 Hi everybody. I'm Ellen Ramsey from the University of Virginia, and I am fully aware that I am standing between you and the reception, so we will start on time. Thanks for coming. I'm the director for scholarly repository services at UVA. You can ask Chip what that means. We made it up. And we thought it would be interesting, and apparently CNI agreed, to talk a little bit about what it's like to live with an open source implementation. We've had one at UVA for a really long time, longer than I've been there, and so some of the things that I'll tell you today are outside of my personal experience, but I'm sure they're all absolutely true. I want to give you a little bit of background, a little history, about the platform and the things that we're working with, and sort of some things that happened before those of us who work on the project now have been working on it. I'm sure that you've heard the history of Fedora and that it was born in Cornell in 97, and 2001 is sort of when UVA started to get involved with this project in a really significant way. We had what we call an alpha test bed at UVA, and it was definitely a proof of concept, something like a half a million images in a Fedora repository. A few things happened over time. We got some grant money. There were some public rollouts, and by 2007, fast forward, we're starting to see some community movement, and so we've got Fedora Commons, and at UVA we did a little thing called Blacklight, and it was really interesting for me to go back and read the history on this. I don't think I was even a librarian yet at that point, no kidding, that there was a general dissatisfaction with what the vendors were bringing, and we wanted to build something that was better. Gosh, I think I've heard that before. So 2008, hey, institutional repositories, they're awesome, we should have one. So like all good research institutions, we wrote a report, and there are options at that time where you're not too dissimilar to what institutions right now are looking at in terms of we could adopt one, we could extend something that we already had, or we could go make something brand new. But there was something else going on at that time, really, for the first time, which was there was a lot of things going on in the open source community. And so the rest of the presentation I want to show you is what was going on at UVA, but what was also going on in the open source community at that time. You might notice though, my little icon there, given that is the holiday season, and given that this is a conversation about community source and what it means at the local level, I'll just say that every time a bell rings, a developer gets their wings, right? So I'll ring one bell, which is one of the original founders of the Fedora project at UVA, Leslie Johnston, left UVA to do much grander things elsewhere. 2009, hey, open access is good. Maybe we can use that to do some stuff. So there were things going on at the community level. We got our faculty senate to start discussing an open access policy, and best saddler left UVA to go, again, do really interesting things in the open source community at another institution. So there was also some community movement. We've got Duraspace coming together at this point. And so really a lot of things, a lot more resources that might be available to us. Fast forward to 2010. We finally get our open access resolution at UVA, and it provides the rationale for our continued development of our institutional repository. There's a link there. You can go read it. Judge for yourself whether you think it might be effective or not. We did get some resources, though, from the university based on that resolution to a higher media shelf, which is now DCE, to help us build what we call Libra. UVA has Libra, Virgo, Leo, you get the idea. The next one is cancer, so we're not building that. So we did do that work. Oh, wait. And so other institutions were building their own Hydra heads as well. And for the first time, we really see a community-sourced project to look at a particular implementation of a Hydra head, a Hydra implementation built on top of Fedora, which was called Hydrangea. It didn't really answer anybody's problems, but it was a nice start. 2011. Hold on. We had, fortunately, it was a good point in time that we had already begun taking deposits into the institutional repository head that we had built, but we did have a wholesale departure of our development team at that time. So things that were going on in the community, there was a lot of discussion about theses and dissertations being the low-hanging fruit, which I can't remember the article, I wish I had looked it up. Somebody said that's a horrible term, I agree. But we were taking that kind of scholarship, we were taking what I call small data, right? What do you get, 100 megabyte data sets? That's not even a data set, right? And we had availability for open scholarship in this tool. Moving into stuff I can actually remember now. At this point, between 2012 and 2014, I want you to notice no developers left. And we got some stuff done, right? So, and you start to see our repository down here, this is what this is, is really growing, okay? So we're up around 3,000 items, something like that. Also that, I would say, are theses and dissertations for sure. So we've got that going and we've got great linkages there. Not too much interest in the other things that the repository can take. And, oh, all of a sudden, man, that code base is getting old, right? So things that had been going on in the community is that Fedora had been, you know, coming out with new public releases. We had one, we had two, we had three, we're now getting to four. Our Fedora instance at this point was at two. So that's where we were, that we were interacting with the community but had some challenges to overcome. And, oh, there's another community-sourced institutional repository, collaborative project, this one was called Hydromeda. Yes, there's some lapses in the back of the room. Wait, you want to know who was in that team? It'll be a good list. So it was five institutions, Notre Dame, Northwestern, Indiana, University of Cincinnati, and University of Virginia. So there's some really familiar names to those of you who are still members of the now-Sanvera community, right? So lots of things did happen. They did not happen through Hydromeda. 2015. Yeah, that code base seriously old. But, oh, I don't know how to do a half a bell. You only had one developer leave, and we got two more. So I think that's a half. So we stopped our participation in Hydromeda at this point. But we were really clear that, man, we got to get off this whole thing, right? The usability for the students with the PCs and dissertations was dropping. And they had to use it anyway. So we had a captive audience, but that wasn't very nice of us. Nobody was using data or open, hardly at all, because it wasn't a pleasant experience. And all of those other reasons that institutional repositories also maybe hadn't been the answer to some of the things that we thought they would be. But in the community, we're getting some convergence on a couple of more successful institutions, institutional repository solutions. So ScholarSphere, Notre Dame's NDQ8, which NDQ8 was actually what came out of Hydromeda, mostly. And we're seeing more convergence on Fedorf. And we finally had the opportunity to go to Fedorf for. And what we took was a modular approach to what we needed to do with our repository. And a couple of things that we did a little bit differently. So we moved our ETD implementation