 Good morning everybody. My clock shows 803, so I'm going to suggest that we go ahead and get started. I'm Sally Jackson. I'm the chief information officer at the University of Illinois at Urbana-Champaign. And I'm going to talk today about a project that we've been working on for a couple of years now that goes by the unmelodic name Cyber Infrastructure Impact Analysis. What I'm going to ask everyone to try to do this morning is to really take seriously the idea that cyber infrastructure is a complex ecosystem involving the co-evolution of technology and scholarship. Multiple studies lately have focused in on the problem of managing this complex ecosystem. And so you see that in terms and phrases like the problem of building a coherent cyber infrastructure from the lab level to the campus level to the national level to the global level. And that's what this project is attempting to respond to. We need to stop thinking about IT as a set of IT and live. We need to stop thinking about cyber infrastructure as a set of IT services, library services, and research administration services. That isn't really what the cyber infrastructure is as it appears to any researcher, any student, any scholar. The cyber infrastructure has properties that are emergent and unpredictable. Multiple sources of control, converging to produce something that no one person can plan or manage. And so what we're trying to do with the cyber infrastructure impact analysis is trying to figure out ways to keep the ecosystem healthy. And I want to focus. There are many, many different ways to look at this. I mean, there are many perspectives that are going to be necessary for a problem as complex as this. But my focus today is on how research and other scholarly activity impacts the campus cyber infrastructure. How bringing a large project to campus, for example, affects its own particular context. And I also want to reflect a little bit on how campuses can respond to the co-evolution of technology and scholarly practice as though this is an opportunity for growth and development and not an infection that the system has to fight off. I'll give you an example of the kind of thing that defines one extreme of this kind of issue. We are currently building the Blue Waters supercomputer at the southwest corner of my campus. It is the tier one system for the nation, a pet-of-scale computing facility that will have enormous capacity to generate and process data on the kind of scale that leads people to talk about a data tsunami. It's too big for our current data center, so we have to have a new building. It's too fast for our network, as Ellen Botecki pointed out yesterday in his remarks. It's too power hungry for our local community. Neither our local utility provider nor our campus power plant have had enough capacity to handle Blue Waters. And all of these are ecosystem problems. They are problems that the University of Illinois has to solve in order to be able to grow this part of the cyber infrastructure. But that's just one extreme, and that's one case. For that one case, there are thousands of smaller projects that present problems that are not so dramatic, but they still accumulate to a very big management problem. And if you think about this happening over hundreds and hundreds of individual scholarly endeavors, you can imagine what kind of a task it is we're going to be confronting. So what are some of the sources of cyber infrastructure impacts on the campus? I don't pretend to have an exhaustive list here, but one impact that any change in scholarly activity can make on a campus is persistent commitments that go beyond the life of the project. And we're all familiar with this. Once you have set up a service, once you have established a data set, the obligation to manage that goes on long past whatever your initial project goals might have been. But there are also stack effects. And by stack, I mean the layers of support underlying any particular technology that's introduced to a campus. So a massive data set doesn't just require storage. It also requires transport. And so things lower in the stack are implicated by any investment that we make. We also have to worry about the byproducts of research. More and more people in the conduct of scientific and even humanistic research are producing things, artifacts like open source software that then have a life of their own byproducts that become valuable property in their own right and that need to be properly managed. And then finally, the thing that concerns a campus chief information officer most pointedly is the cost of operation, the space, the utilities, the staff, and other things that grow around any change to the computational environment, to the size of our data assets, to our network complexity, and so on. And I'll invite you. Can others think of other impacts of scholarly work on your campus environment? And have a feeling that you will think of things that I have not. But our goal here is to try with all of these potential impacts to try to manage a healthy ecosystem. And Alan Bloteky's remarks yesterday were very, very eloquent on this. He spoke at some length about the importance of looking at cyber infrastructure as an ecosystem. And I find that a very therapeutic metaphor. Change in any one element is change to the ecosystem and any deterioration in the ecosystem is some kind of threat to scholarly work. So the imperative to try to maintain a healthy ecosystem is clear. We need it in order to be able to do productive work. Keeping a coherent cyber infrastructure with all parts in balance gives us more possibilities. It allows for more and different kinds of science. And I have in mind here specifically the kinds of things that are happening as large and valuable data sets become open to multiple disciplines. And people find new uses for things than what the original producers had in mind. And finally, again, reflecting my pretty pragmatic concern for the cost of operating the place, a healthy ecosystem