 I'm Cliff Lynch, and let me welcome you to this session, which I'm going to be conducting, which is really about sort of taking another look at institutional repositories. As I say, I hope that some of this session can really be pretty conversational, but let me give you kind of a little background on what brought me to this, and then a quick outline of some of the topics I want to cover in the hour or so that we have. Institutional repositories are an idea that's been around since shortly after the turn of the century. Isn't that a wonderful phrase? Makes you really feel old. Anyway, it's something that CNI has been following pretty intensively over the years as part of our broader focus on changes in scholarly communication, on digital preservation, and the management of scientific and scholarly assets in the digital environment. And we tracked some of the early deployment working with a number of partners of institutional repositories. We've certainly been part of the ongoing conversation, although things like longitudinal data gathering about repository deployment has been passed on to other organizations following on some of our initial work. At the same time, there's a lot of weird discussion going on about repositories now. It's clear that there are some really fundamentally, if not incompatible, then at least divergent ideas about repositories that are emerging that are making the conversation very confusing. It's at the point now where you kind of need to know who you're talking to about institutional repositories to understand exactly what conversation you're having, because people really do mean quite different things. There are some interesting discussions, especially in the press about raising the sort of questions that people love to raise in journalistic environments. Have institutional repositories failed? This is a very interesting piece of rhetoric, because of course it invites you to engage the question of what does it mean to succeed or fail? And I'll come back to that one in some detail. Finally, there is, I think, a perception that, for better or worse, succeeded or failed, that institutional repositories have reached a certain level of maturity at this point, and that we need to recognize that they're here to stay, and we need to think about how they interconnect up with a lot of other components in the broad infrastructure of systems and other apparatus that support scholarly communication, libraries, the intellectual record, a whole lot of things. There's a piece of rhetoric that I find troublesome about how we take repositories to the next stage by thinking of an infrastructure of repositories. I prefer to, and would argue it's more fruitful to think about how repositories participate in the infrastructure and connect up to other infrastructure components. And so that's another place that I want to look. So here are some of the things that I'd like to kind of work through in the next hour. And this is all part of, as I said, C&I's kind of continuing look at institutional repositories in the sort of broader context will probably be producing a couple of articles or white papers or something along those lines later in the program year dealing with some of this, but I thought it would be very helpful at this point to have kind of a community conversation. So that's the immediate context here. So the specifics that we want to, that I hope we can get through are what does an institutional repository mean? Some discussion of this question about success, failure, expectations more broadly. A little bit about strategies for populating repositories and how the landscape has changed there. Some conversation about repositories in the broader infrastructure and in particular in that context, I'll say a little bit about a meeting I was at last month in Europe that was jointly sponsored by JISC in the UK and the SURF Foundation that tried to bring together a lot of people interested in institutional repositories to look at some of the sort of next steps in their evolution. And then finally, if we have time, a few comments about repositories in the broader context of service delivery, especially around de-research. So let me start at the beginning here with this question of definitions. One of the definitions is really about repositories being a very general purpose place to store digital materials that come out of a scholarly community, like a university. That might include data sets, it might include collections of images, it might include software code, interactive models of various kinds, manuscripts, papers, books, learning objects, very broad. And obviously, for most institutions, simply hanging out a shingle like that and saying, send us whatever is not really the optimal strategy for introducing it to the campus community. You do need to, at least the experience of many institutions has been focusing on some specific area of the community or some specific genre of materials or some combination of the two is a clearer strategy than simply saying, we'll take anything that's in need of a home. There is another view of repositories and I would say this is a view that's particularly strong in Europe, which says that repositories are basically where you put the published output of the faculty and particularly their journal articles. And there are a set of games that you play having to do with copyright transfer and how you're dealing with publisher relations about whether it's a preprint or whether it's a copy of the actual published article if the publisher will let you have that. But basically, the fundamental idea is to represent published faculty journal output. These are very interesting, these kinds of repositories because at least to my mind they have a radically different character than these sort of general repositories. They fit in a very different policy. They lead to discussions about things like open access, open access mandates, various kinds of agendas about opening the scientific literature. So they really take us off to a very different part of the universe than this kind of broader issue of digital materials. Now, it's