 I have a message for you on your way from outside. I'm sorry. I was talking about that. I was talking about that. Today from your home. Good afternoon everybody. Thanks for coming. This is a 90 minute session about the question of open source procurement for Law.gov. As many of you may know, Law.gov is an effort to establish a principle that the primary legal materials of the United States should be broadly available. Should be available in bulk and should be authenticated. And there's been a series of workshops around the country starting at Stanford Law School in January. We've done 11 workshops so far at many of the major law schools around the country. Most of those workshops have been legally oriented. They've been issues of intellectual property restrictions on primary legal materials or privacy protection and our courts and even financial issues. And our hope today is to look at another question, which is what if this requirement were in place and every city were required to publish their municipal code in a clueful fashion. What would that take? And more importantly, could there be a reference implementation built on open source software that was available for vendors to perhaps enhance and then support or even for a city to download and use? In the Internet world, when you make a standard for electronic mail in the Internet task force, you don't simply put a proposal out that says all electronic mail shall be as such because you get laughed out of the IETF when you do that. And what you do instead is you have a reference implementation along with your standard and that proves that your standard actually works. And before you can become an actual proposed standard, you actually have to have two interoperable implementations that were done independently and they have to talk to each other properly. And so we're kind of here to discuss that very general principle of could open source software be available for the publication of legal materials. So think of it as a content management system for legal and the exact requirements of what that means, I wouldn't worry about. We had a two-day workshop at Cornell with many of the leading lights of the legal technical world that looked at issues of what is the XML format and what metadata should be required and how do you do authentication. And so we're going to kind of take a step back and our hope today is to look at a couple things. So we're going to start with Jim Onstaddil from Accenture, who is one of our top conveners and he's going to talk in general principles about open source software and the government can it be open source software. We have asked John Scott to talk about how the Department of Defense talks about open source software and procurement. Gunnar Hellexen from Red Hat is going to talk about the European Union and how it does open source software. Brian Bellendorf is going to talk about healthcare and open source software. He's been deeply involved in the HHS efforts and there's been similar efforts in veterans. And then we're going to sort that clean up here and kind of talk about some of those general principles and open it up to be more of a general discussion about what would this take. We are filming this and as we are many of our other workshops. This is the first of what we hope to be a whole bunch of meetings about this general topic. These meetings might end up being in some governmental entity. We hope that at some point this goes over the wall and the government starts to take over how do we publish our own legal materials. So Jim, I'm going to turn it over to you. Sure, so in terms of the question of open source and government that's a really, really broad conversation and I'm not attending to try to tackle all of here. What I was going to try to do was just a few comments about describing a generic pattern that might apply to the kind of situation you describe. So the sort of situation you describe is lots and lots of government entities that range in size from the large federal government entities down to local government entities and it's sort of describing it that way implies an adoption orientation towards which we're trying to achieve. So you can't just sort of, as we all know you can't sort of just sort of mandate and assume that everybody doesn't. There's got to be a high orientation towards making whatever you do accessible and easily adoptable readily adoptable. So in that context I think it's broadly applicable to talk about open source in that. But so a few things you already mentioned, Carl, about the approach to developing standards and how that sort of kick off to anything you might do with this. I think I just wanted to kind of comment that that is a fairly broad conversation itself of the semantics you alluded to with what are the XML schemas or those kinds of things sort of figuring out what the language is and all of that and then there's certainly diving down from there into systems of standard, systems of ability and kinds of standards. How will they actually connect systems together and get done by each other? I think those are actually potentially two separate conversations, potentially two separate standards of attempts if you will. But probably different things in the end. You already also mentioned implementations or reference implementations. I attended a talk by Brian this morning about production ready reference implementations. I think that's a really important thing to add into something like this. If what you're trying to do is ensure that organizations can pick something up and run with it, having it be a production ready implementation rather than an implementation that should demonstrate the standards but could never be performing or operating anyway is a different thing. So certainly striving towards a production ready implementation is appropriate. The other then going from there into whether or not it should be open source and or why and if it's a the or a them is another thing. I think actually I don't think we're talking about a single reference implementation necessarily. I think what we're talking about is building on top of things that already exist and even as we do that this notion of maybe a continental system of law that actually has enough components that we probably want to think about those in an ongoing way as separate pieces rather than as a model. Also that they can innovate independently and kind of progress independently and also to kind of maintain choice across that system as it progresses. But if you think in terms of developing these reference implementations I want to do the motivation for doing those open source in the context of thousands of entities that you ultimately hope will use this stuff is basically the same commoditization strategy that happened with something like Apache and a web server where you're trying to make sure that it's as broadly adoptable as possible from a cost perspective and from all the other reasons what people care about open source because looking inside the code etc I think there's a when we use the word open source we sort of use it to mean a lot of things. It's a very new speak kind of word right. We sort of take this vocabulary and we brought it down to one you know one thing but there's actually as we all know there's product oriented open source or community oriented open source etc I think this pattern I don't think it really has this name but I think an interesting way to think of this pattern is almost like a four double joint venture. You see this in banking for example with things like AMQ and AMQP you know where get together and they say we need to transport broader call for financial information we don't want to continue to pay vendors high prices for this stuff so we'll get together banks. We won't do it the old way which would be a joint venture because it's difficult to sort of sort out the legal stuff up front so instead we'll rely on forking to make it easy to create that kind of relationship we'll do the standards first then we'll build a reference implementation and then in low and behold a few years later we actually have probably five viable competing implementations now of AMQ messaging system out there and they compete based on different kinds of performance metrics or different use cases etc and they've actually generated an ecosystem of businesses that make money on Wall Street delivering that stuff as well as a services ecosystem around it I think that's a really nice model for sort of what you'd be trying to achieve here you want to kind of kick start the development through these reference implementations of a almost a distribution a distro of a system that looks like the kinds of things you want but then allow people to build on top of that to build businesses etc because the tipping point often comes when the people that are selling to your end customers the people generated people briefs etc when the people selling to them are bringing with them the notion of using the stuff that sort of tipping point tends to come so you have to think about that part of the ecosystem so with that you're going to want to choose ecosystem-printing license and this license that will actually let people do that now when we talk the other day briefly about this one of the things we talked about is there a relationship between this open source story and an approach to how you leverage open source in this environment to sort of all the talk that's going on about cloud computing as well if you go back to the sort of level setting at the beginning which is try to make this as adoptable as possible as broadly as possible creating an open source project that is software as a service friendly with multi-tenancy and things like that built in from the beginning has licenses appropriate to supporting that model which is really valuable because then when you develop your ecosystem or when your ecosystem a provider springs up around this hopefully some of them will actually be cloud-like providers that leverage the software you're using but by virtue of the fact that you made it open source up front you keep the potential for decoupling that value chain always available so you always should be able to move both inside your firewall back outside your firewall as your needs change that sort of thing ultimately you can start with external providers external activities you can move into larger providers you can do things like that by sort of standardizing at but I think that sort of like at every step of the way you want the thing you're trying to create to be as approachable as possible and as readily adoptable as possible by those and certainly for a small municipality a runtime permission cloud