 Good afternoon, everyone. Thank you very much for joining us. It's a big, vast room. Feel free to make it look fuller by coming together or something like that. So my colleagues and I are here today to talk about a collaboration that's going on at Yale that involves the libraries, museums, and IT. And so what we'll do through our joint talk today is talk about different parts of that process from the structural changes, organizational changes, what our governance models look like, and then give you a sense of the projects that we're working on or have already completed. And then we hope to have some time at the end for questions from all of you, or perhaps your own sharing of stories of collaboration on your campus. So to start with, I'm Susan Gibbons. I'm the University Librarian, and I'm also the Deputy Provost for Collections and Scholarly Communications at Yale. And I'll explain that dual role in the moment, because that is very much part of the story. But let me get started. So we're going to focus first on structural changes that have happened at Yale. And to do that, I first need to introduce you to the collections of Yale. So we have four major collections at Yale. The Yale Center for British Art is the largest collection of British art outside of the UK, created by Paul Mellon. And he willed it to Yale before his death and has expanded since then. The Yale University Art Gallery is one of the largest university art galleries in the country, started in the 1700s. Yale Peabody Museum of Natural History is both a place where you go to find the dinosaurs and the beetles and the frogs, as well as anthropological collections of Egyptology, Babylonian tablets. And then Yale University Library, arguably the second largest university library in North America, has both manuscript collections and rare book collections, as well as our regular books. To gather, each of us is one of the largest in the country. Some of us have been around for centuries. And so I point that out to say, we have found a way to collaborate, even though our cultures would have been around for hundreds of years and would suggest we could stay in our silos. And yet we found ways to break out of those silos. And although scale is enormous for us, we're talking hundreds of millions of objects, essentially, even though scale is enormous, we are still stepping up to that challenge to try to figure out how do we work at scale. So let me give you a little bit of the background of where we are today. I want to point out, though, that while I'll be focusing on IT collaboration, what is key, I think, to some of the success is that the collaboration doesn't stop around technology. We are having conversations about shared storage, shared preservation and conservation labs. We are talking about how we can bring new mixed collections to campus where parts of it go in the art gallery, parts of it go in the library. We're talking about how we can help teach with our collections together. So we're going to focus on IT today, but I think it's important to recognize that across our staffs, across these institutions, collaboration is happening more and more and it's becoming our norm. So the story really begins back in 2008. And in 2008, the university created the Office of Digital Assets and Infrastructure called ODAI. Today, ODAI does not exist. However, it's important as a first trial for us how we could collaborate across our institutions. What really happened is soon after ODAI was created, the market crashed. And so the money that was supposed to go behind it disappeared and made collaboration a little more complicated. ODAI was created as another entity that was supposed to work within the technological space between the different cultural heritage institutions to really focus on the question of shared infrastructure. What ended up happening is with the market crash, the money went away. And there was always this question of who drives this collaboration? Is it the individual entities? Is it ODAI? Is it somehow shared? And that was really never quite worked through. But I don't want to suggest that ODAI did not provide us with a lot of wonderful things and lessons learned. In fact, we have an open access policy at Yale that's probably quite different than for many of you. Our open access policy is not focused on faculty publications. It is instead focused on cultural heritage. And what it says is anything in our collections that is in the public domain and has been digitized should be freely shared. So we've had a mandate now for seven, eight years that causes us to push out into the world all of our collections whenever possible. So ODAI went away and each of us sort of went back into our different worlds in our silos and we began doing our own work. But then things started to change. One was that we had a new president. And when President Peter Salovey started, what he did was really marked the fact that what makes Yale so unique and different is this breadth of our collections. That whatever the subject is you are teaching, you can teach with collections. And so he wants education at Yale in part to be qualified by the idea that almost every class can teach with collections. What that has mean for all of us is that we have expanded our teaching spaces in the libraries and the museums. And as an example today, the Beinecke Library teaches more than 600 class sessions in its own classrooms, each with special collections in the classroom. So we've done this in silos, but the pressure was on us to say, well, why should it be that a single professor has to move their class from building to building class session after class session to really understand how these collections come together? Shouldn't the burden be on us within the cultural heritage institutions to make it easier? So with that mandate of teaching with collections is of extreme importance to Yale, what is it we could be doing differently? So in 2006, I took on a second role, started as a university librarian, I'm still that, but now I'm deputy provost for collections. And what I'm asked to do is work with my colleagues who are the directors of the other museums to think about where is it that collaboration can happen? What is it that strategically we can do together and raise the boats of all of us simultaneously? So with that, we