is one that uses its resources well. So we know that we wanna try to keep a healthy balance among all the various components of the ecosystem at the campus level, at the national level, at the global level. And this is notoriously difficult at every scale, occupying much, much reflection and study over the past five years. And there are some technical reasons why this is hard. For example, our computational power and our computational capacity has outrun our ability to move data for now. And that's a problem that requires a technical solution. But our problem isn't simply technical, it's also organizational. I like to show this diagram with the principal investigators on our campuses, the campuses themselves as institutions, and the agencies that fund most of our research engaged in what I think of as a kind of a pas de trois, a three-part dance, in which each of the dancers believes themselves to be performing well and acting rationally. But each is in some sense dependent on the others and constrained by the others. And they experience the others as a constraint on their own space for movement. So you have a very complicated set of dependencies in which if you look at it from any one point of view, it is the other entities that are creating the constraints and yourself trying to solve that problem so as to best manage your research and your participation in a broader scholarly community. And no one perspective can solve this problem. This is something that has to be tackled as an ecosystem problem. So we gotta take seriously this notion that cyber infrastructure is emergent from the co-evolution of scholarly practice and technology. These things are twined together. And at any point in time, what's emerging is unpredictable. From any one point of view, it appears uncontrollable. What this calls for in my mind is a new style of planning and cyber infrastructure impact analysis is one element of that. If we wanna protect the health of the ecosystem, we have to have some new ideas about how we manage and govern IT. This style of planning needs to be adapted to the way in which cyber infrastructure is growing. So it needs to be non-orchestrated. The growth of cyber infrastructure is itself non-orchestrated. And our planning also needs to be non-orchestrated. It needs to be embedded in the context of activity that actually introduces change to the ecosystem itself. For that reason, not only does it need to be non-orchestrated, but it needs to be centered on the scholarship, centered on the research. The thing that triggers a cyber infrastructure impact analysis should not be something like an annual planning process or something of that sort, but some kind of an occasion on which change is gonna occur. And typically that'll be research. And this style of planning also has to be much more context sensitive than anything we've ever done before in universities. It has to be sensitive to the local context. It has to be sensitive to the national context. It has to be sensitive to the disciplinary context. And it has to be sensitive to the interdisciplinary global context. And if it has an interest in cost containment, once again with my CIO hat on, that's so that more scholarship can be done. We want to make sure that in fact, we get all the value we can from all that we invest in research and education. So to sum up, the purpose of assessing CI cyber infrastructure impact is three-fold. The first and most important purpose of doing this is to try to cultivate an ecosystem mentality by trying to take any given project, any given program, any given investment and question what are going to be its impacts on all levels of context around and above and below it. We naturally draw attention to the rest of the ecosystem and to the part of the current target within that ecosystem. Second, cyber infrastructure impact analysis aims to amplify the benefit of every single thing we do. At the Coalition for Networked Information, I don't need to preach on this particular subject. As things shift toward greater and greater significance for data sets and data collections, our one of our goals is to try to extend the use of anything we produce and that's part of what cyber infrastructure impact analysis aims to do. It aims to do that not only for data but also for just about everything else that might be introduced in the service of an individual project. So for example, if we discover that a project is going to require some new form of network technology in order to be able to connect a researcher to a remote instrument or a remote data set, what we seek in cyber infrastructure impact analysis is not just the way to do that for that project but the way to do that in a way that'll elevate the overall quality of the network and help us to solve the network problem generally so that we can do that again on another occasion. And the last of the three purposes is to mitigate cost. I have in mind here primarily mitigating costs to campuses. This perhaps is not as high profile an issue at CNI as it is in other organizations but more and more people are coming to understand that our externally funded research actually costs the campuses a good deal of money. Even though we collect indirect, we collect overhead for facilities and administration, we under recover our true costs of research. And one of the ways that we can remedy that is by carefully analyzing what is the impact of any given endeavor on the overall cost of operation. Okay, now, so this is all background. What I've been trying to do is to persuade as many people as I can on my own campus and at other universities that it's really important for us to make this turn toward trying to understand in a very context sensitive way the growth of campus and national and global cyber infrastructure and the impact of individual projects within that. We don't wanna get in the way of research, absolutely not. On the contrary, we want to amplify the benefit that we get from every investment around research but