Matt 01:07:08:15 01:07:40:03

The first thing I wanted to start with is some of the fundamentals, some of the basic definitions that informs the work that we are doing which starts with the different definitions of knowledge itself. In some earlier briefings you talked about two different types of knowledge: one being representative type of knowledge and the interaction model of knowledge.

Ivan 01:07:40:04 01:09:59:22

In the IT industry there has been this use of knowledge management as a metaphor for explaining how you might build collaborative systems. It’s not a metaphor that I particularly like but when we are asked this question what does knowledge mean I think using this metaphor gives us a chance to establish some notions about how knowledge gets exchanged, how knowledge gets created, using again knowledge as a kind of placeholder. When we stepped in to the early development of some of our systems, knowledge management systems the questions that we were being asked and presented treated knowledge as something that could be codified, captured, put into documents, stored and then retrieved. This is a classic model and it is a model that should stay with us. It’s the classic library model of knowledge where, and later on the lectors and the lectures would grab these books and they would stand in front of folks and they would take this codified knowledge and pass it on. But I think the second part of knowledge which often has been ignored is that based on communication, on interactions with other people and on the attention that other people place on particular artifacts. This is what we called initially the notion of an interactive aspect of knowledge, that you can learn as much about what was going on by watching how people communicated about things than you could about capturing and codifying knowledge itself. And this interactive notion which has informed the next generation of the web starting with Search where you can analyze and derive products of attention such as linking structures and then begin to see what should be weighted and later on social software which says who’s paying attention to who and who’s an expert in what and how I can drive those kinds of conversations were central to us trying to use the language of knowledge management to make a simple point. Which is that there is as much in the communications interactions of people and in understanding how communications flow if not more and that is a superset of this idea of the codification of knowledge into things or documents that would later be retrieved in some kind of massive digital library model.

Matt 01:10:00:06 01:10:30:12

Based on the interaction model of knowledge is a keeping that you see in this greater concept of the web being another aspect that influences the work: what is the, how do you see and define the web and how are you seeing it affect business, education and things in general.

Ivan 01:10:30:15 01:12:23:24

It’s interesting that people now talk about the web as opposed to a web which is, a web is a structure or a network and is something that we see in many different domains. Now, explicit models of social networks, the ways that ants behave, the ways that cities get built: I mean all these things are related. Christopher Alexander used to write in a pattern language how cities would organically emerge and E. O. Wilson would do these analyses of how ants and these biological systems would interact. The web is that, it’s organic. It is an imperfect, explicit digital representation of this organic element of the interactions amongst people. This is nothing new; I mean this is not any extraordinary revelation. I just think that too often when we take a technology centric view we forget that really it is … the endpoint of the nodes in the web are always people. Tim Berners-Lee, the inventor of the web, has said this time and time again. What becomes interesting about the web now is that as it codifies and creates these explicit models, the question becomes what does it tell us about us and how do we architect those things in order for us to set the spaces and establish the frameworks of how we want to communicate and reflect in those again imperfect and explicit models ourselves to us. This is I think the role of architecture of these kinds of systems. I think what we are seeing now is the maturation and emergence of a greater kind of consciousness of the permeability of this technological thing we call the web.

Matt 01:12:23:25 01:12:38:06

And in part of that web you mentioned that the end points are people and their behavior, and so why is it important to support this so called emergent group forming behavior and what is that?

Ivan 01:12:38:10 01:14:49:24

Well, I think a lot of it is about the changing metaphors that we have for how we interact and how they affect the way that we view what is valuable and what is not valuable. Again, going back to this notion of the codification of documents. We used to create these top down hierarchies and these single views that helped, they needed to be reference points for how we dealt with information, communications and each other, but what we’re seeing now is that this multiplicity of views is economically, much much easier to create. And that out of that multiplicity of views and that extraordinary diversity of views, we are being able to understand how knowledge floats through a system, or you know some people call it memes, how ideas, you know the genetic equivalent of genes but in idea form, permeate and flow through a system. What a conversation looks like when it is not just a pre-defined codified conversation but when it really begins to reflect this kind of many to many conversation that we continuously have. And as we develop those kinds of sociological understandings as mediated through technology they tell us a lot about ourselves about we’re doing, about what our goals are, about how we align them and they allow us to interact with information in what I would claim is a novel way. And this is something that we haven’t really began yet to really understand but this is the point of designing these kinds of participatory architectures, these generative architectures, that the web is about, and about supporting that emergence, that horizontal emergence, that it’s been called a rhizome sometimes that kind of spreads and creates organically innovations, adjustments as opposed to some kind of predictive model of what is going to happen. So, in a lot of ways it’s giving us an ability to re-shape the landscape of how we interact and understand much more effectively how changes happen and how they affect what we do in a much much faster time. Again, in imperfect ways.