onto Sophia, which is what ScholarSphere became. And that was where the community was converging. We did something really radical. And our library IT director is still mad at me about this, which is we said we're not really sure that small data is that useful. And so we're not going to build anything for that right now. But there's this awesome thing called Dataverse that we're going to use for a while, thus the modular. And that has been a reasonable solution for us for the last couple of years that we've had it. And the other thing that we did is we took away the discovery layer from the institutional repository itself. If you want to know what's in our institutional repository, that's awesome. You want to know what's in the library. And so it's in the library catalog. It's in Google Scholar. It has DOIs. It is in data site. So go find it there. We don't want to silo this content because if the objective is to make open scholarship in context with all the other kinds of scholarships that libraries provide, that's great. Let's put it in the same place. And we had it there. But by forcing people to, essentially, by forcing people to look for it in the library catalog, that really makes it clear that this is really in the library. And it's valid stuff, valuable stuff. So we're watching the community. As I said, ScholarSphere became Sofia. And you curate became curation concerns. And while we were launching Dataverse, we're also watching what Michigan is doing with deep blue data. We're still watching. And they've got wonderful stuff. And that may be the direction that we go. But there's some other stuff that's happening in the broader community. Elsevier starts buying stuff, like SSRI. And we're seeing some more Mellon investment that we think is really relevant to us. So Hydra in a Box. And I'm sure that you have heard of the presentations that have been around the conference about that. And there's some really interesting things that Mellon is funding in terms of what might we do with open monographs. And initially, that research was about how much does it cost? But then that has, in the last year or two, moved into, OK, now that we know what it might cost, how do we do that in a community way? So here we are in 2017. We finished building the modules to our institutional repository this year. We launched the faculty piece, the open scholarship piece, of Libra in August of this year, Libra Open. It's got some stuff that we didn't have before. We were actually pushing people's deposits to Orchid. And that's pretty neat. We didn't give DOIs before for open content, but we did for ETDs. So we added that. And the other thing I didn't put on the slide, but we just had it on Friday, was we turned off the original Libra instance. We had a wake. That's fun. So other things that are happening in the community, we've now got high racks instead of two separate instances of Sophia and Curate. People are looking at Haiku. And we're seeing some interesting things going on with open publishing. More on the journal, sorry, on the open monograph side than on the journal side in the community. And so that's what we're watching. So what's next? We are really proud to announce that, can I announce this, Dave? He's looking at me. That we are piloting Ubiquity Press as an open journal service. And again, it's very similar to our decision about Dataverse in that we're really not sure how well this is going to take off. It might be fantastic. Dave thinks it's going to be. And if it is, that's great. And we may invest resources later into building something more custom. Or we may say, hey, this is awesome. And it is an overlay of an open system built on OJS. But it is also a vendor. So we have made these decisions about what we can do to be modular, to be agile, to be responsive, without making enormous investments in. So we will be migrating to Hierarchs this year, given no ringing of the bells. And other things that have been happening in the community, I think you've heard about those a lot during this conference as well. Hydra is not a thing, but same there it is. And I don't know if I've heard this today, but that Haiku and Hierarchs, the code bases are actually going to move. And so there should be some more alignment. And that will be really interesting for institutions that end up going with sort of the more out-of-the-box solution of Haiku and want to do more customization with Hierarchs. I don't know that I would go the other way, but you never know. And so we're watching. Now we have the great advantage of watching our partners already migrate to Hierarchs. And they said, oh, sweat. And our repository has continued to grow. So some of the culture shifts that have happened during this time, as we talked about moving from an all-in-one repository, Libre One was just everything in one bucket to some more replaceable components moving to that single search so that we can have lots of sort of invisible sources for the content coming to what feels like the same place to users. Internally, we moved from a more traditional, a little bit larger investment of time of an epic and story model of development to a more agile model. And I had to learn how to do that. And I think that it has been very successful for us. It's also something that the community was doing. And so we've been able to align better with the open-source community. I think we have to recognize that as a partner, we're really, as it says, just altruistic enough to participate with the community but make sure that we've got our eye on what's useful to us. And so there's a lot of pressure in the San Vera community, not a pressure, but there's encouragement to contribute to the community code base. And we've done that. We are doing that. And we're doing it in most cases where we see a real outcome for what our repository needs as well, what our institution needs. And that's a little bit more pragmatic, I think, than some institutions have been in the past and that some institutions are now. But, you know, just saying. There's some things that have happened in the community that have been really helpful and moved a little bit away from purely, from just a technical community to really how to manage these projects, how to transition to, okay, we built it and we still have to keep the thing going so that we don't have happened again, you know, an orphaned code base and aging software. And it really has been sort of what I would term agile production. We are still doing sprints for our Libra implementation, but we only do them once and once now. And we sort of gather up a little bit of stuff and roll it out. And one of the benefits of that has been that we had, our ETD deposit period is always like right after Thanksgiving. So we draw straws, who has to cover over Thanksgiving. We had no questions this year over Thanksgiving. And so, you know, we really have been able to respond to our community while still keeping current with standards. And I think I made the last point already. There's a recognition of things that you can get from the community, but it also, that it costs. And it's not always the solution in every case. Either right out of the box, or maybe something that we go to longer term when we feel that we need a more custom implementation. For your reading pleasure. And the UVA team is so sick of this graphic, but it gives you an idea of where we started and where we're trying to head. So the gray areas are what we are thinking about and the white areas are what we have already rolled out. I would be delighted to take your questions at the reception. Thank you. Thanks everybody.