we do need to do something because as Alan was describing yesterday, it does us no good to create a resource that presupposes a capability that we don't have and that's the sort of thing that this analysis is designed to try to address. So let's talk about the design of Cyber Infrastructure Impact Analysis. We've gone down one path at Illinois but we're aware that there are many other paths that we could go down. And I'm gonna try to organize what I talk about this morning around three critical design tasks. One task is finding an occasion for the introduction of the assessment. Another is the content of the assessment and then finally you gotta figure out who's gonna be the agent of the assessment. And for all of these, well we've chosen one path and we're looking at one initial sort of stab at this. We're also very, very interested in trying to see what could be done if we were starting down one of the other paths. The problem is anyone institution really can't do all of these experiments and so that's part of why I'm here today. I am looking for others who are interested in running parallel experiments in what I think of on an analogy with a multi-site clinical trial. This is an intervention we're gonna make with our various populations of researchers. But any one institution really can only choose one style of intervention in order to see what happens to practice this as a result. So I'm looking for others to buy the same sort of idea and to engage in these experiments in parallel with Illinois. Okay, so let's talk about the occasions on which we could do cyber infrastructure impact analysis. And there are many, many, many that we could choose. One idea that we thought of and that I like a great deal is to try to conduct cyber infrastructure impact analysis as whole research programs come onto campuses. And if we were gonna implement this at Illinois, the way that we would have done it would be to reach out one new faculty member at a time or one new PI at a time. Each time a new person appears in the environment, reach out to them, do an initial sort of essentially a business analysis around whatever kind of work they're gonna do, and then update that over the years to try to understand what the introduction of that research program is gonna mean for the campus. Okay, Illinois isn't doing very much hiring right now. Lots of universities aren't. So we didn't start down that path. We are doing things with all new faculty who come onto campus and we are reaching out to the highest productivity principal investigators. But the way that we have tried to situate cyber infrastructure analysis, given what our possibilities at hand have been, is to do it in one research proposal at a time. Rather than one research program at a time, which is something attached to a lead researcher, we've gone after the individual proposals. The advantage of that is we can capture every single one of them as they enter the pre-award routing for the proposal. Okay, so our implementation of cyber infrastructure impact analysis situates it in the proposal routing process for grants, external funding. That's very limited, but it gives us a way to start and try to figure out how this can help us to create a healthy ecosystem. What I've done here, I don't remember whether this is part of our routing proposal form at Illinois or one that I snagged from somebody else's research administration website, but pretty much all of them are the same. You have, for the researcher, routing a proposal to get it out the door to an agency. There's some kind of approval routing that you have to go through. With check boxes saying, yeah, I've had it go through the institutional review board. Yeah, somebody has signed off on my facilities commitments. Yeah, we are a drug-free workplace, and all of those other things that a researcher has to certify in order to submit the proposal. What there isn't at any university that I know of is a check box for cyber infrastructure impact, or for anything IT related. Nobody ticking off, yes, we have data center space for this equipment. Yes, we have a network backbone capable of carrying this size and volume of data. No one doing that. And to even suggest that is a little bit inflammatory to our chief research officers. They're worried that this will interfere with getting the grants out the door. Everything you add to the approval workflow does have that risk. So part of the design test here is to figure out how you're gonna mitigate that risk and convince everybody that it can be done in a way that won't keep a proposal from going out the door. Okay, we've worked on this very, very hard with our research administration office. I had an opportunity to shift a very senior person in my staff to a role that we simply called a cyber infrastructure analyst for a year. We were in an organizational transition, and he was the chief architect for the campus when I arrived as CIO. I didn't want a chief architect in my CIO office. I wanted the chief architect to be in the IT organization. So I offered the chief architect the opportunity to take this on and he did. That's Mike Grady, who many of you know. Mike spent a good part of that year proving that we could do cyber infrastructure impact analysis on the occasion of proposal routing without interfering with it. He went over and sat in the research administration office and took in proposals, paper proposals since we don't have any electronic handling of pre-award materials and worked through them, essentially conducting this analysis. I submitted a grant during that time. He reviewed my grant. That was invisible to me. Nothing in what he did was visible at all to the researcher. And that's because we didn't require that this be a check box. We only required that it be an FYI, okay? So, and he was able to manage the workload associated with this quite nicely. We now know that although it would be very boring for a single individual to sit there doing this all day long every day of their work life, in