Matt 01:15:32:00 01:15:53:14

Key terms, key define characteristics, what they still mean and are defined as and then later get to how we are defining the capability and really for the layman try to explain what a data platform is.

Matt 01:16:03:15 01:16:59:14

With this background perspective on what knowledge is, an interaction model based on people and their conversations and what is going on on the web: there are basically two challenges, two problem spaces: one is what is happening in this age of information and how we deal with that, perhaps, the tools of today email, chat by themselves maybe not helping us solve these problems completely and then applying information age challenges to geo-political complex operations that government/military has to address. In a brief a couple of years ago you said there is an incomplete understanding of the problem space; is there any better understanding now of the problem space than four years ago.

Ivan 01:17:00:13 01:17:18:23

Is there complete…I think there is still incomplete understanding of the problem space. I think that over time our perspectives change and the challenges become something different and we discover new things.

Matt 01:17:42:00 01:17:53:20

Do we understand the two aspects of the problem space: our information age and our strategic complex operations problem space.

Ivan 01:17:53:25 01:19:41:03

The problem spaces that we are challenged with; I think we have begun to understand a little bit better the challenges and how difficult and vast they are. It’s a good thing that we didn’t understand them quite the same way four years ago, we might not have started. But, there are a couple. One, is this thing we call the web and communications in general. We’ve seen an extraordinary proliferation of that capability. And we’ve almost become unstuck in time in the sense that not only had the capability to generate communications existed, but the ability to search and archive it and for it to be created, and what this has created is a crisis of attention, people have referred to it with the notion of attention economics, a catchy phrase, but again it does lead to the question of how do we keep up and what do we focus on when we’re presented with an extraordinary amount of information. This part of the problem space applies to everyone. The second part of the problem space I think which is that in our environments that is complex operations, you’re dealing with contradictory requirements, temporal associations, that is associations that happen out of convenience given certain contexts, so a lot of context driven associations that don’t have organizational permanence or institutional permanence and how do you accommodate for communications that don’t have locuses of control and are distributed. Those two aspects I think we have begun to understand a lot better and are I think central to a lot of the design principles that we’ve tried to put in place and the experiments that we’ve tried to put in place.

Matt 01:19:41:17 01:20:18:25

Using a real world example: we have challenges ranging from a multinational response to an earthquake, to a human disaster, as a result of a tsunami. In those situations where there are many organizations involved, they also need to get information; that’s part of the operations flow. Do current ways of doing things work?

Ivan 01:20:19:00 01:21:59:13

I don’t believe that the current way of doing things works, although there is a dual challenge of understanding that it works for particular sets of people, particular points in time, but kind of holistically it doesn’t. Part of it is that in industrial models the entire idea of how you dealt with complex problems was centralization and redistribution. You see this in your power infrastructures, your sewage infrastructures, classical responses that needed centralized coordination. The issue becomes though that at a certain level of scale and participation, especially when there are contradictory requirements, those approaches are brittle, they break, they don’t work. This notion that you’re going to have some kind of hierarchy that can globally coordinate things doesn’t happen. Those things are much more negotiated, they are peer to peer, and at the same time they are imperfect. We also know from advances in science and the complexity of science that there are problem spaces that we can scope, we can identify, we can potentially in theory measure but that we can never calculate. And collaborations is one of these super massive problems, it’s a grouping problem, it’s a 2 to the n problem, so that a hundred different organizations create 2 to the 99 or 2 to the 100 possible subgroups. And we know that you cannot calculate, predefine a lot of those things. So, we’ve got to come up with novel ways to say how does this web behavior, this peer to peer behavior, this emergent behavior accommodate and find solutions …

Matt 01:22:12:04 01:22:39:20

Riffing off the idea that there’s a hierarchical model based on old industrial age principles, the way to deal with collaboration basically is to have an enterprise solution within your own group, but if I understand what you are saying if you have a situation in which you have lots of these groups trying to respond together then you can’t get out of your own box that can lead you to work together.