quite striking to me that when you look at these e-print repositories that what you're dealing with here doesn't feel very digital to me. It's actually mostly electronic paper because that's exactly what's coming out in these journals still as they've migrated to digital delivery. The articles still look a whole lot like electronic representations of paper. And arguably at least from a stewardship perspective, this is the least endangered kind of material because we understand it very well. It has publishing venues. It has systems that are set up that are connected to the traditional scholarly publishing system to archive this material. Think things like locks or portico. It's really very familiar stuff. So the issues with e-print repositories are really not much around stewardship. They're really about around access, at least the way I read it. Quite a striking discrepancy. Now, I will tell you I've had some very interesting conversations recently with some of the folks who are very committed to e-print repositories who are now starting to talk about, well, the sort of maturity, maturing arc for these is that we'll begin to consider articles with supplementary digital materials. So we'll begin to introduce the truly born digital content, the kinds of things that do need lifecycle stewardship and that aren't being well accommodated in the publishing models that we're familiar with. We'll treat that as ancillary material, as supplements or enriched articles, but the foundation of our repositories, the sort of point of departure, will still be collections, articles. I find that very, very interesting and I personally don't think it's liable to be a very fruitful way to engage eScience going forward, but it's an interesting approach to try and open up a little bit that the e-print view of the future of repositories. I'd note that there's sort of one other niche area which has motivated a number of repository programs and that's electronic theses and dissertation programs. To the extent that your electronic theses and dissertations are PDF files, for example, it's not a big intellectual or technical leap to accommodate those in a system that's designed to deal with PDFs of articles, but it's quite reasonable to see people identifying an electronic thesis and dissertation program as one of the initial motivators for creating some kind of institutional repository service quite separate from questions about capturing faculty publication. It actually comes again from a very different kind of policy and intellectual frame. Now I'm a little nervous that it seems to have become an article of faith at least as I travel around and talk to institutions that virtually every institution should have an institutional repository and that this should be done without much consideration of what you want to capture or the amount of material you want to capture. I'd suggest for example if the main reason you have this is to do theses and dissertations you might want to count up how many theses and dissertations your institution actually produces per year and think about whether a locally hosted thing really is the most cost-effective way to handle that. I think that there's a lot of confusion around this kind of question. What I guess I worry about is that institutional repositories and the creation of such a service has become kind of a way of an institution signaling that it's serious about engaging changes in scholarly communication rather than a sort of a well thought out piece of a strategy for that. Occasionally I go to places and I will talk to colleagues who will tell me proudly we've just gotten approval to do an institutional repository and I'll ask what are you going to put in it what program is it going to support and I get very nervous when there's not a clear answer for that but I think we actually are at that stage sometimes. Now let me flip to this question of have institutional repositories succeeded or failed. I would suggest that if what you're trying to do is run an e-print repository that's blessedly a pretty easy question to answer. If your stated purpose is to capture a record of published faculty output your major issues are one getting faculty to put their material in the repository and their various approaches to that outreach mandates research assessment exercises in countries other than the United States which really provide very strong motivation for universities to get a very structured kind of documentation of their faculty output. You can actually make quite credible estimates of the number of articles published by faculty to give an institution in a given year and you can actually do quite credible statistical things looking at a repository of e-prints which permit you to make statements like we think we've got about 70 percent coverage this year last year we think we had about 65 percent coverage we're doing well and do and steadily improving this is a great success or to allow you less happily to do things like the National Library of Medicine did before mandates and said oh we have about two percent coverage this is not so wonderful that's great if that's what you're trying to accomplish I'd suggest also that you know if you have an ETD program and if what you've done is mandated that everybody put their thesis in there in order to graduate and you know how many people you're graduating you have a pretty good measure of success there and can can test and enforce compliance pretty effectively but let's take that broader picture for a minute that issue of repositories as a tool for providing stewardship for material throughout its life cycle in the digital world how would we test success there does the fact that a given repository doesn't have a million things in it yet mean that it's failed does something with only a relatively small number of contributions per year signal failure I don't think we really know the answer to this indeed the answers to this are somewhat intractable but I think we need to be honest and fairly aggressive about about leaving open the questions about success or failure at one level I think you could argue that