fact it is about a one FTE job for a place the size of Illinois. And that's, I mean, that's a pretty important thing to know. Okay, so one possible occasion is proposal routing. Another possible occasion is the introduction of a new research program onto campus. Another possible occasion for cyber infrastructure impact analysis is any major change in curriculum. And I'm sure that there are others but those are what have come up in prior discussions on this. So let's talk about the other design task, the elements of the analysis. Again, this is something that could be done in many, many different ways. What Mike and I decided is that if we were going to be evaluating cyber infrastructure impact one proposal at a time, we needed to have an extremely efficient way to do this that lent itself to triage. So somebody like Mike could run through the proposal summary and then shoot the proposal to multiple different readers to evaluate all kinds of impacts, data curation impacts, network impacts, whatever specialized needs the proposal might have. And so what we did is we developed a scoring rubric and then tried to apply that to proposals. Okay, it's not important to read the details of this and I know that you cannot but what I want you to see is that what Mike created is a sort of a table with a whole bunch of categories to remind whoever is doing the scoring of the things that they should be looking for. And I picked out there are many, many lines in this rubric and we've been through lots and lots of complicated versions of this before we've come down to the fairly simple version that we now use. But the section that I've pulled out here is the data section and the analyst is simply asked to estimate what the impact of the proposal will be for storage, for database, for all kinds of things having to do with curation, rates of change and so on. And there are other sections for networking, for space and special computational needs. There's a section for management of software artifacts and other things. And here again, the advantage that we would get from doing this as some kind of multi-site project rather than as a single experiment by one institution is in enriching the rubric that we use for evaluating proposals. In each of the main sort of technical domains that make up the cyber infrastructure ecosystem, knowledge is changing very, very rapidly. When we started doing this, there was no such thing as a data curation profile, for example. Now, we definitely need to try to align what we're doing here with the things that we're doing as we try to cope with people's new needs for data management plans for NSF proposals. Again, we're looking to try to find partners who will join in investigating this kind of planning methodology because it'll make the whole system healthier. Okay, so we've talked about the occasions for cyber infrastructure impact analysis and we've talked about what kind of elements you want in it, what kind of content. And the last sort of design problem that I wanna talk about is who is the agent, who's the analyst? Is it gonna be the researcher themselves or the proposer of whatever activity is going on? Or is it going to be a professional, somebody like a cyber infrastructure impact analyst? There are advantages for both kinds of approaches. We think we are on the path to a professional analyst and that's partly because I have an interest in trying to figure out ways to use surplus talent in the IT workforce as we move to higher productivity methods for all the operational things we do. What I'm trying to do is shift more of the IT professionals into work that is directly relevant to research and teaching. And that means creating jobs for them, creating meaningful work for them. And these kinds of roles are ideal for that. They help the IT professional to really understand the research enterprise in detail, give them an inside view of it that most IT professionals never have. But there are lots of other ways to do this and some of them have some pretty enticing advantages. Asking researchers to be the analyst and to essentially do a cyber infrastructure impact analysis statement of their own, just as we now are going to see them ask to do their own data management plans, helps them to think about their own work in relationship to the ecosystem. It helps them in discovering where all of the research stack is located. So if you think of their research support environment as having the things they control directly layered over all kinds of other things in the technology stack, having the researcher do this has some real educational benefits for the researcher, helps them to understand their context better. But I am pretty sure that researchers will not want this task, even though I would like them to do it. And so I'm not starting down that path, but I'm keeping it in mind. Okay, so if you think about this, the idea of a multi-site clinical trial, we have many dimensions along which the interventions could vary. The occasions for the analysis, the elements of the analysis, the agents who perform the analysis. So there's, so I'm not going to be honest but this doesn't work until I've seen trials of quite a few combinations. But I'm confident it will work because we're already seeing quite a lot of interesting result from it ourselves. So let's talk about what the benefits of this would be. The cyber infrastructure impact analysis, it isn't an easy idea to sell on a campus. The first reaction of everyone is, oh, PIs won't stand for that, or oh, the vice chancellor for research won't stand for that. And you do have to overcome that initial reaction. You have to take the conversation just one move further than that and then everything starts to be clear to people that you can do this without causing disruption. And once you get past that, that first little bit of resistance, it isn't a hard sell because it's beneficial to everyone. I managed to, oh, thought I disabled my navigation. But individual researchers will