Ivan 01:22:40:10 01:23:19:08

Right, the questions of hierarchies and the role of organizations is becoming increasingly interesting. Again one of the key points is that I would never argue that hierarchies or those kinds of organizational boundaries would go away. It’s just that they are a subset of this larger thing that we understand, this complex adaptive system that we understand has to exist in order to effectively provide unity of effort in these contexts which is much more about harmonization of these hierarchies and their touch points and the people within them and these emergent structures as opposed to these top down structures.

Matt 01:23:21:07 01:24:03:02

You talked a little about the background of knowledge and the problems of information in specific situations and therefore about 2005 started creating a possible solution to address and described some architecture with some design characteristics basically. Unless you want to respond, I’d like to go through some of those characteristics and then quickly what does each one mean to you. And so the first one, what is modularity.

Ivan 01:24:04:09 01:24:27:07

Well, the notion of modularity and there’s great literature on that out there is this idea of a box in a system that would allow it to behave more effectively and that this is based on a standardization of interfaces that would allow you to spawn off a richer style of system.

Matt 01:24:27:10 01:24:31:03

And we can wrap them all together at the end, but another one: simplicity.

Ivan 01:24:31:03 01:24:46:13

Simplicity is really about complexity hiding. Which is to say that at some point technology should have all the properties of magic; it should just work. And the complexity should only be there for the people that want it and not for your average user.

Ivan 01:24:47:22 01:25:02:20

Diversity says that strength comes from the avoidance of a monoculture which leads you to weakness. And also an acknowledgement that diversity fills niches that you cannot fill….(cut off)

Ivan 01:20:59:10 01:21:16:07 (CAMERA B)

Again this has to do with the aspect of modularity but standard script and early protocol and things speak are a key part of this as well as the notion that these modules are objects, need to carry with them their own DNA, the information that describes not just what they do but how they’re built.

Ivan 01:26:20:11 01:26:32:23

The notion of embrace and extend is that if you think about systems as being evolutions you don’t want to get rid of the models that existed before, you want to embrace them and bring them forward.

Ivan 01:26:46:03 01:27:09:24

This came from the notion that too many of our technologies right now demand your attention based on the creator of information as opposed to the consumer of information. It’s actually borrowed from Dr. Dave Alberts and it says that architectures where you allow for a filtering system that is negotiated where somebody posts and then somebody has to define the context for pulling that information would be more effective.

Matt 01:27:10:06 01:27:42:02

And so these design characteristics were part of the initial architecture of a capability that is sometimes described as a data and communication brokerage platform or a data sharing rendering framework or a multi-modal platform. How did you get from design characteristics to a function capability.

Ivan 01:27:42:11 01:28:21:04

Well, these terms were architectural principles that formed the framework for us to then address the question of the codification of knowledge and for us that meant data and how it would exist across this network that would include systems and people. And based on those characteristics we made a series of assumptions and looked at existing systems and said “ok” here’s a model for convergent communications and how it would flow into social architecture and what it would mean for the codification of knowledge.

Matt 01:28:22:01 01:28:25:19

What has been built?

Ivan 01:28:26:04 01:29:15:21

In a lot of ways what we’ve built is based on what the web has built. A lot of people use terms like Web 2.0 and or interactive operating systems; there’s all sorts of terms. But, really it’s about looking at the web and web architectures as communication platforms as web architectures become the point of convergence for almost all of our communications and as web architectures become extended beyond laptops to all sorts of devices. As the web permeates everything and other networks and as communications converge around Internet protocols and the web how do you then enable using social architectures and all the assumptions that we see in web design patterns into a common communications platform and capability.

Matt 01:29:18:00 01:29:44:16

And so on this term service oriented web architectures or distributed infrastructure; what does being a web native capability or a web based platform entail just in very basic terms.

Ivan 01:29:44:19 01:31:00:11

There is a new term to address this question of what is a web architecture and people are starting to talk about web oriented architectures. The notion of service orientation used to be that particular capabilities on the web or in an architecture in any information architecture could provide a particular specialized service and that they could describe it and that others could consume it. Now what those services are is an open question that’s one where you say: “what is the value with what I do with this particular system.” Is it expertise service, is it a document service, is it a financial service? At what level of granularity should it exist? Our point of view is that the web has been the best example of how loose coupling and services orientation allows for what people would call now mashability. And then the question is what do you want to mash? If you can make everything mashable, if you can make every data source mashable, how can you re-combine, re-compose, re-create things in order for you to consume these services and create something that’s valuable to you or to your organization as part of a conversation about that data itself.