service that requires nothing from the IT department up front will be much better than building source so I think that's a really good point of remarks there I think the one other comment that I would make is that from the beginning I think it's important to think about this as a process that will be forever in this equilibrium think about it as incremental not as an instate that you're thinking for so both the standards bodies should start with the things that are most important up front and then move from there and the implementation should start with the simplest things that get the most back to the buck up front and move from there and that what you would hope to achieve as the incremental process was happening was not only an ongoing increase in the functionality available to the users of the system but sort of seeing the breadth and scope of the community developing and evolving as that happens as well Well that's great I love these workshops you just learned so much on these different topics so Department of Defense I arguably is the best at procurement because they do the most so maybe oh my maybe they do the most totally untrue I think if you look at the quality of the quantity line that's one great thing about DOD is it's such a huge organization you can find pretty much every example of how to do and how not to do it out there so I guess a little background about open source and DOD you know it's always been there from the early in stays and everything else I think that it really got a lot of play about what it was in 2003 when a certain entity tried to push through a memo on DOD-CIO about the media for this open source software at DOD and people at DOD said well how much open source do we have and so they just went out and did a little survey and it was just an email survey, not really too scientific just email it out, blah blah well if you want to just destroy the entire defense department then they get a little bit sourced it stopped so the memo turned into one of like well it's okay to use open source and so since then we've had Department of Defense memo come out and make clarification that open source software it meets the criteria of commercial software we had a memo come out last year is what open source software is, it's okay to use you still have to go through the normal wickets to kind of use it so I think we're in the mode of you know letting people understand it's okay to use open source software and working it from the outside so it's kind of the first step of moving along with what we're actually using now we're in the part of like well how do you actually interact with an outside community you know patchy likes and actually push bugs out there you know with Red Hat does a lot of that for what it's for the military you know I see but for other packages or applications you know what are the rules and regulations around kind of interacting with an outside community that you're not paying so you still have it seriously or still so I think we're kind of getting into that other mode we're nowhere near you know that point where we're kind of at least the people who really want to do it at least now have some good ammunition to kind of run around and say we can do this now I think in DoD the case where we've seen work best are a lot of the people who want but just can't pay for they can't pay for it because the existing capabilities, the existing system you might buy are just so expensive that they still have it in their budget so usually what happens especially with the other generations coming out in the military and the civilians they know how to build software or they're not intimidated by actually going out building it themselves and how to bring it in especially some people need it more maybe we've got to talk about that since history has done great work out in the field because they had access to to actually being able to get inside the source code so I think that's where we see a lot of it I think also DoD's been good to kind of be outside being inside that's probably where the toughest model is in terms of classification issues and I see as well and the other thing you should also keep in mind that the intelligence community uses a lot of open source software that made a conscious decision about we're going to go down that route because we have to just period stop and we have to be on that route I think that the next thing for the military is really how you can develop and publish open source software so for things that just doesn't exist out there in the marketplace how you actually build, run communities, there are a bunch of experiments going on right now there's 4.0 which has gone along the inside and kind of started to run and I taught and distributed that source code there have been a number of various government agencies just for example the Department of Energy allowed the program managers to publish to create projects as open source code for the people who have paid for them so it's interesting a lot of the source code that went into that migrate over to IH4 like Proteomics and the apparel of the beauty came out of DOE labs just basically the publishers open source basically I think it's called the CAID project which is basically taken off and the funding goes back to C funding that DOE pumped out so there's a really good language there that I can send to you about allowing the program managers to actually require the contractors to put in an open source license as a community I think the one thing that we really push in DOD is getting the communities together first because I think a lot of people come to this like here's a technology solution when you go out and do this beyond in my mind the first thing is you really have to get the people together who want because they probably know more than you about that solution they probably just already exist out there and first I'll tell you what's really important day one not something else you know like Jim's point about finding a little hanging fruit you know just having something as simple as maybe need contract identifiers for a contract it's a great value versus having an entire content management system for RFPs just having that unique thing and fixing the accounts so I think finding those simple things that everybody really really wants that you can land last together really quickly you know and it gets you towards that reference implementation I think from the government side of government funding source code development you know a lot of wickets to kind of go through in terms of getting you actually need to engage a hundred to build your system actually putting in special language in RFP you will publish this as a software code as a community Brian can probably talk about that because it's just the idea of IP stewardship there's not there's lots of government has limited and unlimited government use rights around things they build but even the gubbies don't understand how to do it and the gubbies do understand how to do it but then government for 30 years and they think they can work for a law firm or they've been there for 10 years and think they can work for a law firm that went to a and it's kind of a spy or a paid one to a legal seminar about protecting your IP and they walk through all the various ways you can protect your IP and one thing is you know file a ton of before you sign the contracts file a ton of informational patents and you've got 18 months to basically go through and add in whatever you want copyright everything because eventually if you put copyright in on enough stuff it becomes your whole hope so I think that's kind of one of those first design patterns kind of thing you really have to kind of take a look at and be upfront rules like here's how you really want to treat our IP with the community and I think that then you'll end up down the road, it's kind of like staring at the sun you want to stare at the sun you want to kind of get the corner of your eye about where you want to get to and get the community together first so I mean the militaries I think you know they're very good partners in the military they're kind of going down the route of using a lot more of this function a lot of it is because they where they need to build stuff I think in this case less money is actually better because it's worth the attention going on European Union yeah sure so the European Union is is actually about the same place where the U.S. government the office management and budget and where the DoD is in the sense that commercial software open source software is commercial software and should be evaluated on the sun now some member states are a little bit ahead of the others maybe too far ahead asking for preferences for open source software and open standards, that's a highly contentious issue and I'm not sure if we actually want to direct it here but the interestingly the open source observatory and repository which is a kind of research organization inside the EU published a set of open source procurement guidelines which it was a funny to read a 64 page document that sounded awfully similar to DoD open source method so good on the DoD for actually out running the EU we got it in three pages but it took the EU 64 pages but the we do have an 850 page constitution so wow that's right this was only in one language I don't so the OSOR open source procurement guidelines were very interesting they start out very much as the DoD memo now which says okay yes open source software is the same as commercial software you have a set of requirements you can evaluate them you can evaluate them equally and then it spends the next 63 pages telling you how to engineer a procurement to ensure that open source software is used which is pretty fascinating which the DoD does not have so you did not get quite that far yet and so the so this is an end environment though they say this up front 19% of the software that the EU buys is commercial software 52% of it is custom developed or done by contract by someone else and 24% of it is developed completely internally which I don't know what the numbers are for the US government but I suspect that they are inside out the government actually has a depends on how you measure what has an express preference for commercial off the shelf software and I would hope that more than 20% of the software used by the federal government in terms of dollar spend the government will buy a lot of closer software and then pay 500 bucks an hour to change it so what does it become the only reason why I mention the statistics is because it seems they themselves mention the statistics because this is an acute problem for them they are looking for a way to but encourage