success is recognized would be recognized by the fact that over a substantial period of time scholarship and evidence to support scholarship survives for reexamination reinterpretation and reuse in settings which it would not have done in the absence of an institutional repository that would be a success measure now let's let's just consider how horrible that is to apply in practice in the first place it involves long unspecified periods of time in the second place it involves a negative material that would not otherwise have survived always a easy thing to prove that's important material that turns out to be valuable for scholarly reuse and reinterpretation and reexamination across time something we're notoriously poor at predicting even in the physical world over long periods of time yet I think if we're really honest about the way we frame institutional repositories something like that would have to approximate the success measure and it's a very uncomfortable success measure but it's exactly you know at some level the same success measure that we apply to stewardship around cultural and scholarly materials going forward independent of format it's exactly the kind of uncomfortable success measure around stewardship that we've always struggled with I'll say one more thing too if you look again at the very comfortable kind of e-print interpretation of a repository not only can you measure success at some level in terms of the amount of faculty publication it represents you can actually you know what the time horizon is most faculty publish fairly often so taking let's say annual snapshots or something like that is a very reasonable thing to do I've actually even seen some arguments coming out of some of our colleagues in the UK that say that if you look kind of annualized across the academic year pattern of deposit as opposed to you know one big spike once a year when somebody comes round and says we're gonna audit our deposit mandate in two weeks is really the sign of a truly healthy repository that people are feeding it as part of the normal routine publication workflow as distinct from this sort of separate nuisance that you deal with once a year when you're reminded of it now in this kind of broader context if we want to try and track deposits and contributions as progress this takes us into a set of questions about what's the life cycle of different scholarly material how how when when is a faculty member who's produced it or captured it prepared to relinquish control when does it make sense for he or she to move a copy of it to the institutional repository for longer term stewardship I'm thinking that the answer to that certainly is not one of months it's at the very least one of years it's a cycle that perhaps is tied up with the completion of grants that run multiple years or even series of grants the shifting of a scholar's interest from one topic to another in a sense that they've you know they've finished that book or exhausted that data set or a line of inquiry and they're ready to move on to something a bit different perhaps it even literally is tied to the career life cycle of scholars that these kind of deposits are something that you think deeply about as you start thinking about you know concluding much of your career's work and you recognize that you've amassed a set of material that is going to be far beyond your capabilities to do justice to you've you've worked with other colleagues on some of it and you want to make sure that it moves on and survives in an orderly way I don't think we know these questions about life cycle in the digital world well at all we there there have been a few investigations in very specific areas I think that they lend themselves to a degree to be studied empirically although they have one of these bad characteristics that some of them are fairly long-term what one would really like to do is studies of considerable time duration in this area and those are always harder to do than you know your long kind of snapshots just in terms of mobilizing research or instant interest but I think that when we really talk of success and failure these may be the kinds of criteria we need to think about I've at least not been able to come up with ones that are substantially more quantitative outside of the sort of straightforward e-print model one can imagine you know in perhaps another decade as practices around e-science and e-research mature being able to get some insight into it through mechanisms like this it's it's it's becoming more and more clear that funding agencies are calling for data sharing and data management plans as part of grant proposals now basically what these begin to do is identify an inventory content assets that come out of grant supported work to the extent that these are accurate in predictive inventories one might be able to gain some insight by going back and looking at what happened to this material now simply saying is it all in the institutional repository obviously isn't going to give you the right answer because in some cases these kind of funded data sets will end up in disciplinary repositories in various settings rather than institutional ones but you can ask questions about are at least some of these represented in the institutional repositories and did the institutional repositories play a role in their survival I'd I'd I'd qualify this by saying that while this may be a route to some insight the institutional repositories are perhaps most desperately needed by those scholars that don't get a lot of grants that aren't in disciplines with good infrastructure for disciplinary management of content assets that really have no one to rely on except their institution in terms of the preservation of their scholarship and their evidence and those are the folks who won't have data management and data sharing plans on file with your office of contracts grants and sponsored research so those are a few reflections about success and failure in this environment which I think are worth at least thinking hard about I personally am very uneasy with this these sort of binary notions of success