benefit from this in all kinds of ways. Whether they do the analysis themselves or whether a professional does it for them, there's the opportunity on any occasion of the kind that we've targeted to try to call resources to the attention of the researcher. Undiscovered resources and every campus has them. At Illinois, we have the advantage that our campus backbone connects directly to our own dark fiber, but it's not dark anymore that runs up to Chicago to connect to a thing called the Omnipop, which is shared by a consortium of schools, the CIC, and from there directly to internet too and other research and education networks. That means that we have some ability to control every link on the network path between our researchers and any other researcher in the CIC and beyond that to a number of other potential collaborators as well. Well, most of our researchers don't have a clue about that. They don't know what we have. They don't know that we're capable of giving them end-to-end control of the network if we need to. And so cyber infrastructure impact analysis allows us to see that, allows somebody who does know to see that and call it to the attention of the researcher. The research community as a whole benefits from ecosystem effects. If we can take large projects of any kind and figure out ways to amplify their benefits for the rest of the community, then that is something that'll produce more opportunity, more possibility, different scholarly opportunities. And then for the campus, cyber infrastructure impact analysis has a very obvious benefit, the ability to plan and to budget. I've been a CIO at two large research universities and the defining experiences of this job have been the things that have come on campus with huge infrastructure needs that we're never budgeted for. Arizona had two of these, a mission to Mars, which created a need for a network backbone twice as big as the rest of the campus altogether and in a remote location from the campus and a free supercomputer that only needed us to build a new building to put it in. And of course, many things like that have happened in Illinois too, but I won't bore you with my war stories. But being able to plan and budget for things like this is not just of interest to a person like the CIO or the chief financial officer. This is also very important to the research community. It allows you to be sure that you're properly resourced and it allows you to make sure that you don't find yourself in a position where you have to refuse a project once it's been granted. And that can happen too. It was touch and go on whether our vice president for research would accept the free supercomputer that we were offered at Arizona. And there's a really sad end to this story. The agency that was going to give us the supercomputer at the last moment, changed its mind and didn't deliver. Okay, so for campuses, the ability to plan and budget is more than just taking care of the money. It's taking care of the science and making sure that we actually can do our business with integrity. And then finally, there are benefits here for support staff and I've already hinted at this by being on the inside trying to figure out what are the IT needs of a research project. Support staff get a much better grasp of the needs of researchers. Whether those staff are librarians, whether they are IT professionals, whether they're people in the research administration office, these things are important opportunities to learn and to adopt the mindset of the researcher. And then scaling up, agencies should be interested in this, even if they're not, because it means that they get more product per dollar invested. If we can somehow better manage the bad cost impacts of our evolving subjects and cyber infrastructure, we can get more research for the money and if we can avoid the fragmentation that comes from getting the ecosystem out of balance, too much data, not enough bandwidth, for example. Again, we get more product per dollar invested. Okay, so this is highly beneficial all around. There are no issues and I've hinted at some of these already too. One is the potential for more bureaucracy depending on what specific path we go down toward implementation of cyber infrastructure impact analysis. We might find ourselves with very bureaucratic processes that require every PI to jump through 20 hoops or every department head who wants a new curriculum to talk to six new offices. But we don't have to go down that path. When we look at this as an ecosystem, that kind of tells us that we should not be trying to bureaucratize this, but should be trying to embed it in context. Those things have an affinity. And so I hope that what we are doing at Illinois doesn't look to anybody like more bureaucracy. I believe that when we implement this full blown, it'll be invisible to the researchers, except in those cases where it actually benefits them, where we're able to call them up and say, hey, we got an idea. There is the potential for researcher resistance. And I believe that as long as you don't let that be the end of the conversation, that part you can overcome. And it's very, very easy once you get past the initial reaction that this is going to slow down my research to get people to see that it's all benefit. And it is labor intensive. You have to be willing to allocate somebody. You have to be able to allocate somebody good, not just a body, but somebody with a lot of brains. So an unanswered question for Illinois and for everyone is whether the value of doing this will exceed the cost. I believe it will as we confront what we see happening. Yesterday, Alan's presentation repeated something that many of us have already seen. The idea that we are unprepared for what's about to happen to us with the change in scale and the change in complexity of our research and education environment. We need to start finding tools that are adapted to that increase in scale and complexity. So we're not gonna know for quite some time whether