started having regular meetings. And just the fact that we all got together and had a regular meeting was like an aha moment for us. That hadn't happened before. The library museum directors didn't sit down for coffee on a regular basis. So through that, what started to happen is we realized we had some common pain points. And with those common pain points came opportunity to try to tackle some of these problems together. As I said, we'll focus on IT collaborations today and those pain points that got us started, but we have pain points when it comes to storage, when it comes to security forces. So we have a security force in each of our museums, how can they work together? We have common pain points about not enough space and talent for preservation and conservation. So we share many. The next thing that happened was in 2017, we had a new chief information officer, John Barton, who joined us. And within the portfolio of the CIO were many individual projects. He would focus on different projects. And what he was finding is there were just this vast number of projects that were asking for a finite bit of money. And what he wanted to do instead was create pillars to create within the IT portfolio different pillars that would focus on different areas, academic technology, administrative technology, student systems. And he saw the need to create a pillar for cultural heritage. And so this became a carrot for us. There's a bit of money that's available to us if we work collaboratively within the university's overall IT project budget. And that was another reason to bring us together to say, why would we pass up those dollars? They're not enormous dollars, but there's some dollars there, when in fact we already can see the need for us to work together. So we now meet regularly within the cultural heritage IT pillar to think about projects, plan out those projects, and you'll hear about some of them in a moment. So what has happened is we now have the regular convening of the cultural heritage directors. We have the regular convening of our cultural heritage IT and collection representatives. We want to be careful in the work that we're doing that we don't have it driven just by IT and we lose track of our users as expressed by those who work most closely with them, which are curators in our librarians. So they're part of our planning meetings as well. We now have a financial incentive from the university to collaborate with University IT and Lewis will talk to you about that as well. And what God is going. So when we got together and say, where are our pain points? The biggest urgency that we had at that moment was a solution for the museums around digital asset management and digital preservation. Their solution was at end of life and they hadn't planned what was next for them. Meanwhile, the library had brought in Preservica, which is a commercial preservation system. We stood it up, it was working well and that became the first point at which we could come together to explore what would happen if museums and libraries worked around the idea of digital asset management and digital preservation. So with that, I'll turn it over to my colleague, Lewis King. Thank you, Susan, I'm Lewis King. I'm an enterprise architect in the IT organization, Full Disclosure. I came to Yale to be part of the ODAI initiative and through those organizational changes, I ended up in IT and I worked very closely with that pillar because of that history and my interest in collections. Today, the arc that I'd like to take us on is from architecture through convergence to sustainability issues. So as Susan mentioned, we had this end of life situation. It didn't have to be end of life, we could have just spent a lot of money to get there and continue on that platform, but we felt there might be better opportunities. This is our overall architecture, don't look too closely at it because you're too far away, but overall, our collection management systems are on the far left. We aggregate metadata and digital representations of our collections or digital materials into a central aggregation. We make them available to our web presence through a content delivery network and a search interface with APIs on it, and then we integrate those collections into the web presence of the university, whether that be websites or different academic tools, our learning management system or research platforms. So as you can see, the core here, one of the core components is digital asset management. In 2009, when we built this, there were no really strong open source opportunities for us, so we went to a commercial platform and we went with that for nine years. In looking at that more closely, we thought because of the principles that we had followed, that we had the opportunity at this point to reevaluate whether that's what we wanted as the core system. These principles really are suggesting a modular system, one in which we felt when the end of life came of any particular module, we could easily exit that module and put another one in its place that had the capabilities that we wanted. One of the big pieces or few of the big pieces that we considered here is that we did not customize the applications that we licensed. So anything that we did customize were workflows outside of that system. So we owned what we changed and we didn't have to reconfigure that when we bought new systems or integrated new systems. The second piece that I think was really important is along the way we recognized that the scales of the participants were really different and what that translated to is substantially different workflows. Some organizations have 20 people in a workflow, others have two and so if we forced conformance on the workflow, we would have a bunch of unhappy players and we would not be enhancing productivity in any way. And so that architecture really gave us the opportunity to standardize only on the aggregate. Hit targets that allow us to bring things together and make sense together but before that you're doing your own thing. So we took these principles and we had this really great exercise. We took all of the technical thinking, design thinking and architectural thinking IT folks from each of the units and we got around the table and we wrote every system that we have in play and we threw it on the table and