more commercial off the shelf software and also figure out a way to not be saddled with the maintenance headache of the software that they've paid to develop or develop in-house so they mentioned the usual benefits that we've already talked about that cost the transparency and one interesting point that they spent a great time on is the sustainability of open source software and this was due to a few attributes that are unique to open source software which is an argument I think we're all familiar with the second was the interoperability they actually make the claim that open source software is more likely to be interoperable than a proprietary solution which is another argument I think we're all familiar with there was also this the characteristic of independence of open source software that is my ability to actually tender out the maintenance of a piece of software to multiple vendors and not be inadvertently locked into a single vendor just because they happen to be the guys who have wrote it in the first place the chances that I could get a marketplace or an ecosystem of software is a much better chance of doing that if they were to present factory open source and finally the flexibility they are more likely to be able to have software that responds to their needs if they themselves can change it or if they can find literally anyone to change it on their behalf so the sustainability arguments was interesting and not an angle that I really elaborated on in the DoD memo or anywhere else in the US we did try that we did try to have it and I think for the DoD the model is much more around the agility of what you can do all of that in terms of just being able to respond and build new systems to not be locked and that's the real value because we have enough money and so the rest of the memo discusses and I believe it is framed around the fact that we have rules on books that say that you are not allowed to mention a specific vendor when you issue a tender I'm not allowed to say that I need semantic software I'm not allowed to say that I need hardware that is Intel or equivalent I can't even get that specific there is a very narrow set of rules around when you are allowed to mention a vendor specifically and instead it's better to describe a set of requirements and then allow the market to figure out all of those requirements the process we are all very familiar with and also very familiar with how this gets gained which is why it's interesting to have 64 pages of this document to tell you how to gain it and the author is actually to say here there is no obligation on the part of a public body to adapt its requirements to the business models of particular firms which is a rather strong language for coming out of the EU but Anika, again so they talk about three specific requirements that you can include under procurement that would compel the respondent to include open source software as part of the solution and the first is having a transparency requirement open voting is probably the best example of this if everyone is going to procure a voting system it wants to have some assurance that it knows how that voting system works otherwise it is simply not as good a voting system otherwise another requirement is for redistribution want to be able to procure the software once and then be able to hand it out to other agencies once I've solved the problem or for that matter distribute it to the public once I've solved it the third requirement that you can include in that tender is the ability to alter source code and you can do this for avoiding the lock in for encouraging new markets around a particular piece of software so if the government is doing something in particular cutting edge it wants the ability to allow many vendors to pop up around this particular problem space and so you can actually include that as part of your procurement that this would require then they discuss the importance of standards and Jim and John both touched on this and they talk about the tricky everyone agrees on standards everyone agrees that open standards they define open standards and as we've seen in the hurricane interoperability framework it is impacting possible to define an open standard like me and so the author suggested instead let's focus as we did on the open source software itself focus on the characteristics of an open standard so I would like to cure something which implements a standard which is transparently open I need software that implements a standard which is available open networks and so in kind of engineering and describing the attributes of the thing that you want rather than saying explicitly I need this to, my open standards are I need this to implement open source software you can actually guarantee that first of all you are improving the procurement process itself because you are in fact following the rules and this is what you actually should be doing in the first place and second of all you're creating a level playing field on which open source software can compete so I will summarize by saying that it was interesting to see that the European Union would and obviously there is diversity of opinions in that large body but to actually have a document that comprehensive and it does all but recommend open source software for most software projects and clearly they are encouraging especially for government use and most especially for purpose built software that will be run as open source projects Okay, great, thanks Brian, your experience not only in healthcare today but certainly in the Apache Foundation seems somewhat relevant to this discussion I thought I would start by talking about the project that I'm particularly focusing on but also talking about its history long before I showed up because this is a project that is within the federal government called the Nationwide Health Information Network it was actually started under the Bush administration started in 2005 when the administration said we want to create a network for the exchange of healthcare records to provide many benefits provide every citizen in the United States with a personal healthcare record a digital healthcare record by 2015 right their first focus on this was to well they created what's called the National Coordinator for Health IT within HHS and through the OMB EGOV process went and interfaced with the VA and DOD and 28 federal agencies that have either possessed health data or have a need to see health data and said let's focus on solving the internal exchange problem some agencies such as the VA have had a terrific digital health data system for a long time other agencies like Social Security all of their processes were based around phone calls and faxes because when somebody applies for disability to the Social Security Administration it takes on order four to six months for them to be able to build that in some cases over a year for determinations to be made because it involves going to clinics going to hospitals and asking for data so there's a perception that there's a big problem here and they sat down and tried to say what are the existing standards in the space there were terrific standards for data formats HL7 and CDPQ and other digital systems that were just starting to be used but the way I characterize it is imagine if we spent the first 20 years of the web debating HTML versus PDF and we forgot to do the URL we said oh we'll hand around these HTML files on USB sticks or something so the realization was we need to talk about services and better or worse they pick soap as a platform for building network based services that were really about saying if I've got the DOD and I've got VA and I've got these soldiers who are going from their entering service they get injured down the battlefield they get treatment perhaps at a hospital in a different branch they come back and then the end of the VA system that has them crossing quite a few health IT systems and that's what helped them crystallize actually movement towards this virtual lifetime electronic record for the soldier and that finally created the use case in fact a year ago we had a conference around to run this project and had a living example of such a soldier whose spouse had become his living health care record because she was the one who carried the boxes of paper records from doctor to doctor meeting and sent those to the doctor who would prescribe treatment eggs and she'd say well maybe not maybe this other doctor said something else or when a medicine would be prescribed she'd have to say no he was actually prescribed that six months ago because the new doctor is not going to look through all of them right so tremendous opportunity here to save money to improve the quality of health care Peter Orzag at OMB believes that if we get to the system where this becomes natural that we could cut 30% off the cost of health care in the United States and combine it with comparative effect of all these kinds of things for us the project they started coming up with these standards and said to test it let's actually build some pilots so they took a big pile of money and funded ten separate implementations of the first round of these standards as part of it was we've got ten different organizations that have different IT systems of course they're all different and they went they went forward and built it demonstrated that okay there was some value to this and they went forward with scenes predictive the cost of actually maintaining independent implementations in different languages that was seen as a failure of that process so they took a step back and said what's a better way to do this and they realized that architecturally all they really needed to build was a common gateway that the VA could wire into VISTA to their VISTA systems that SSA could wire into theirs that other agencies could wire into the private systems and if we just built one body of code that talked on one end to the public standards and discovered other servers and exchanged these data it could be like what Apache was whereas a real simple gateway to back in services that could all be different but it's standardized and harmonized that on the way out over the time they also said we can't just look at the federal use case we have to look here at building this in conjunction with the private sector because A, our needs are not unique B, an organic agency like the VA spends 70% of its healthcare money in the private sector and the real problem in terms of cost savings and lack of adoption of electronic healthcare records is with the small clinics it's not with the Kaiser's they do a great job it's with the fore doctor operation in rural Kansas or downtown San Francisco