and failure in this setting and more comfortable with questions about asking questions about the the contribution that these services have made across time and then in evaluating resource commitments to these services by weighing those contributions versus the contributions that could be made by redirecting resources into other kinds of activities so those are a couple of reflections on kind of the current state of the discussion around institutional repositories I think what I'd like to do is pause for some reactions and and pushback or other views on that before I go on to talk about repositories in the broader infrastructure a bit I know that you know many of you have probably been following the same debates that I've kind of summarized and perhaps a little bit caricatured here I'm sure at least some of you have had people come to you and say did you know that repositories are a failure or how do you know your repositories is success I'd be very interested in takes from you all on some of those comments we'll get you next floor yes Debra Ludwig University of Kansas Cliff I was thinking about the quantitative quantitative measures and wondering if anybody is looking at qualitative measures related to faculty publications in institutional repositories and I ask it because we're very close at KU to a faculty Senate vote on an open access policy for the entire faculty and one of the criticisms that has come out is well the institutional repository represents something of lesser quality and frankly I as faculty X don't want to include my materials with the rest of that dross so let me throw that back to you well I I've heard the same kind of question and I think it's a very important question I personally I think that you know one of the very scary things we've done is interconnected institutional repositories with some of these questions about open access mandates because I think it's a tremendous source of confusion especially for the faculty I mean personally I would be happy with an open access mandate that said you know put it anywhere public put it in the physics archive put it in put it in the PubMed central archive you don't have to put it in your institutional archive or and actually it would be very easy later on to let these systems feed each other when we come to you know our discussion about infrastructure sorts of things it's really clear that we want a certain amount of replication in here that happens fairly automatically but I think that the the core difficulty there is that we're starting to give signals to our faculty that conflate putting things in the repository with publication and that leads them then to think about well how do we you know what what's the what's the level of quality and prestige of our repository as opposed to other publishing outlets I might select and you know the notion of taking a repository and comparing it to a journal seems to me to be apples and oranges particularly to the extent that you you start dealing with things that don't even look like journal articles you know we really don't have a good way to answer that question but I think that I think that we need to be very careful about the the suggestions that are sometimes made that repositories can represent a a form of publication that captures the evaluative characteristics of traditional journal publication for example the the other thing I'd say and I've run into a few places who in response to this you know start thinking about well maybe we should run peer review for the repository in some fashion that seems to me to be very problematic for two reasons one is that the kind of credibility you're going to get from an institutionally based peer review is limited the second is that it seems to me to run counter to the whole purpose of institutional repositories I mean that might my you know personal preference on this is subject things to some kind of limited sanity check for example you know if someone wants to give you 30 petabytes of you know CNN news that they've captured off of TV in high definition you may need to talk with them about things ranging from the cost of storage to copyright law but fundamentally you know I think your argument should be your institution is selective in its hiring of faculty in the first place you hire them and you give them frameworks for operating like notions of academic freedom as you know as a consequence of this and probably affiliation as a faculty member is probably a perfectly reasonable sort of first cut at admitting things to repositories probably for students you want some kind of arrangement where they're either sponsored by a faculty member or or doing something that's you know structurally reviewed like a thesis or dissertation but I think we I think we've really run the risk of talking ourselves into a real bind there people are trying to apply standards that I think aren't really relevant my institution may be a little different in that we run a university archives but carrying out your argument that all faculty at our institution must have some kind of quality we in fact reject any number of requests to deposit papers using a set of criteria so really the same kind of question about how do we determine or how do we determine what we select and what we don't and it's a harder question for us to ask I think at this point because many institutions are looking for things but if we reject out you know however many years from now do we want it to be filled with everything or anything except the obvious tapings or do we want to start with some set of criteria I think that's a really really interesting question and putting it in the context of things like faculty papers going into university archives which I know many universities have various kinds of qualifications on that they they accept selectively both in terms of what they'll take from a given faculty member and which faculty members they'll accept from in many cases that I think that's a very interesting kind of you know test point to put out here and is exactly the kind of thing we probably need to be considering if we're going to have a nuanced