said, so what's going on? And then we started tearing strips of poster paper and using that to make connections and then everyone was talking about what, no that system doesn't do this, that system does something else and there were about 15 or so people doing this and we did it for a number of hours and a few sessions and in the end we came to the realization that we were able to transform our architecture in a way that considered the holistic view of the technology footprint that collections had at Yale, cross the board. We weren't going to touch them all but we knew what they all were. And so what we did is we moved the digital asset management system out of the core, the central component, collecting component and instead we put Preservica. We converged on the library's mature service for that and we would make that have the most highest quality resolution images that we have and we would also continue to use a digital asset management system but this was for professional use. So this would be for people who were using exhibits, building exhibits, shooting their own videos, shooting their own images and they would have a repository with all of our collections not at the highest resolution but at professional resolutions that would be a working environment for them and that would meet all of our requirements. That also turned out to be a cloud opportunity. So we refactored the architecture to include these things. We decided to do that a year ago and over the last year we've completed that capability nearly 70 or 80 terabytes of data migrated, millions and millions of records in all. But as we started thinking about the digital preservation service, we realized early on that we didn't have the staffing that we needed for that and so there was a discussion like well everyone's coming in, how are we gonna scale this up, what's fair? And the directors decided that they would all chip in a quarter of an FTE to add the staffing resource necessary and that was just like hey we have to get this thing going, let's do this but that made us start thinking about what are our structures? Like how do we make these decisions and who gets to pull those levers and how are we gonna resolve it? And so we put a working group on this and created this governance structure which is fairly simple. The steering committee is the group of directors in the conversations that they're having. They provide guidance and direction and oversight to a program committee populated by their representatives essentially and in the program committee people are thinking about where are the real opportunities for collaboration? Where could collaboration really advance the mission of the institution and how can we drive the experiences that we're trying to drive together and where are those points? But then come along the services, you say well how do you manage the services without really having to change your whole configuration? Are they gonna start reporting to the program group? And the answer that we came up with is no. Every service still reports to one department, okay? And they go up their hierarchy to that department. So the program group ends up being advisory to the service teams. So right now we talked about digital preservation as one of the service teams that's in the library and the other service team is the team that's doing development in ITS that are providing services for this group. So both of those teams have to negotiate with that program team as to what they're able to do but they all report and have one voice guiding them. The other piece that really helped the structure is thinking about how are we going to fund this? And we came up with these five funding categories yet another working group to do this. The idea is that once you put a certain amount of money into something that has to be repeated year over year you need some funding mechanism that will give you sustainability in that. So there's this piece, the annual subscription rate, that quarter of an FTE that everyone chipped in for that person they're gonna need to be there next year. And so that's right now that's the threshold of the annual subscription rate. But the directors themselves get to negotiate this, they get to negotiate who pays how much into that but what's important is we're saying that this is an annual thing that we're willing to pay into. We agree that we're gonna sustain at this level of activity. On top of that there's a commitment to in-kind support. The in-kind support is really the collaboration cost. It costs money, it costs staff resources to come in and have the conversation of how could we change and how could we create a better environment. And so that in-kind is a commitment to that. The commodity charge back answers to our different scales. If you use a lot of storage you pay for a lot of storage. Everyone pays it essentially the same rate whatever the gross rate is. The value add investment is an opportunity for two or three units as opposed to all of the units necessarily agreeing but two or three agreeing that this is an important thing we wanna look at. We're gonna bring resources for this but we wanna do it within the context of the collaboration because we're hoping that it'll advance the collaboration. Maybe that will become a shared service once we explore it or maybe it'll be chopped. And the final thing the transformative investment is really where the university or external funding could come into play such as grants to fund some of the transformation that we're driving for. And the last component that we put in organizationally is a model for how we could assess those shared services. So in talking to the shared services you really have to understand what's going on, what are the capabilities, what are the needs, what's the next two or three years look like. And this model lets us look at five levels of maturity against five specific attributes. And the attributes are detailed and we use it as an assessment model to review the service annually and to report out where the priority changes need to occur in that service year over year. And with that I'll ask Michael Appleby to join us for the next steps. Hi, I'm Mike Appleby, director of software engineering at Yale University Library. Until about two weeks ago I was head of IT at the Yale Center for British Heart. So