in my case who never saw the financial incentive to actually move to a digital system for patient records for the billing as many of the billers said you've got to use a digital system but never saw the incentive to do that for patient records so the framing for this end-in project moved from one that was purely about standards where the audience was the federal sector to something more that was a combination of standards and software for which the federal partners were the primary customers but which now there is a greater incentive to socialize this with the public to with the private sector but how do you scale that and in 2009 they made the decision in springtime to open source the platform and I think they realized this is just before I joined I think they realized this wasn't the magic pixie dust that would automatically establish it as the universal platform but that at the very least putting it out in code was the first step I joined after I spent some time at the White House working on the open government project and Nisha Chopra pointed me at this he asked what do I know about health IT and I said well I went 10 years without talking to a doctor so the doctor that I have does not have an email address because he doesn't want to hear from those patients so I assume it's all like that he's like yeah that's all pretty sad so I dove in and spent the first six months every day learning a new acronym learning a new standard body learning a new constraint but discovered that here was a project that was just on that cusp they had just released the code and now it was time to build a community around that code base and the real question was how do we do this phase change a phase change from a project largely built by contractors because FHA the federal health architecture which is within the office of national coordinator has a little bit of staff mostly what they did was collect requirements from the federal partners and then boil that down into code to write right in the form of deliverable time and materials contracts to some great vendors by the way some really good vendors I drove a really agile process released every quarter but where they saw their primary customer as the federal partners I felt that rather than forcing to put something into a nonprofit or bootstrap it as an Apache incubator project that I felt that there would be a gradual way to do this phase change which would be about opening up their development processes so we took this aversion repository they were using made that publicly accessible took the issue tracker they were using the JIRA instance and there's still some process planning tools that are tied into how they build the federal sector that can't be made public because it's data about who's working on what for how many hours most of these systems most of the processes involved in actually writing the software and we've done this largely because the chief architect of the project actually now project lead Dave Riley wrote into the contract the ability to basically make this an indefinite deliverable indefinite timeline if you have the term IDFQ so he can basically create new requirements all the time and as long as he's signing checks for the contractors it works right but that's that kind of agility I think is essential to I mean it's how open source communities tend to work is by really focusing on a more agile priorities driven kind of thing so FHA is still playing this role of bringing the agencies into a production situation there with the new leadership at ONC with a bunch of the contracting work there has been funding on a continued basis but not the kind of reboot of the whole process that we've been needing for a little while so you know I'm hesitant to point to it as a model just to the public facing side we've been more successful by putting this code out there by recognizing that we can engage even the federal partners in a public process where the guy the VA is integrating this in can be asking questions that otherwise might be happening on a phone call or private email asking in public it doesn't involve confidential data it does in the etiquette private but if it's about how is this code supposed to work and we answer that question in the public it creates this public base of knowledge this Q&A back and forth that makes it easier to scale up and bring new parties in we now have I think the number is about 12 different vendors who are building products and services on top of Kinect we have 4 core contributors now outside contributors to the Kinect these are the Apache car lines committers right on the project who we see it appears in the process and we have even though the project originally started as kind of an integration of big systems we have examples of small companies that have gone and delivered this as a product as a solution to the needs of small regional healthcare information exchange needs there's a what's called an HIO a health information organization called Redwood MedNet based in rural northern California Mendocino County and Sonoma County who built basically a network to get small health clinics both within the HR as well as sharing data locally most health data exchanges local some of these things specialists going to a hospital local area and they built this in partnership with a company called Merth who had their own open source platform which is kind of a master patient index you plug that together with Kinect essentially you have LDAP plus PostFix to use the S&TP kind of comparison and they put a solution like this in a small little 1U rack server go out and sell 100 at a time now you have instant healthcare messaging and healthcare search that's for me a success they didn't need any federal funding to do that they didn't need any oversight by us they didn't need to sign an agreement with the federal sector and that to me is an example a successful example of scale coming to think about what we have learned that might apply to law.gov I would say obviously there's the federal case but let's also look at it was at Tocqueville called states the laboratories of democracy I think cities are even more so I think we see Toronto, we see DC we see San Francisco and New York if we can when it comes to data.gov even more so going faster and doing more creative stuff and even the data.gov federal level I think just as with healthcare a law.gov solution I would sooner have looked to getting responded by a consortium of cities or a consortium of states and saying we've got peers peers who all have a common need we were lucky with Connect with the healthcare project, the NHIN that those 28 different federal agencies all essentially have the same need if you're building an air traffic control software for one company though you're immediately going to realize oh I should partner with the other airports that need the same software whereas states naturally especially in the era of declining budgets are going to naturally want to see how can I cost share with other parties who have the same needs so that's the first thing and the second thing is finding a neutral ground to be able to have a conversation to bring them together and say let's have the providence of this project take place here I don't want to say governance it starts to introduce a whole lot of issues simply here's the neutral ground where we can build whether it's Google code and you know a simple wiki or it's under the auspices of a non-profit or an NGO that simply acts as kind of an oversight and one thing the Apache software foundation does is provide a legal framework for collaboration contributor license agreement and that way anybody downloading code from Apache knows that this isn't code that just showed up story one day this is actually had a chain of signatures back to the original developer so that I know that this license I see on it is actually trustable right? Likewise an NGO or I mean even an Apache incubator project could provide that same level of trust that as this thing has been built it has a legal history I would also say be aggressive in building on top of what's come before you know even if that means making a technology choice that some people might throw their hands up and say no you didn't pick my choice right? The Drupal community for example you know Drupal has emerged as a real solid content management system with lots of atom modules very well known now in the government space I would imagine 95% of what you would need to build for a law doc has been built already right and that perhaps might be the thinnest sliver of work to do but you know there's other options out there and finding a way to collaboratively build that between the parties might be the fastest way to get to a federal solution I have a question actually for Brian and actually with respect to what Jim mentioned as well is in the governance of the project things that as you say 28 peer states collaborating on something that is different than having 28 peer states and then for vendors who stand to gain financially from the outcome of the project contributing to it and so how do you manage how does one manage first of all mission creep right because it's community people and anyone can add whatever they want and so suddenly we've got suddenly it's sending email and it's a web server and it's a you know and then also like how do you prevent in some doing like how do you prevent also accidentally locking yourself in for a particular event sort of like running an android running an android right so the answer to the first is that one thing we've struggled with within FHA which is federal health architecture the office within ONC is conceptualizing themselves as a pivot point in the way that Red Hat is a pivot point let me describe it this way Red Hat has two communities that engages in it but let's say two buckets you know the financial services customers the health care customers basically the people who pay Red Hat for a support relationship who have things that they need to see and are willing to pay to have it all and that's part of their relationship with Red Hat you have the expertise and you can customer manage them and all that kind of stuff and they pay for that then there's another community of the open source community and you are this pivot point because requests are coming in you're managing them and trying to say well no that's really silly or yeah that makes sense for the Red Hat we never say well get to that next quarter and basically you product manage but when there's a new piece of code that makes sense you either advocate for it or go ahead and commit it on the other side of that question FHA is for you to figure out that they have to be in this