the value of this exceeds the cost, but this is so plausible and so likely to work that it would be foolish not to try it. And then I'm done with the formal part of this, but I've been referring at various points to a desire to try to create a multi-site clinical trial in which multiple partners try to implement versions of cyber infrastructure impact analysis. And we work together on this as a project trying to track what the consequences of this are, not only for research productivity and the success of the research enterprise, but also for the quality of IT environment. And on that note, I wanna stop and invite comments and suggestions from all of you, especially anybody who is interested in being part of a multi-site clinical trial. There's a microphone in there taping, so it might be good. For example, building that new building. So how do you broach the issue with someone when you come up with something like to do that requires a new building and maybe there's not space or capacity, which might actually be harmful to the researcher, but actually might be a necessary part of maintaining a sustainable ecosystem? That's a really great question. With something like needing to be able to provide data center space, if the researcher doesn't know that the university doesn't have that, the agency certainly will. So calling that to the attention of the researcher and saying, look, you need to be able to show that in fact we're gonna be able to successfully bring this project on to the campus and execute, that is beneficial to the researcher. It may not feel that way at first and we'd certainly want to try to make every effort to approach the researcher with a problem solving mindset, not with a veto kind of mindset, but it's simply a fact that sometimes researchers do have to be told, no, you can't do this. The campus can't allow this. With Blue Waters, being able to predict all these things and to understand all these impacts and knowing all this stuff very far in advance allowed us to find lots of sources of support for building the new building and also allowed us to organize very effective ways to use the prior building that NCSA has been occupying. So in part, that experience of watching how the campus has planned to bring Blue Waters on has been part of the inspiration for wanting to do this on a much more routine basis. You identify the problems and then when the thing is important enough, you solve them, but you're quite right, Rob, that sometimes this could end up being, sorry, but we're not gonna be able to allow you to do this. Hi, Ken Stafford, Kansas State University. How decentralized or centralized are you? Because most of the campuses I've been on are so decentralized this would be very difficult to do this. We are more than average decentralized. We are about one third of all the IT staff on the campus report up to me through the IT central service organizations and about 40% of the IT expenditure is somewhere in my organization. So we're not as decentralized as some, but we're a lot more decentralized than most. And you're quite right, Ken, that makes a huge difference to both the occasions for using cyber infrastructure impact analysis and its acceptability within the culture. But there is one point for research, there's one point where you can always be sure to capture anything that's gonna involve expenditure that doesn't go through a budget process. And that is in the proposal routing process. So that's one reason we chose this because we are so decentralized. There are many other possibilities for how to do this. On big decentralized campuses like Illinois, typically there are college level officials associate or assistant deans or even college CIOs who are closer to the research and know every researcher within the academic division. And it's possible that this kind of thing could be driven deeper into the research preparation process if that kind of organization exists so that the consideration of cyber infrastructure needs and cyber infrastructure impact could be done collaboratively with the researcher as they prepare proposals years ahead. We can't capture, at Illinois, we cannot capture activity that early in the process. We simply don't have the means to do that. And so, but in no matter how decentralized or centralized a place may be, I think there will be some method for capture including as a last resort just ignoring everything that's in the past and starting with the things that are coming new to your campus such as new hires, new curriculum, new buildings or whatever. So anyway, it's a good question and definitely something that'll make a difference to how any campus tries to find a positioning for this but that'll make it even more interesting. Two more, Perger-Rodgers. How do you plan to conduct a multi-site clinical trial? Can you say a little bit more about that, please? Well, I'm a social scientist and enough of a social scientist to actually really wanna do this. Try to gather data on behavior and try to measure the impact of the impact analysis itself. But that's not a very realistic ambition. What is realistic is to have several informal experiments of this kind going on with people comparing notes. So it'll be analogous. I mean, it's just a metaphor really to call it a multi-site clinical trial but it's analogous in many respects. Have an intervention, it's implemented in a situationally appropriate way here, here and here and then you look and see what happens to the patients and pool the data and try to come to conclusions about whether on balance this kind of intervention is positive or negative for the patient population. So anybody wanna be involved in trying to do this? I and my staff are really interested in working with others who will try it. I believe we're about out of time but if there are more general comments or questions, perhaps one. Okay, well thank you very much for being here today and I enjoyed the conversation that we've had and talked to a couple of people afterwards.