discussion about success failure impact etc. I'd suggest maybe that one of the there are a couple of things that are a little different about archives one is that usually once you take something you're kind of taking it forever you know the notion that you take a faculty members papers and then decide 30 or 40 years later that and nobody's really interested in them we'll just put them out in the dumpster in the back you try not to get into that situation with some of this digital material I think it has a natural life cycle which isn't necessarily forever and we need to be able to talk about managing through life cycles particularly when we look at some of the material that comes out of research data sets and observational data and things like that I think we're starting to grapple with the notion that these things don't simply either get tossed immediately or kept forever that there's a whole range of kind of intermediate possibilities and those are I think in play in the repository setting I think another piece of this is that some real consideration of the the basis on which an archive would make those choices would provide very helpful guidance for repositories and thinking about things but when you think about this just just recognize how different the level of interaction is here and I think this gets at some of the the extreme bifurcation between these e-print things and the sort of broader idea of a repository in the case of an e-print you have a lot of faculty who are probably publishing three four six ten papers a year so they're interacting on a very transactional basis with this and it's a kind of a transactional basis that makes and frequency that makes human review very expensive and in fact impractical in many cases other than to do some you know simple checking did I get the file in a readable way does it have some minimal metadata on it the discussion with an archive happens once in a faculty member's career typically this is a discussion that actually can involve some you know more elaborated human appraisal somebody might actually consider well what did that faculty member do what's in all those boxes who might be interested so you know the notion of having a repository deposit that's at that scale and frequency level is something that I think has been rather alien to our thinking so far and is much more akin to the sort of thinking that drives the acquisition of whole sub collections into an archive or special collections environment but is one that I'd argue at least perhaps we really need to move into our our thinking about institutional repositories that that's a that I think that's a wonderful you know kind of test case to hold up for perspective maybe take one or two more comments on this David you're talking about distinction between PDA type positive thing like data set but increasingly people are wrapping data sets and so on in work flows and services and so on as opposed and that needs a very different kind of repository infrastructure than you can just say okay here's the data set here is a bit are people looking forward in that direction or is this something we're going to end up with somebody walks up to the repository with this popular service that a lot of people depend on and say well you know I'm retiring or I'm moving to another institution here it is I think that that's intellectually certainly in scope I think pragmatically institutions are having so much trouble just dealing with if you'll sort of allow me a very loose term document ish or data set ish things as opposed to services that the kinds of things you're describing where we truly don't know how to handle them technically in a lot of ways are really you know people will admit intellectually yes we're going to have to deal with them but they have no idea what to do I would suspect that you know we're going to see a set of kind of boundary cases that are a little bit more manageable that people will be poking at those will include in my view things like simulations of various kinds games computer games of various sorts although I think that particularly when you get into multiplayer games we are rapidly again into we know this is a problem and we have no idea what to do about it pragmatically territory I think the other the other area you'll see some progress is software I think people are starting to recognize software as a kind of a legitimate form of expression that's that's worthy of preservation in a range of different tactics and levels you know at one level the the large source management system archive systems like you discussed yesterday at other levels just executables that can run within a some kind of specified emulation a environment but I think that you know when you get out on the the frontiers there we have got some you know enormous foundational questions about how to approach preservation of that that are unanswered as well as the the kind of more tactical things about how to do it so I think it's going to be a while before you see many services that effectively set up for that one of the things I worry a lot about is that it is a kind of an intermediate stage there where you've got large-scale systems that have been built up by a faculty member or a team that's anchored by a faculty member and that when that faculty member moves on or retires the the pathways to institutionalize those are very very shaky I mean I look at something like oh the work Greg Crane is doing on Perseus for example and you know I really wonder you know how many more how many years will that go on if Greg lost interest in it for some reason and didn't you know continually drive the funding of the platform and the you know rejuvenation of a team around it yet imagine depositing what what would it mean to deposit Perseus in an institutional repository you know yeah I'm particularly interested in the way you've been seeing people teasing this out between the sort of universal touring tape storage device that's short-term preservation it puts everything where preservation is just a backup and these large institutional issues because everyone has this this need to