pivot point because the question is how much of what happens over here should actually be happening over here quite a bit but then there's other things that shouldn't I mean these federal agencies expect FHA to deliver on a deadline to their promises because they've got pilots that they're spinning off they've got providers and other upstream things that they've got at times so when as the state wants to expand the work that they're doing with others it's like they have to plan ahead for them and that's a promise that only a vendor can really make you might be able to commit to delivering Red Hat branded code but you can't commit Linus to something like that we're not that far nobody's that far the relationship is a standard contract in relationships and getting them to think that there's also this community-based thing that there are more than that's where the phase change might have more of a hard clip it's actually an interesting twist to how these models end up rolling out so in some of the DOD projects that we've seen that have been open source for example there have been some cases like there's a project that the Army runs a simulation system and they sort of set out internally to make it an open source thing they wanted a community to develop around they wanted companies to use it but they've explicitly built into their contracts cleansed and licensing to say this is how you license it these are activities we'll pay you to do so we're going to pay you to run a repository we're going to pay you to run a community and do those things because if you forget to do that it doesn't happen part of the requirements for that program were to sort of do community and do that stuff and then you have other cases where it's actually been vendors taking initiative to create open source where they recognize that there are serial projects that will have similar needs and they take it upon themselves to sort of be the product management of the stuff so for example there's an open source project called Delta 3D you know cockpit simulation and stuff like that and what those guys have done is actually the corporate entity that said nobody ever asked us to do this but we're going to create a distro of this stuff put a bunch of stuff together we'll pay out of our own corporate IR internal R&D dollars to manage the repository and manage the ultimate product development priority list and then what we'll do is each time we get a contract where we're going to deliver the contract some of the overheads of those dollars to support this it's worth pointing that out just to say that there's different models by which something like a lot of God kind of environment can play out but if it is government deciding to do it one of the bigger question ultimately is how do I get these coordinating entities to agree on how to do this you ultimately want to make sure that whatever contracts they write explicitly like ask for the things that are not automatic sort of the master services agreement stuff these are the licensing terms you want to use you know we're going to pay you for it to do a repository we're going to pay to do community activities we're going to pay to run a master product list and this is how we're going to pay you to do governance around how it gets prioritized and all that kind of stuff that if you just sort of leave it then it doesn't happen but Brian said it was really important this is the only thing we struggle with in the military is they'd like to run things more open but how do you run things open because all of those projects they have a really strong leadership team either leader or leadership team in terms of people who are like here's what we're going to do that's what you know even if I'm laid off but you know government you end up with people who are like they run into their jobs they'll be here for three years oh healthcare okay I was going to miss one before but you end up with especially with government funding and what they're doing there is how do you kind of build that process and that system to where people are somewhat interchangeable you know such as the codes are interchangeable people have to be interchangeable as well so how do you kind of lock them into a mode of contractually bylaws or memorandum of understanding amongst the 20 agencies especially something like this where you move forward with something that could end up being multifaceted it might be services, it might be cloud code, production code you know how to kind of get the government to understand yeah you can run this and just what is that you know governance is the bad word but just kind of keep this thing rolling where the personalities don't really matter so much you mean like there's a great case of the polling in the system called SOSCO which kind of built like a basically an operating system for the military and at some point some girl came in instead of running on Microsoft he ain't got all the Linux stuff and he was there for two years and then like he left me, now that I came it's like okay I don't know what that was all about never went back to Linux and that's the word I do but you know in that case we're starting to just kind of come in and just change everything so it's like you kind of want to have you know change is good in terms of moving on so I'm really interested to see this is the toughest problem that in many of the toughest nugget for things like I mean connect or something like this going forward is keeping that okay so I think what we're going to do now is Vince going to say some a few words and then I'd like to open up the discussion a couple reminders we are videotaping this so if you're speaking if you can belly up to the table simply because we don't have an incredibly wide camera and if you speak for the first time please introduce yourself so that we know who you are I was a bit if you can maybe summarize what we've heard so far I'm Vince I'm Google's chief internet evangelist but I promise I'll try not to be preachy today let me start out by making the observation that you began this meeting with a kind of implicit assumption that open source was a solution to a problem you thought you had and so I want to step away from that with just a moment and only get there by some kind of syllabistic reasoning the first thing I think we want is to get the laws in a form which renders them easily accessible for everybody at very very low costs to access start with that there may be a cost to get them into that position we all I think appreciate that the implication that all the laws are somehow equally accessible says that they are going to have to be rendered in a form which makes them easily accessible which suggests they have to be in some standardized form and I want to use the word form in the plural for just a moment not to be trapped into a specific choice if there could be a few forms that could be widely understood and interpreted that might be adequate or okay or one might be okay although I've just come away from the meeting that talked about the bitrock problem which has to do with the fact that you store digital information away which is only interpretable if you have application software that knows how to interpret that in a particular format of the data and if that software becomes inaccessible or unavailable for some reason then the bits are going to be useful and they become kind of rushed so standards are going to be really important here even in the absence of open source the standards are going to be needed so if we have to start there there has to be some agreement about that it's also clear that you want to make it as easy as possible for the sources of those laws to render them in this form and here I'm not yet fully persuaded that open source is necessarily the solution to that I can easily accept the theory that there are proprietary packages capable of rendering things in a form which makes them interoperable or easily accessible one thing which is however an interesting question has to do with how to keep the cost of maintaining this growing body because it's clearly going to change over time so laws get new laws are enacted, more laws get distorted, they get revised and everything else so there is this turmoil and evolutionary effect which means that whatever this is not a one-time thing and haven't even said anything about peace law which is interpreting these things but that too probably has to be taken into account because even if you only have access to the law as written and you don't have access to the law as interpreted then the information that you get may not be as useful to you and that could be quite misleading if this has been somehow misunderstood in the absence of having the case law this is about primary legal materials on all three branches of government so if the law is as compiled in the code, it's the opinions as issued periodically, it's the federal register notifications as compiled in the code of federal regulations and that's what the national national level national level and then we have our states we have our county so you are including the case law primary legal materials in the united states so that means that we have and we're thinking not only very broadly but we just made the problem harder for the good of all mankind so I think that the the storage is not limited by the possible that's a t-shirt so now the capture of those laws and the ability to get them conveniently somebody's going to have to store them away and make that accessible because we all know we aren't going to go to each city in order to get copies of laws the whole idea was not to have to do that and there may be good reason for wanting to be able to reach the laws and case law on widespread basis so I may need to know what the laws are in 17 different states in particular now in order to render an opinion or formulate a plan or a policy this is where you get into a very interesting question at google of course we build major systems to store a lot of information away which can then be accessible to those who are authorized to get that data and in some cases it's probably being interested in indexing the whole of my web so there may be more than one way of which to get all that material how does it have to be in one central repository it can't be in our system it's going to have to be a distributed solution now there may be people that will aggregate it commercial vendors such as lexis nexuson west or the law library of congress but the source production of these legal materials by definition is going to have to be a distributed system that's the only reason it has to be by the government in one place and it would be very hard given the distributed