Matt 01:31:01:00 01:31:16:03

So, in some sense you might say that maybe you haven’t built anything but instead they’ve found ways to integrate and combine the best of what’s out there.

Ivan 01:31:20:07 01:32:57:12

I would say that, what have we built, I think that we’ve glued a lot of things together to propose a model of integration that promises the future, well that promises what the web can do. In that sense it’s a little bit of a different approach to building, where instead of looking at things the way somebody might look at them on a cell phone, with a list by list, you look at them on an i-phone and you say that the sum of integration is greater than the individual part pieces because of the way that things flow in and out of each other. And in taking advantage of those relatively simple things, for example, location and social networking so that I can know which restaurants my friends recommend in which places when I happen to be in that place is what the web is starting to bring to us. And that is what is unique about the flexibility and modularity of the web that allows itself to be glued and re-combined to create very individual niche services but also potentially very large services like Search. And so web architectures are I think in this way kind of fundamentally different to the traditional enterprise architectures that we’ve dealt with. And all that we’ve been trying to do is bring these concepts and a tool kit of these web architectures into enterprises so that they can begin to deal with this notion of information turned inside out and this kind of need to share infrastructure, this mashable infrastructure, that we’ve seen has been a real challenge and a bridge too far for most enterprises, especially our own.

Matt 01:32:58:08 01:33:26:10

I guess continuing on that note for enterprises that have need in some cases, the goal to protect information or to keep things bounded, that’s been our situation to a certain extent, what have been the challenges to this inside out approach to DOD related and government related….

Ivan 01:33:26:13 01:34:40:18

Well, when you turn your information inside out and you participate in this kind of organic data commons called the web and you try to draw boundaries it presents all sorts of challenges. And I think we are only just starting to scratch the surface. But there are questions about identity, about security, about classification and they all matter. This is where we, honestly, when we define the problem space, are starting to see the real complexities in how moving to a need to share and participatory kind of open system, open collaboration model, changes the internal DNA of your organization and begs the question of what’s really truly important to keep private, what do you have more benefit from keeping inside than from putting outside and how do you move from this kind of fail-safe notion of information assurance to this probabilistic notion of mission assurance when the need to share outweighs the need to protect and if that happens where is the balance and mode of interaction. So, all of these questions are enormous challenges that we will have to kind of inevitably face in the missions we are being asked to participate in.

Matt 01:34:40:23 01:34:56:06

If we weren’t asked, meaning through strategic guidance or policy choices to take on different types of missions, would be talking about this stuff in the DOD context at all?

Ivan 01:34:56 01:36:14:20

If our particular mission set didn’t presuppose distributed, non-centralized collaboration, right, this need to engage multiple unanticipated partners at times, I think it’d be harder for us to be having this conversation but I think we’d have to have it anyway. The truth is is that at certain sizes the complexity of organizations kind of again asks the question of how do you best organize yourself internally. And industry is starting to do this, right, I mean Cisco when they adopt Web 2.0 behaviors and Google, when they allow these 20% projects and communities of interest, they all imply this notion that emergence and informal collaboration need to be dealt with much more effectively to really leverage and transform, re-invent enterprises. So, I think that the conversation would be happening; I think it probably would have happened much later because that act of recognition is a much more difficult one than dealing with this particular set of missions that says: we assume, going in, that there will be no centralized and key body.

Matt 01:36:21:08 01:37:08:06

This architecture, this problem space, has yielded, I guess we would call it, a manifestation of all these concepts and there is a capability, one may call it Scholar or there is a capability that has been created. Now, what’s the difference though between this platform, or what aspects are different than just a portal on a website or just something that you see, kind of re-treading some ground, what you see on a website, how do you succinctly explain that there’s a lot more background to the capabilities.