I'll put on my museum hat here and talk a little bit more about the collaboration that the museum's embarked on, the infrastructure that they built and where we're going is next steps in partnership with the library to enable eventually a cross-section discovery platform that will serve library and museum collections and provide both a web interface and hopefully a data platform for future use. So recently, as Lewis mentioned, we've upgraded our digital asset management system, we've migrated to a cloud-hosted solution, we moved 1.1 million images and about 80 terabytes of images into Preservica and put access copies into our digital asset management system and we now have this true digital preservation capability with the library's Preservica instance. Now, as we look ahead, we're looking at two areas, one of which is metadata aggregation and the other is our online content delivery of images and when we talk about metadata aggregation, I just want to give you this high-level picture of what we're talking about, which is the three museums, the library, multiple metadata formats. You can see that even the art gallery in the Center for British Art, which you might think might use the same metadata standards, do not. So we're in a situation where each unit has its own expertise in metadata and has its own technical expertise in how it works with its collection management system that manages that metadata and all of these units have to collaborate together with Central IT, that cylinder on the right is represents the Central Information Technology Services Unit. It's again a very small unit within Central IT that manages the centralized database and index that this metadata is aggregated into. So there might only be one person for each of those museums who has a lot of other work to do, but they join this collaboration and they've been participating in weekly meetings as we've migrated our digital asset management system and they're highly engaged in this collaboration. When we started out, you might have seen OAI mentioned on one of the little stickers and Lewis's picture of our table as we were thrashing through the architecture. This is one example where the museums and library and IT were able to get together, talk about existing APIs, talk about how to build this metadata aggregation, agree on an API, which is great because it kind of lowered the level of effort that was required, we weren't inventing something new, we were able to embrace some open source software around the OAPMH protocol. But over time as we pushed this aggregation forward, we hit road bumps where we needed to look at more performant harvesting, we were able to have conversations with all the museums about things like Resource Sync and now we're looking at another protocol, activity streams that we might use to aggregate our data and at each step of the way, this partnership has been strong enough that the technical metadata resources from across the library and museums have been able to sort of reconvene or reevaluate the potential of making changes. In some cases say, well no, we're not going to adopt that but it's okay if the other museum wants to do that and we move forward. So it's been a great technical collaboration. And once we had metadata aggregated and we really wanted to meet the goal of this open access policy that was introduced to make our high resolution images available online. And each museum had its own approach to doing that. It wasn't very standardized and there was no guarantee that an image of a certain resolution would be available even if an object was available under the open access policy. And so the museums collaborated to develop a custom API that we would all use to access web derivatives of our images and as well as like higher quality downloads of TIFF images for print purposes. We agreed on the API, we agreed on the image sizes that would be made available so there was sort of set menu of JPEG and TIFFs. And it worked pretty well. We could make our paintings and sculpture and other anthropological cultural heritage artifacts available but it wasn't perfect. So there are outliers in our collection that just didn't work well with this format. You can see that the full screen image of that scroll does not in fact fill the entire screen. It only gives you a small ribbon. So we hit certain limitations and while we do make these images available online now in our online catalogs, it's clear that we can do better. And so we've been meeting again as a collaborative team to examine the use of IIIF, which some of you may have heard of, International Image Interoperability Framework. And we see this really as a way forward for our image delivery. One, it's an international standard and so not only will it benefit us to use it, it will allow researchers around the world to have ready access to our images through a well-understood API. And the API also supports activities that researchers might want to engage in, such as extracting regions of interest from images or annotating images using standards like the W3C web annotation data model. So we see IIIF as a great way forward to building out the underpinnings of this future cross-collection discovery website but also again, thinking of cross-collection discovery as a data platform that opens our collections both to other internal users like our DH lab or exhibition sites for the museums and so forth and the global audience who would be interested in using our collections. So I'll turn it over to Dale to talk about the future of cross-collection discovery. I'm Dale Hendrickson, I'm the Senior Director for Library IT at the Yale University Library. So you've gotten a good background of all the work that was required in order to make things start to come together, groups work together. Well, what do we do with that in the future? So one of the big efforts that was done as part of the ODI originally was examining all the metadata standards from the different organizations and coming up with a common cross-connect and identifying those key data elements that needed to be part of the index while recognizing that it couldn't be a comprehensive picture of all the metadata available at all the institutions. So one of the next steps for us is to review the work that was done originally with that ODI group, check it for