nature of the government to do that there are entities such as the law library of congress that have a dream of being able to aggregate large amounts of information but even they understand they're not going to be comprehensive but you wouldn't need that right if all of that all of the primary law was publicly accessible google would make it searchable yes although it's also very important other than google to do that too we get to this need to have an architecture that allows lots of opportunity for let me use the word competing that I'm really trying to say multiple parties can find the capability multiple parties can fold the information as long as we can discover them I think it's also really interesting absent the widespread use of urns and the understanding of urns today urls serve as that point of authority right that to be able to say like if we're going to if it's you know law. say dc.gov it's the law.gov repository it's the district of Columbia right and then there's some regular format to the endpoints of that that becomes a reference that ends up in other legal opinions that want to cite prior legal opinions right that becomes the basis of the hypertext connecting these things and eventually we'll get the urns but for now url space actually connotes some semantic meaningful things it does although I am a very strong believer that urls are a very weak stopping point because of the fact that they are not necessarily stable the Cornell workshop by the way looked in depth of some of these issues and for example the urn Lex scheme coming out of Italy it shows great promise for legal material and has been adopted by a large number of entities one could imagine the administrative conference of the United States for example developing a urn Lex document ID mechanism which might then get adopted by a number of law-generating entities as a way of doing unique IDs inside of each of the documents perhaps the 7th that should be building a fire box module for resolving Lex. And so a lot of this as we've heard is about standards development but then one of the issues is there are going to be a lot of entities producing court opinions we have courts all over at different levels and the question is if we have a general principle that says if you are a court you should make your opinions available for bulk download and they should be authenticated. So the basic principle is that each of the bodies who are issuing portions of the primary legal matter legislative administrative and the court who are who are issuing case law and if each of them individually publishes a body of legal material then other sources can advocate for index and that will take care of itself in court. So the important part here is that we have a set of standards that allow for this degree of flexibility and distribution. And we're not worrying about the downstream part as much as the initial production of those legal materials so that the downstream providers can do things. So we clearly have a big challenge I think we have to identify what things need to be carried what standards need to be in place what practices need to be instituted and by referring to a predicated documents, things that use integrity and devalue we're implying a bunch of other mechanisms that we're going to have to specify. And to make matters even more complicated whatever it is that we choose will have to be acceptable and satisfactory to all parties is to use a very rough metaphor when we drive in states other than the ones that have issued our driver's licenses there is a presumption among the states that you've been adequately tested to drive safely across the United States we shouldn't take time to debate whether any of the tests are adequate or not but we make the presumption we make the presumption because anything else wasn't implementable so there are a set of presumptions that we will have to build in that either certificate issuances can be trusted or the technology that we've chosen to validate the documents is strong enough to modify material so this is a systems designing problem in my point of view even before we ever get to questions of local source so from there was an earlier comment about first we do the data and then we do prototype for reference implementations and then we try to do something that's operational and I'm sort of okay with that observation but I have a great desire to try things out somehow before I think it's safe to say everyone should do it this way most successful internet standards started with code and then became standards and it became standards because there were two implementations out there and people decided they wanted to talk to each other sometimes although there are different standards to figure out what should be held in common it was exactly what should be held in common it was avid issue it wasn't trying to get SNA to talk to Deccanere it was recognizing that SNA and Deccanere would never talk to each other except for the LOPS N squared implementations across the stop didn't work for the same reason that the licenses are accepted that's where tests make sure that you drive safely and every state you might show up so there is work for standardization but there's also work for validating what you chose to standardize actually made sense it scales up and everything else so to come now to the open source question it has advantages that have been outlined in some cases first you can know what the software is now this when you scrutinize that assertion you can discover very quickly that you aren't necessarily sure because they can show you the source code and they can give you the object code and what you don't know is whether the object code and the source code are you can say well I will compile my own and there are a lot of certifications are you capable of doing that I'm a little free I don't know what the word compile means so we need to be a little careful about wibbly assuming you know the equivalence I'm not trying to make this doomed from the start by any means I just wanted to be realistic about what assumptions we make so the open source also has to be maintained and there has to be the source of it which is why the Red Hat is terribly important the Apache example is terribly important because the fact that it's open doesn't mean that you should accept the next version from an arbitrary place because you might do everything that is supposed to do with something else that they didn't tell you about and you don't have time to go look so you need to have sources of open source that you promise to maintain that filter the suggested updates modifications and everything else which says it isn't terribly different than the kind of proprietary software that we use to use so the mechanics and the dynamics and even some of the costs are different here is that there is an possibility of someone else competing on the assumption that the open source is really important I think that's a key point man it's not necessarily that the entire stack is going to be an open source solution and everybody will be able to download it for free and use it I think the question is whether the core requirements for example if one could use a Linux operating system and could use Apache and could use Bind for the DNS and then a vendor could in fact build some authentication XML management on top the core issue though here is whether one could do this without issuing an RFP for a single vendor that might just to use an example specify Documentum and Oracle as the core data repositories and that all future uses of this software are hardwired to again just for an example Oracle and Documentum as their use and again the key here is that we're going to have a lot of small jurisdictions cities counties water districts that are going to have to be able to issue these materials and perhaps they're going to contract with a cloud like vendor a code management company that offers the cities to maintain their codes for them but once again we want to make sure that we have it hardwired that into for example the Amazon cloud I think it just said open source is maybe the wrong word lock in is maybe the right word it's two specific problems the first is making tools generally available to even the poorest jurisdiction and then second is being able to compete which we're actually two separate things ecosystem is a word I really like I think there's a third too which is that you said you want to avoid having to hire some big integrator or whatever to build this whole thing and as you described the problem there's sort of a vault slot kind of thing which says I want to have a big system eventually and I wanted to have grown up in an evolutionary way to meet all these needs I can do that one of the reasons why I might be getting from open source for this is I think that it helps develop that from a round up in short incremental cycles that eventually gets you there whereas if you do go say dear large company I need you to have a system that ultimately will satisfy 7000 municipal entities and please go figure out the requirements for that you just automatically occur in a situation you can't fulfill I met the CTO of the city of Santa Rosa in California really good guy really knows his stuff and if he is required to put their municipal code online he might RFP this out right and get software but it would be nice if there were some version of a code management system available that was available and perhaps it's not the full future version but one that's available at minimal cost if not free and certainly unencumbered by license protection it's if nothing else and open source implementation of the standards that we're talking about is an implementation of last resort so I think frankly if Oracle and Documentum and General Dynamics and Lockheed sell something great that's great like they may in fact do a great job and it's going to happen in places like California court system for example now here's the interesting thing the California court system has a system though it's more than simply dissemination of data it's also management of their data but they're spending three billion dollars on building a software system for management of the California court system just have a question in terms of who what are the companies doing that anybody actually built I mean Europeans has anybody built maybe the structure or the framework of something like this already it's out there in the open source land there is not a complete equivalent to WordPress for the legal publisher any other countries have done a much better job than we have in building open access to the law functions in Africa if you look at the LII movement from Cornell for