Ivan 01:37:08:11 01:39:04:14

Well, I mean to come up with an explanation of the difference between an architecture like the one we’ve put together, a web native architecture vs. your classic enterprise architecture or traditional portal architecture, I think you have to ask the question, I mean it goes back to what is a portal, an entryway into information. And then if your assumption is about how you enter into information and what the role of information are fundamentally different are than your architectures are going to be different. Again, if you look at most traditional portals in enterprises the place that they start is the organization chart and they end up codifying the very kinds of stove-pipes they are trying to break through collaboration. Our model is much more individual-centric and data-centric which allows for this emergent structure to occur, by emphasizing again explicit models of social networking. And when that happens it allows the question to be raised at the senior levels, this is the way that my organization is telling me it behaves, how do I then re-organize the org chart to take advantage of it, how do I formalize those informal structures as they deal with problems as opposed to trying to enforce the kind of top-down organizational hierarchy that I assume is going to work. Now, both need to happen. But the second type of system, the type of system, this web-native system that we built, is this open, participatory system, allows for a different set of qualitative questions to be asked and a different set of qualitative behaviors to take place in a formal model that don’t happen in traditional enterprise systems today or if they do happen they happen inefficiently and out-of-band. That means that they’re not available for the enterprise to really leverage as a resource. Trapped in emails. Trapped in phone calls. And so enterprises don’t effectively support this many-to-many conversation that web systems like ours are hopefully are architected or intended to support.

Matt 01:39:11:24 01:39:54:03

I’m trying to figure out how to ask a good question because in the past it’s been called it’s not really a fair apples to apples question, it becomes an apples to oranges question, when you start trying to explain the capability that’s being developed and worked on right now with other tools and applications that people use everyday. And so I guess we can go back and forth on what aspects can you compare and contrast between what you’ve been working on in terms of capability and with let’s start with Facebook.

Ivan 01:39:55:21 01:42:00:14

Let’s take a step back before we talk about Facebook and let’s look a little bit at a key transition which is that Microsoft and traditional enterprise architectures that focused on personal computing and a combination of personal computing with classic enterprise computing that was represented by IBM ideally, this centralized managed systems, are now facing a new kind of challenge which is there is this technology now that allows for personal networking and for managing networks of data very ineffectively today across a series of services both inside the enterprise, although not as much, and outside the enterprise. Facebook is a great example; Linked-in is probably even a better example in a professional context for a certain generation where you establish and you have an identity online that allows you to create associations, manage events, do all sorts of other things. Those capabilities that individuals more and more especially of a certain generation are looking to leverage, whether it’s to organize my music like Last FM or organize my professional network like Linked-In, organize my social network of friends from the past like Facebook, and a whole series of other ones that are beginning to emerge and become much more adult in their output are the kinds of behaviors and patterns that to date have not really existed inside of government or inside the enterprise. The whole conversation about government 2.0 and enterprise 2.0 is about what about these behaviors is applicable or worthy to what we do inside. And so this is again this notion of how do you deal with personal networking and data in the context of an enterprise and outside of the enterprise where those things matter. Our tools are really about giving the enterprise an ability to have a lens of that and a role in that, which it hasn’t had, and giving individuals those same kind of tools and capabilities that they seem to be embracing in the open Web.

Matt 01:42:01:10 01:42:13:15

Another example, a slightly different twist than Facebook, but something like what Google has put out in terms of applications accessible on the Web.

Ivan 01:42:13:19 01:44:08:04

Well, again, Google is an example of more and more computing moving toward services on the cloud. For enterprises, however, that is sometimes not really a choice, and for us especially when you talk about issues of data custody. But, when Google and other folks begin to talk about things like open social or this concept of the social graph; this is really the question: how do you manage which kinds of services are individuals going to use both in the context of enterprises and outside the context of enterprises to manage their data as it exists out there. And what is it going to mean to them and how are they going to leverage it. And the truth is that enterprises haven’t really had tools to do that. They don’t know other than completely informally who’s on Twitter, who’s blogging, who’s forwarding emails to folks outside. And they don’t know what’s good and what’s bad; they don’t know how to adjust to it necessarily, they don’t know how to manage it as part of their collaboration. It just kind of happens. And there are no explicit models, governance or control. Those are all needed. At the same time individuals need to be able to permeate those things because they will find ways to do it and need tools to better share that information with others and better tools to filter the noise from all of this massive proliferation of services and sites and things that are out there. Those tensions are hopefully the sweet spot that we are trying to address with a particular model and architecture that says: here’s the role of the enterprise, here’s the tools for the individuals, and here’s how these things interact in a permeable world, not in a mote driven notion of enterprise computing where you drop the drawbridge whenever you want to go communicate and share data with other folks, but in this notion of cities establishing commerce where individuals are the carriers and ultimate unit of participating in that commerce. Same happens for an information economy.