validity, is it still current, is it still something that we can use and then start to move into whether we need to update those standards or can we move forward with what we already have. So this is a very big and important component. Another aspect of this is that we're also including physical holdings that do not have a digital mate to it as well. So the idea here is that the cross-collection discovery platform, this index, will be able to actually present items that are held that are not, that don't have digital objects that we can then connect to. But at the same time, if we do have digital objects, making sure that they're easily accessible through a one-click operation. Another big component of this is the user research and persona creation. Once we build this index, you need to be able to interact with it at a basic level. So the central ITS, our central IT organization has put forward some members of its user experience group and they're beginning to move forward on figuring out how to identify the people that we need to talk to to figure out what the user experience should be based on the personas of, say, a researcher in a particular area or a student in another. With that, it's the index development and implementation. As was brought up by Lewis, the original index, I believe, was a commercial product. Well, that product is now aged out. It's time to update it. Again, going back to the idea of the metadata standards, are they still valid and do we need to modify them? Do we go with an open-source tool, such as Blacklight or another product that might be available that fulfill that service? We see that as definitely a key component that we need to have in place first. Then from there, we will do the user interface development and implementation. Again, this is really focused on being able to use that index to have, look across the collections, but it's not intended to be too deep, where the real power we hope will be developed is in APIs, the ability to actually access that index data, pull it into your own system, and then be able to use it in the ways that we can't even foresee right now. So with that, there's going to be a lot of work with that, developing the microservices necessary to open it up. And again, this goes back to that idea of the open access policy, by having these APIs available, making them open to the world, that we'll be fulfilling that policy that we're working with under. So we're hoping that this leads to some great innovation and different types of uses of this that we may not even have thought of as of yet, as potential users throughout the world access these collections information and use it in innovative ways. At Yale, we're looking for innovative uses of the collection within teaching and learning. How could a cross, you know, a cross search of the cultural heritage institutions reveal interconnectedness between the holdings of Yale that may not have been obvious had you had to go to each individual site to be able to pull down that information. Research, what could come out of that? Where might insights be made? One of the examples that we use is if you're doing a search for Monarch butterflies on this cross-collection discovery platform, well, you'll get hits from the YCBA, you'll get hits from the Art Gallery, obviously from the library, and from the Peabody, and perhaps you'll be able to get a different picture of what you initially thought you were going in for to find but see something very different and be able to make connections between the collections that are available at Yale. And then the real exciting component is if these APIs are up and available, what kind of machine learning might be connected into this index and into these collections to be able to, again, create new insights and new opportunities for learning and to exploit the great collections that Yale currently holds. So that's all I have. So at this point, I'll turn it back to Susan. All right, so I was just gonna wrap up a bit to emphasize that what's going on is still for us new territory. We have a general sense of what we want to do. Often it's the holy grail of discovery across collections, very, very different collections. So there's a lot of R&D that's gonna have to happen for us to get there. But we're in an environment where we have this institutional mandate that says make sure our collections are being leveraged for teaching, learning, and research. And that gives us the impetus. But that institutional mandate doesn't answer the question of how. And so that's the excitement because we get to figure out the how. So what is changing is now we have colleagues who were in different silos, who are now working together and exploring things together. And we also have different organizations popping up. For example, we have a digital humanities lab now. So what can the DH lab do with this new content to work across the collections and help our students and faculty do that as well? So all of that's in our early stages for us. But we seem to have gotten over the hurdle of that. Why would we collaborate together? Often it is we come together in room, we say, ah, we're all doing okay. Thank you very much. Let's try again in a few years. And in some ways I think that early experiences with ODAI, that was our first attempt. It didn't quite work out, but we had lessons learned. So really for us, this is the second time around to try to reach this ideal of collaboration and we're much better positioned for it. The other thing that happens is as we bring a new leadership, we can set the stage. So the Yale University Art Gallery has a new director that started just a few months ago. Through her search process, it was very clear this is about collaboration. We're now searching for a new director at the Yale Center for British Art. Same thing is happening. So as we bring new individuals into this, collaboration is established from the beginning as being essential to what we do in the cultural heritage groups of Yale. And that's really helping to set the expectations going forward. So with that, what we'd like to do is turn it back over to you. You may have questions for us or you may be undergoing similar types of collaborations at your campus, which you can use this form to share with each other. So with that, I'll turn it over to all of you. Thank you.