example they've helped lead that as kind of an international revolution now that said those jurisdictions are potentially much simpler than the case of the United States where it's truly mind-boggling I mean we think DOD is complicated our judicial system for example so why would it be complicated if I could slip in there's a list of at least three making tools available oh I'm sorry David A. Wheeler IDA there's making tools available to everyone even those who can't afford say that three billion dollars system when you lock in and growing up in a revolutionary way now it's out of fourth which is similar to the flexibility to the growing in evolution where that's a whole flexibility the ability to change to new environments I know completely agree it's related to the first one frankly I think that's one of the big draws from the DOD world from the DOD world you know the adversary doesn't attack the same way they did last time necessarily we need to be able to adjust we get new systems we need them to be able to work with the old ones if we can't change the software on A and we can't change the software on B the only thing we can ensure is it's going to be really hard to work together whereas when you control the software on ideally both sides suddenly things like integration using a new situations is it's suddenly a rational economic thing you can do as opposed to ideally I could do it and I've had a billion dollars to spare I could do that also and so in theory you can do that other ways you know I'll build a new box and translate everything in squared ways or other stupid approaches if you like but it's a whole lot simpler to be able to say I need this software to work with that software I will change some of the software so they can actually work together and I think that is an argument if you want to go down this path for being able to do open source because it means that not only say I don't think you're going to talk about the central side but that different folks can adjust things for I'm a local municipality and I do things this way with this local legacy system and I now can integrate it into my system I guess I threw in a little more comment too I'm talking about open source as a potential approach I guess I was implying with that saying I'm assuming that the software that these government entities are paying to build can be developed in license that way in a modular design if there are components like to your point the things that are creating the documents in the first place wherever where there already is an answer to be sort of boundary implemented into the overall system and it already exists as a proprietary thing and on an economic basis it makes sense to do that it may be easier to take that off the shelf and do it but if you are the state of California and you're going to spend $3 million developing, implementing and operating the system I think it's in their best interest to do that in an open way ultimately because it will give them more flexibility in other terms First of all I think you should let everybody else have the space but I do want to over-emphasize that it's the standardization which is critical here the implementation question is in some respects secondary and important because of its enabling character but without the right standardization and architecture none of this stuff works it doesn't propagate it doesn't evolve or anything else so I would caution you to be very careful about being overly specific about the fact that this subject matter happens to be wrong and the only reason I say that is that if you want something with generalizes then you want to be a little careful about how special it is that's why HTML which wasn't specific to any particular expression of any particular information turned out to be so important because that you could generalize lots of people could use it there was a gigantic support platform because it wasn't specific to any one application so I would say while you're thinking through whatever gets standardized be careful not to over express the subject matter here Can I jump in here? How many came in on the Chewbacca Electronics? I think we did it on the head there because just we've got speaking from the kinds of things that we do speaking as a vendor and I'm not trying to say what we did I think this whole industry has to evolve but we've got some experiences to what works and what doesn't work and we've got case management software and I know where the complexities are developing case management software they're in the fact that every single jurisdiction and every single state wants that $500 an hour consultant to do it slightly different and it isn't that a single case management software would suit so if you think about it from that approach and it's generally the way I would think about it it's what can you do from a central body that would create enough standardization that you still have the motivations of the individual jurisdictions at heart but you get enough standardization that maybe market dynamics kick in maybe you've got incentives I mean when I think about open source as a commercial vendor I don't force open source there are times that doing open source is potentially a good idea for me and I've got projects that I do open source and I've got projects where I have intellectual property and I've got reasons for doing both I suspect the same would be true here and I would go back to the principles you talked about for Redhead if you go back to the attributes you're looking for which is big sophisticated systems for big sophisticated states or jurisdictions that can't report it I suspect the answer will present itself right the one thing I was going back to was a lot of good technology eventually becomes a commodity how do you kind of push forlaw.gov how do you create that driving factor like you made today I will pay for this but in three years maybe I'm not paying much more how do you kick off that commodization of that market because we all have cards but there are states who will basically say we know we'll do that how do you kick off that commodization and they think open source is a great factor for driving some of that but in the day if you could buy the same system at $5 versus $50 million it doesn't matter whether it's open source or whatever but you just want commodification I think the word open source is maybe the red herring and I'm certainly responsible for that but the issue is we have a lot of common elements in publishing many of the elements of the law and of course many differences in the jurisdictions and the courts are absolutely famous for that each one has different local rules but if you look at the process of producing annotated statutes for example the question is are there issues of standardization or reference implementations or encouragement of ecosystems as we've seen in the healthcare world that can help spread that on a national basis in a very complex system of lawmaking and the points of not getting overly specific on the law are vitally important the kinds of standards I think about are you should use HTTPS if you have a web server you should consider secure DNS if you're doing the DNS and those are not law standards so much as best practices that might apply to a wide variety of things but also would be very useful if you are a clerk about to publish a municipal code and looking for a specific idea I actually don't think it's a red herring but I would agree just focusing on just open sources by itself it's not right on the other hand focusing solely on we're going to make a standard is also not right you need a holistic way of approaching this I actually gave a presentation on standards years ago and someone years ago challenged me and said basically if it's a successful standard it has an open source implementation and I said now that boy I can't think of any counter examples I haven't think for years and I have still haven't found any very good counter examples you can probably find a few niche areas but in fact I think these exceptions prove the rule there aren't any open source implementations the problem with specifications and by the way I'm part of standards bodies I'm a big leader they don't do anything useful they sit there unless they are implemented and open source is a wonderful way of getting them out into those ants so it's worth considering as an approach I would say that I got a lot of the examples of open source that are okay for verifying functionality but they don't be very good for scalable production operation like lots of cases like that I think there's a whole lot of proprietary software which is exactly the same I would say that's a license I think you should not assume that the existence of an open source implementation necessarily means it is also suitable for production that's a new thing but we have people I want to make sure that a few people like Ed Walters and Mike Walsh have a chance to comment we're running out of time here Brian, I think the place you had to comment I'm famous for my equanimity about open sourcing I believe it's not a right hearing here to the degree that the vision for law.co is perhaps unlike its title it's not about a single site but instead about a type of site that every governing body would eventually have if that is the vision that is absolutely the vision that we're producing the law I think open source is actually key and I think it's the only enabling way that we allow these jurisdictions to get to that vision at their own pace at their own technology rapidly enough and I think it becomes a platform for either the formula or the de facto adoption of standards for the data types that it manages because you'll bootstrap it with some common things that are all about moving the judicial law data PDS, whatever or I'm sure there's an existing trough style kind of like data formats but I haven't heard that before eventually it'll be here's the data format for here's the data format for grounds for new laws I could even see it growing eventually become a CMS system for that the standards and governing bodies would use as a part of their process of defining law of going from raw graph to actually voting just like we use the Wikipedia state but that's a far off future I think again it's about many peers who all want to converge on a common platform and eventually talk about standards on top I think open source is totally key to that it's something that is hard to otherwise okay, Ed do you have any statements anything to say about this or I don't want to put you on the spot no I actually do have a few thoughts Ed Walters is the CEO of Fast Case which is one of the vendors in this space and has some experience