Matt 01:44:08:20 01:44:40:18

Just one last quick example for our DOD setting, in working with more and more types of actors, and beyond just military to military affairs, we’ve tried to develop something like a partnership network, in some cases a harmony web, for example, as an example can you draw comparisons and contrasts between these approaches to collaboration with other actors.

Ivan 01:44:41:07 01:45:47:09

Well, I think it is really important to address how these technologies are built in that the open processes for open technology development are a key part of this in developing partnership networks. And the reason for that is we’ve often had you know this prescriptive model for dealing with people that are not like us, in terms of we will build something and you’ll love it because it’s the best and you will use it, and the truth is that isn’t an effective way to deal with local context and that the participatory part of emerging a solution with partners is important so that open collaboration has to extend in my opinion at least to the building of the systems that people are going to use. It’s a process that has to be normative or it will fail. So again as we build a partnership network it’s about building into the architecture the modularity, the capability, the extensibility, that allows our partners to build it with us or we with them as opposed to kind of giving them a prescriptive solution.

Matt 01:45:48:03 01:46:33:09

I guess segueing to development of the capability, one could say that that’s just as new and different, the means of developing the capability, as the capability itself. I mean that’s sort of been the challenge to explain to people: not only is this a different approach, but actually the way we create it is different as well and therefore wanted to address, we can go back to the social production or the partners but I first wanted to address the idea of open source and how the open source development process and how that has impacted the capability, and what are the benefits and challenges of, and what is open source.

Ivan 01:46:34:07 01:47:46:00

I mean, that’s a huge question. A lot of people have different opinions about what open source is or isn’t, but there are a couple of things that open source does that you can recognize. One, as you mention, is the notion of it being an example of social production so that open source which started in software is extending to all sorts of other things in order to do kind of sustainable social production and a very different style of social production. And this is a term that I think Yochai Benkler put together in this book, The Wealth of Networks, a very interesting book. The second aspect is that open source and this term of social production leads to something that another Professor, Eric von Hippel, talks about that is democratizing innovation. And the notion that in these kinds of open collaborative, open source style systems, you have democratized innovation, a very effective way of codifying that innovation and a social production capability that allows it to be sustainable and participatory is really an interesting and different approach than the way we’ve traditionally kind of developed products over time.

Matt 01:47:46:15 01:48:04:09

And so I guess maybe bringing it from the theoretical aspect of open source to the actual process you’re involved with day to day: how does the open source development process work for this particular …

Ivan 01:48:04:10 01:50:12:11

Well, for us, and it’s very interesting to try to measure the network externalities and effects of community driven participation, but for us the open source development process was about developing externally, it was the same as this kind of notion of collaborative process, the same theme of turning your processes inside out, and engaging with people outside your organization to jointly develop capabilities. And so we very early on identified communities, engaged those communities, and took on the social aspect of saying: hey look we want to be a partner in this development, how can we do it. These are the resources and skills that we bring and let’s be very open and honest about how to negotiate those social aspects as we develop technologies together. And, it’s been great. For example, just the other day I saw a data-preparedness initiative by USAID that worked with someone doing DRUPAL development, one of the communities with which we developed. And they used technologies that initially we contributed to the community. This same group has contributed technologies that we use, in particular the SMS framework, in the next generation of our architecture. And I think that shows this kind of effect because without our participation they would have been worse off, without their participation it’d be much more expensive for us. This is, I think, open source, open collaboration, the way that it is supposed to work. Now I think it is very important to draw a line between that social model and the legal licensing models that usually get associated with open source. Those are related. They’re important. But, those behaviors don’t necessarily necessitate some of those things. I mean, it would be possible, for example, I would hope to have those same kinds of behaviors and ecosystems around proprietary software that doesn’t have these licensing capabilities.

