David Newbury: Pointers, Proxies, and Modeling Complexity – Episode 59

David Newbury Modeling data and knowledge always involves trade-offs. At the Getty Trust, David Newbury manages what he calls a “complexity budget,” focusing on the concepts his users are most concerned with. For example, they carefully model time for art historians while downplaying geospatial data. Getty shares the resulting structured knowledge in two ways: as simple reference documents and as a rich knowledge graph that helps researchers answer questions that no one has asked before. We talked about: his role as Senior Director of Public Technologies at the Getty Trust how Getty settled on semantic technology to manage and share knowledge about its extensive collections the differences between how an organization like Getty thinks about its collections and how its users access that information the unique benefits that an independent cultural heritage organization has over commercial entities when it comes to knowledge architectures his academic work on "pointers and proxies" — knowledge organization that helps users find things vs. metadata and other information that lets researchers digitally model items in their collection to conduct research the philosophy and practical considerations that underlie their modeling practices their allocation of a good portion of their "complexity budget" to modeling time, since that is such important information for art historians and other researchers the crucial role of linking in their metadata and modeling strategy the origin of the concept of provenance in the field of art history the two different ways they talk about provenance at Getty: their Provenance Index, a huge collection of art-transaction receipts, and their historical records of the people, places, and things collected their goal to serve as a source of ground truth about objects in their collections, providing the same information both in reference documents and in complex graphs that let people ask research questions that Getty staff may not have thought about the broad adoption of their Getty Vocabularies, a collection of cataloging taxonomies that are a standard across the GLAM space, and their upcoming updates David's bio David Newbury is Senior Director, Public Technologies at Getty, where he leads the team responsible for using technology to share Getty’s knowledge, collections, and campuses with the world. He oversees digital imaging, collections platforms & data, and digital visitor experience, and has led Getty’s transition to a linked-data first organization, working across teams to develop data infrastructure and access mechanisms for Getty’s knowledge and collections. David is the former president of MCN, on the executive committee for the IIIF consortium, and one of the original creators of the Linked.Art standard. Prior to his time at Getty, David's work included Art Tracks, a provenance data standards project with the Carnegie Museum of Art; and the American Art Collaborative, working with 14 museums to connect collections through linked data. He formerly worked with Carnegie Mellon, American Eagle Outfitters, the University of British Columbia, and PBS. Connect with David online LinkedIn email: dnewbury at getty dot edu Video Here’s the video version of our conversation: Podcast intro transcript This is the Knowledge Graph Insights podcast, episode number 59. Modeling data and knowledge always involves trade-offs. At the Getty Trust, David Newbury manages what he calls a “complexity budget,” focusing on the concepts his users are most concerned with — for example, carefully modeling time for art historians while downplaying geospatial data. Getty then shares that knowledge in two ways — as simple reference documents and as a rich knowledge graph that helps researchers answer questions that no one has asked before. Interview transcript Larry: Hi, everyone. Welcome to episode number 59 of the Knowledge Graph Insights Podcast. I am really delighted today to welcome to the show, David Newbury. David's the Senior Director of Public Technologies at the J. Paul Getty Trust. Welcome, David. Tell the folks a little bit more about what you're up to these days. David: Hi, as you said, I'm David Newbury. I work for Getty, which is a museum/library/research institute out in Los Angeles. It's always important for me to start these things by saying that I am not part of Getty Images. It turns out that generational wealth tends to create lots of institutions and they name them all after themselves, but we are a museum and an archive that collects data and shares out with Los Angeles and the world. We don't sell stock photos. It's always confusing because we do collect photos, and so there are millions and millions of photos here, but not those ones. Larry: Got it. Yeah, no, I think it's always important to get that out there. David: Yeah. And I have to start every presentation that way because no matter how many times I say it, someone asks me about royalty images. Larry: Can you get me a deal on... I want to just briefly tell folks how we connected. I was at the Semantic Data Summit in New York about three or four weeks ago, and you did this really great presentation about your work at Getty. And one of the things you talked about there is you have all these things, you have collections of things, and you opened with this example of you had an illustrated manuscript, a photo album, and a book of poems, and at any one point in time, each of those things, it's a piece of art, it's a research source, it's a collection in its own right. Can you tell me a little bit about why you opened your presentation with that and how these sort of perspectives influence your work? David: So one of the reasons Getty got very excited by semantic technology is we always have this idea that it would be really interesting to understand the connections between all the stuff that we have here, because Getty does two things. It collects things and we have warehouses full of things, sort of Indiana Jones style, and then we generate knowledge about those things. We have hundreds of people here who are the best in their field who know about pottery or know about 17th century silver making, and they create knowledge about that. And so we know huge amounts of stuff about the stuff that we collect, but we tend to know it through the lens of the various disciplines that generate knowledge. And so we have a bunch of librarians who are really good at thinking about the way that you catalog books and the way you capture those. David: We have a bunch of archivists who are really good at managing collections of ephemera, papers and receipts and the stuff that humans collect over their life. We have curators who describe artwork, but at the end, they're all talking about things that we have on a shelf somewhere, generating knowledge about those structured in really different ways, but that are all connected to this larger idea of how do humans create material culture and how do we talk about it and how does it influence our society? David: When we are thinking about what we are doing, it is taking that data and trying to do something to make it useful and useful in the aggregate. And so we often have to go not... My job, I think, when I'm thinking about our data infrastructure is really to say, how do we go below the level of those disciplinary practices and figure out a way that we can talk across all of those infrastructures? What are the connecting points between them? And how do we make those connecting points visible and how to make those connecting points interesting as a way to build interfaces and applications and generate new knowledge on top of it? Larry: That kind of gets at what we were talking just before we went on the air about in another context, but that your charter there is about really getting this batch of knowledge ready for other people to use. So you have a website, there's a way to access the work you do via your own thing, but this sounds like this is mostly about creating a shareable, reusable, connected set of both the information about the artifacts and then all the cultural stuff that goes around them. David: Yeah, it turns out that most of the interesting questions that people want to ask about art, like the Getty's collection does not answer that thing. People are interested in impressionism, people are interested in the art market in the 1850s. Getty is not a representation of the art market in 1850s. We just hold part of that picture. We're nonprofits, our job is to hold objects and knowledge and trust for the public, and so we're not in competition with people like The Met. They're friends, they're peers, they're helping us in this larger project. David: And so Cultural Heritage is this amazing place to do data work because my job is to give the information away and to give it away in ways that make it easy to connect with other people's information sources. And so I think again, the promise of linked data has not been fully realized, I think often because of the tension between the idea that you connect all the knowledge in the world and the reality that knowledge is commercially valuable and maybe giving it away is not the way to capture that value. But in cultural heritage, we don't have that second part, and so it means we can really play with this stuff in ways that some of our commercial colleagues have a harder time doing. Larry: Oh, that's so interesting that you're unfettered by the commercial part of this whole thing and you can be true because we're celebrating the 25th anniversary of the semantic web this year, and that's one of the ostensible failures is the inability to have a common understanding across the whole thing. But if you're in a non-commercial space and you're just concerned with the facts about a bunch of stuff that you've got,...