park their data somewhere and the I think the criteria for failure is exactly that that point of which it's not going to meet these particular notions and as various players in the institutional domain begin to recognize that it's not gelling around their particular idea they're going to say well I think actually that's a that's not only a good question it makes a nice connection to the last bit of stuff that I was I was going to talk about of course there's a large middle bit of stuff about repositories is infrastructure we've missed but let me let me try and link up to some of that I think that there is a view that repositories are coupled on some level with an institutional commitment to stewardship now that's not the same as a commitment to preserve forever I think I'm more and more here very explicit recognitions about life cycles that are attached to various genres of material and trying to develop ways to think about life cycles we did an executive roundtable yesterday for example talking about lecture capture and that's that's a perfect example of material that is starting to get produced at some scale by many of our institutions so you can ask some really direct questions about well now what's your asset management strategy on this and where are you going to carry out that asset management strategy well you could certainly put it in a repository but you need to think about if it's just for a given class for the duration of a given class for the consultation and review of students in that class it probably doesn't belong in a repository probably belongs in the learning management system or in some other special purpose thing if you're going to keep it for a few years but not necessarily archivally it might go in the repository with a notion that you know you're going to refresh this or junk it every few years and you know if it turns out that before you junk it the the lecture wins the Nobel Prize you'll move it over to the stack of stuff you you know keep longer term not necessarily because a lot of people are going to want to watch it but just because it seems like you ought to keep it so those are you know examples other examples are very much going to be driven I think by data retention requirements on grants where you make you'll make a commitment to 10 years say and after 10 years people will have to decide is it really worth tying up the storage space is someone really likely to come back to it so I think that that idea of life cycle is very key in one of the challenging things we face at our institutions as we try and think about digital assets is processes to define those life cycles now on the other side there's this question of how close are repositories to usage environments for data in other words you occasionally will run into this picture that says that you know a repository is really just sort of like network storage and you know every time you hit save on your word processor another version goes into your repository and all of your sort of working temporary versions of everything you do I don't think people mostly want that I think that's probably a bad idea you want some some slightly higher bar for placing a version in a repository than hitting the save button how you get at that is not entirely clear to me I think what you don't want for example is phenomena where while someone's working on something they have 90 versions of something in a repository and then when they final it they collapse those down to you know a final version to intermediates and throw the rest away I don't think you want that much volatility in your repository fundamentally but how we explain that and differentiate repositories clearly from network storage to potential users for the repository I think is a big challenge because it doesn't lend itself to really bright line things though the two last points I'd make here are there's a lot of discussion about what the repositories how repositories link up with e-research for example and the use of repositories to manage data sets I think that we need to be very mindful that the performance characteristics of a repository may be very different from the performance characteristics that you want for the storage of data sets that are an active part of an e-science stream so you know if you're looking at at doing computation even read-only computation on a large data set in an e-science setting you may very well I think end up wanting to stage a copy of that to higher performance storage storage with better connectivity to national high performance networks or grids to do that computation then to have the expectation that you can just beat on the thing in the repository some of the sort of lower end versions of that are very interesting too you hear people talk about well what should we do with video in a repository one extreme case of that is well you put it in as a file and yeah pull it out as a file and that's all the repository knows how to do with video drop it in as a file pull it out as a file and if you want to do anything clever with transcoding it and for various streaming settings or something you do that outside the repository on a video server that's provisioned for that in terms of software connectivity and cycles so I think we have those kind of performance constraints too that need to be factored and explained into the repository setting and some of those are are a wipe all as we go farther downstream to have some serious cost implications because the notion for example if you've got a big reference data set of moving it into a computational environment that's set up for that in order to really make use of it and that redundancy has a real cost attached to it other comments I just wanted to go back to this question of publishing the publishing idea and we've actually taken that approach because the idea of the repository is a place to put stuff didn't seem to be very appealing frankly and so to conceive of it as a publishing alternative of publishing platform with peer review capabilities that are not just in the institution they're done through the relationships that the faculty have has seemed