with this well let me just say so they say that anecdote is the singular of data so maybe not illustrative about 10 years ago I left my law firm here in Washington to start a legal publishing company with the idea that we would take public domain law and make it available for free the absence of a law.gov in 1999 made that impossible to get the back file to collect the data from the thousands of sources that it comes from every day is hugely expensive and so as a result we made a service that I would say has kind of a premium model probably 95% of our users have access to Fast Case for free 5% of them end up subsidizing to 95% so it's a little bit of a nod to the fact that it's super expensive to collect this information both in the back file and on an ongoing basis that's not going to be the way it works in the future I think law.gov is going to be instrumental in making the law on an ongoing basis available to anybody who wants to publish it so I think that's hugely important I think Ben really nailed it when he said that we should be focusing more on the standards and less on sort of how we get there I think there's going to be a lot of different ways that people implement those standards I think a lot of them are going to be open source I think that's going to be a a really good way to get people into it I think there will be probably proprietary standards I'm sure, yeah, it will access what I'm not listening to proprietary implementations proprietary standards proprietary standards it's kind of an oxymoron there are those part of the problem I do think that there will be people who put together very nice implementations and very nice solutions for governments to kind of build into these standards but at the end of the day you have to make sure that the next fast case is able to take the judicial opinions from all of the courts and format in a way that's purposeable, easy to use simple, there's a lot of power in that and I think that's really important I think kind of beyond that one big question that we need to ask how do you create incentives for people to contribute to the system and here I'm thinking about those municipalities that are going to have a hard time paying the money doing the training getting the stuff online I think back in the 90s when my law firm had this bright shiny KM system that they had rolled out it was great and they did a bunch of trainings and a lot of money developing it and nobody ever contributed to it and so it was just a beautiful tortoise shell there is a kind of answer to that question specifically and I believe that people will adopt the use of something if it makes their job easier to do if it's incremental, if it's another thing they have to do there's less incentive but it actually helps them do what they need to do and it may also be the case that there are some forcing functions if it is required by some means or other that these laws be available if we have state level requirements or even national level requirements that they all have to be made available in some standard form the source is going to help because of the cost but it also helps them do the job which is now required of them so we may have to have some forcing function here plus the availability of software they can adopt to meet that requirement I think it's a key part of the equation because just having standards out there just having open source software because a lot of people say look I've got enough to clean out 5 every day and I'm leaving at 5 o'clock is it not already true that organizations that cannot publish jurisdictions that are not able to publish their legal materials online that probably belies some set of technology I must say failures but their technology situation is probably wanting and so by creating or inculcating me the ability to publish this stuff in a standard way they can actually feed themselves in this way I'm going to say the Court of Appeals of Georgia as a paywall they charge for the opinions how are we going to convince the Court of Appeals of Georgia to contribute to a system like this it's the right thing to do but they're going to look some on it this is just like trying to get IEEE and ACM to make their online versions of the journals openly available at least to the members of that incremental cost absolutely and the one difference between IEEE and ACM of course is that you don't have to obey the ACM so we're almost out of time I want to make sure before we close this session that we hear from one person Mike Washu is the Chief Information Officer of the Government Printing Office and I believe the only person in this room who actually produces primary legal materials and makes them available in bulk for people to download and does so in an authenticated fashion so maybe you can close this out Mike and I apologize for being late thank you we gossiped about you while you were done we signed you a video we signed you a ton of work great we've just been developing the RRP for you right now great I've been listening to the last half hour and I completely agree with the comments about standards I'm also very much an outcomes based type of thinker and when I think about law.gov and the work that I've been involved with the car on our conversation you know when I think about the outcome and we did this with our system at the GPO is when you focus on the outcome of freely available materials and being able to share and operate with information it really drives you toward what would be the standard what you know could eventually emerge into a standard and opens up I think doors for solutions that would eventually become open source or open source applications you know at GPO we've didn't have the luxury of using some of the open source tools early on these weren't really ready at the time so we're using document and we're using Oracle databases for some things and we're using fast research and things like that but the things that we did was particularly around a lot of the way that we manage our data has been to make it common and make it normalized and those you know whether you call those standards or not you know we forced ourselves to do things in a way that we can easily drive flexibility, expandability and extensibility into different types of things and it's already paid off and some of our packaging concepts for information we also drove ourselves to self describing information package which I think could emerge into an interesting standard because we didn't want to be locked into documental we didn't want to be locked into anything but our package could be if it truly is self describing it needs to improve it but if it truly is self describing it could work in a number of different environments in open source tools you know how Fresco for example should be able to work with our package and be able to completely recreate all of our access and publishing and I will second that by the way I spent a bit of time looking at the government printing office and spent quite a bit of time looking at this and I'm a skeptic when it comes to large government and I convince myself very firmly that I like FDCIS and I wasn't prepared to like it but I liked it because the system is architected in a way that one could in fact rip out any one of the pieces and all the hard work went into things like the metadata description and the packaging and the workflow and the provenance and that to me we've begun pointing to FDCIS as an example of what we mean by law.gov and in fact when we do a workshop with which John Podesta is hosting on June 15th there will be an entire session that Andrew McLaughlin is chairing and Mike will be at and the director of the office of the Federal Register and that will be an examination of not only how to do it right but when you do it right the innovation that happens on the outside we're bringing the Gulfpulse.us team in to do a demo of what they did with the Federal Register next and now. Sam will be able to talk about more stuff that they're doing. Absolutely. In fact there's been a feedback cycle it's fascinating because not only did the Gulfpulse team on their own innovate with the Federal Register the office of the Federal Register team is beginning to incorporate those innovations into their next cycle of publishing the Federal Register. So thank you everybody for coming we're just about out of time here I know Big has some other obligations I have some to do at 6 o'clock. This is the beginning I hope of a very long ongoing conversation that began here but I'm hoping that the government will begin having that and the government I mean is broadly rid will have that conversation much in the same way as Brian shown us have been helping in the healthcare role. One small comment because I'm looking for ways to facilitate this happening is of real interest. Speaking for Google here there's a guy named Don Horowitz who is a former state Supreme Court judge justice in the state of Washington and he has a project which he calls Access to Justice and he's talking about digital access his belief is that in the absence of having digital tools that help you get access to information about the justice system that there will be people disadvantaged who won't know what the law is or what their options are or anything else that sort of resonates with what you're trying to accomplish so an offline I should introduce you to Don Horowitz maybe there will be some... Well that's actually the flip side of this we just did the Chicago Cant Law Dock workshop last week and Professor Stout who's one of the leading experts in this area and is the author of A2J Access to Justice began to deal in great depth with many of those issues and it is a key part of the Law Dock of story. You know that over 50% of the people that apply to the legal services corporation are turned down because they don't have the resources more than 80% of the legal needs of the poor are on that a large number of the things in the courts when you go see a federal judge of filing requirements, they point out to you that the vast majority of their cases do not come from lawyers they come from pro-same representatives and so that's actually a huge issue and it's part of a workflow because if you're going to disseminate your data properly you've got to get it into the system properly and so that's a huge issue. Okay so thank you for being aware of the Louisville Washington DC office that this is exactly the kind of thing that we always hoped would happen here once we established this so please feel free to reschedule another meeting if you wish here I can help you do that. Thank you all. Thank you very much everybody. Let me stop the recording. Sorry I don't know.