Matt 01:50:14:00 01:50:48:01

The idea of with this type of architecture that’s been created is the open source development process the only way to be able do it effectively and efficiently in terms of cost because of the fact that you’re trying to reach out and grab what’s the best and plug in…getting to the point of the cost effective nature of developing this

Ivan 01:50:48:06 01:52:24:20

Well, I think there’s a lot implied in the notion of what is cost effective about social production and what the actual benefits are and I think there’s a lot of work that has to be done there. I would say that what it has allowed us to do is create partnerships and return value in very rapid ways. And the other thing that open source has allowed us to do because it is an open environment is negotiate with and find talent and expertise very very quickly. It’s opened up a global market for us in terms of who is talented and it’s allowed us based on negotiating the interestingness of the problems we were tackling to engage the expert. Which is something that has traditionally been hidden a lot from projects of our kind because usually the expert is guarded and kept in a proprietary place building for their enterprise and there hasn’t been this kind of interaction between the expertise for example that’s built the core of the operating system and the public at large that’s possible in open source. So, those benefits I think have been extraordinary. They are very difficult to measure and this is one of the things that leads to another question about this problem space which is how do you measure them and how do you make sure that they are being effective. And we need better models for understanding what that means so that we are not just making these choices based on anecdotal or intuitive evidence but that we’ve actually looked at what it means to measure and model social production and open source capability. We’ve started to do some of that work as well.

Matt 01:52:35:12 01:53:01:24

Strictly one or two things, one is, what is the latest result of the architecture of the development process into the latest release or the up and coming software release. What distinguishes it from previous attempts and what’s out in general, what’s out on the web in general.

Ivan 01:53:02:03 01:54:44:05

Well, I think, that our latest release if you look at it closely presents a miniaturization of the web. That is, that a lot of the capabilities of the web in a toolbox. It has things that allow you to bridge to cell phone networks and SMS, Twitter style capabilities. It has an ability to plug into geo-positional information location based services. It has this ability for social networking so that you can use your social networks as filters. It has this ability to plug itself into distributed data sources, spreadsheets, Google spreadsheets, other structured data on this open-linked data model so that you can create and mash things in a very flexible way. And really it allows for three different kinds of integration to happen I think pretty effectively. One is at the application level so things like Google mashups that say give me all the restaurants, restaurant location capability, or in our case a field hospital location capability. That’s an application style. Secondly it’s a data integration platform a lot like Yahoo Pipes that says everything is mashable and this is the way that you can create your data that way. And the last one is this infrastructure mashability that takes advantage of cloud based computing which is to say I can take my infrastructure now and distribute it over the cloud as well. Those capabilities which again represent the best of the web and all these ecosystems that are generating these kinds of solutions I think are codified and brought into this tool kit in the next generation of our software and so now the question is going to be how do you use that tool kit for specific domain spaces. And that’s kind of the next series of challenges that we face.

Matt 01:54:44:12 01:55:01:21

That’s a good point and it goes to the idea of adoption and usage. Will, can, are people able to use the technology now? Or is it too ahead of where people are at?

Ivan 01:55:01:24 01:56:33:00

Well, so what are the challenges of adoption? They’re a lot. And at each of those three different levels there are challenges. I believe at the application level we’re starting through feedback and through seeing a lot of these tools like Facebook emerge start to be understood much better. At the data level I don’t think that that’s quite there yet. But the networked world of data is I think emerging and really to end users it’s a question of this kind of complexity hiding, of simplicity of the system. And then finally at the cloud based computing layer there are challenges but those are probably more for a professional class of folks as they understand what it means to be able to think differently about infrastructure. The challenges exist and there are two ways to address them. One is that I think that the world at large is starting to mainstream and catch up with these things we’ve been talking about for the last five years and has been called 2.0 or whatever. The second one is as that happens the geeks if you will are working very very hard I believe to simplify the technology so that you get to this tipping point where it’s simple enough and mainstream enough for these models to be incorporated. And I think we are starting to see both those trends at every one of those three layers that I’ve described.

Matt 01:56:33:10 01:56:54:11

And then finally what’s next. I think you’ve touched on two things: the emergence of more data, the importance of data and the mashability of data in the future; and the emergence of cloud based computing. I guess what’s out there not just six months but further.

Ivan 01:56:54:12 01:58:11:05

Well, technologically or socially? Well, technologically I think that we’re starting to see an extraordinary set series of innovations that are going to be there. I think the I-Phone begins to lead to a model of what is going to happen with say location-based services and a whole series of other things. So technologically I think mobile is going to become central and very very important. I think you’re going to see the desktop begin to be less important and that mobile applications and kind of this notion of data in context is going to be very important. But I think that socially a lot of the things that we’ve talked about are where the real questions are going to be. What does privacy mean? What does anonymity mean? Who owns what? What does copyright mean? Who owns what pieces of data? And all of these questions area also going to need codification and explicit models as we move forward. So, there’s going to be this interplay between what the technology brings and how we codify a lot of the open questions that the changes in technology are leading us towards.