to give us a different way to think about this and I think it was also born out of the approach that we started with that our relationship wasn't with the individual faculty member it was with the department the research unit and those especially the research units have relationships outside of the institution and just recently we started we have about 25 journals with another four or five to come and we started a new one where the editor is from Stanford and she wrote a little introduction explaining why she chose to use this tool because her relationships were with a research unit at one of the UC campuses so I think this you know sets a whole different framework and how to think about the value of the repository and how to evaluate its success so I just wonder what your thoughts are on that I'm glad you raised that actually I think that you know what you've done setting up a publishing platform of that kind that can be used by departments by inter-campus so are you research units and other things is wonderful and I think you know the example you just gave of having a faculty member from outside come in and you know place a journal on that platform is a measure of how well you've done in engineering that platform and getting it out there I think that setting up those kinds of platforms in the context of universities makes a ton of sense and I think we're gonna see a lot more of it I it's often affiliated with library operations I think rather than presses traditional presses because the libraries have more of a technology base more access to grants and things although I know of a number of cases I believe yours is one of them where at least for some of the journals you have the press involved in in various ways as well the only thing I would say about this wonderful activity is I really wish you hadn't called it an institutional repository I wish you'd called it a publishing platform or or a platform for for you know new scholarly communication because I see it I see what you're doing as most essentially a publishing and dissemination activity as distinct from a stewardship activity that's the that that's the contrast I draw and I admit that's not a you know sort of a bright line where you can say it's all one or it's all the other but I my personal view is what you're doing is more more about publishing and and I think the apparatus like peer review kind of emphasizes that character to I think it's wonderful stuff how are I guess I feel like we were we are vacillating between two like very distinct different approaches and over the last decade and we we're we're we're kind of we're gonna end up in the middle but we're at one end and the other I'm just thinking back to ten years ago when when we were working on for instance designing a preservation repository for University of California and we had huge fights between people who believed that a preservation repository should be only for preservation zero access and me who lost who believed that there had to be some access permitted to it mainly as a what I what I've always believed is the best way you have of knowing that something goes bad is to have some degree of access you know so so that that there is a necessity for some degree of access and then you know now we have things like the happy trust who you know basically are talking about providing full access to the repository and they're not thinking of any kind of thing that's out there that's protected in in some ways but it seems to me that we where we really should be is somewhere in between where there's there is there's enough access to assure that that nothing is going bad or that you catch it while you can but that the real you know hammering on it access needs to have to happen somewhere else so I think that's right I think that you know what's always awkward about that is you're now into very relative terms about what constitutes hammering and that of course changes as the price of cycles and the use practices on materials change I think that you know one other aspect that we didn't get into it all here but that what has been a very significant part of a lot of other discussions about the balance of access in preservation systems is in is one about intellectual property and so-called dark archives where really licensing and intellectual property concerns shape a lot of what you can do as well I I do think that you know as as we deal with material here that is much more explicitly computational in its use data sets that you know that relativism gets clearer because you know last year's big data set is today's thumb drive with all of the you know computational things that come along we are very just about at the end of our time which is awkward because I was going to tell you about repositories in infrastructure and I was particularly going to mention things like identifiers and name authority and some and inter repository replication as areas that seem to be emerging as calling for attention I guess the good news though is that we will have an opportunity to return to some of those things later there is going to be a report I believe in the next few weeks out of the meeting that I mentioned briefly that took place in Amsterdam that gets into some of this I'll make sure that a pointer to that gets out to see an eye announce when it's released and I imagine that we will have some opportunities hopefully in fact before the next CNI meeting in December to pursue some of those I would just like to really thank you all for joining me this morning and helping to talk through some of these ideas I would invite you and indeed urge you to think about some of these issues hard over the next few months because this whole question of stewardship strategies repositories and evaluation I think is one that's coming up to be a very fundamental agenda item for many of our institutions it's one we need good answers to and I hope that this conversation if not providing answers has at least helped to flesh out the agenda of questions that it might be worth you spending some time with your colleagues at your institution exploring so thank you very much