 So we'll just get started on the last session. So our last speaker has traveled all the way from Australia to be here today. And Mike Jones' talk is titled Somehow The Vital Connection Is Made, Developing Generous Collection Data. So Mike is a consultant research archivist at the University of Melbourne's e-scholarship research center, a PhD candidate with the University School of Historical and Philosophical Studies. He's also a research associate at Museums Victoria and a freelance consultant. So thank you for bidding in this talk and your very busy schedule. It's really great that you could have made it. So since 2008, he's worked as a researcher and consultant on numerous paper-based and digital archival projects and with academics, government departments, community organizations, and the GLAM sector to explore the potential of structured and formatic systems to support organizational and public knowledge. His PhD research looks at the history of archives and museums and investigates new approaches for capturing, managing, and disseminating interconnected, artifactual, and archival collections. So thank you, Mike. And I'll let you do your talk. Thanks. That biography sounds like a bit of a mouthful, so I probably should change it. Thank you all for coming. Although I am going to get to talking to digital, I'm going to start a little earlier than that in 1784. So in 1784, in London in December, it was starting to get cold. It was going to be a really cold winter. It was going to be one of those winters where the Thames froze over. The parks were closed down. The pleasure parks in Vauxhall were closed down. And there was a bookseller down from Birmingham called William Hutton, who had locked himself into a ticket to the British Museum. That's William Hutton there. And the British Museum, at that stage, was in Montague House, not the building that it's currently in. It was a house that had been converted and this was the entrance gate. And he gathered outside with about 10 other people to have a tour through the British Museum. And he thought he might be in there for, say, a couple of hours. He ended up getting hurried through the museum much faster than he expected. And he stopped partway through the tour and asked the guide whether they could be told what they were actually looking at. The guide said, what would you have me tell you everything in the museum? How is it possible, besides not the names written upon some of them? And so he fell silent and got hurried through. But afterwards, he wrote down his thoughts. He said, the history and the object must go together if one is wanting the other is of little value. I considered myself in the midst of a rich entertainment consisting of 10,000 rarities. But like tantalus, I could not taste one. It grieved me to think how much I lost for want of a little information. In about 30 minutes, we finished our silent journey through this princely mansion, which would well have taken 30 days. I went out much about as wise as I went in. So from very early days of museums, people have been worried about lack of information and lack of knowledge that's accessible to them. I think they've changed quite a bit since then, both within museums and in the way that information is captured. And we've been through different technologies, including internal technologies like catalog cards, this one being for a spring gun that's used to deter poachers and burglars, a fairly brutal device. They sort of laid cords out from it. And if the poacher tripped over the cord, the gun would swing in their direction and fire at them. As sort of befits the catalog card, it's got a very little bit of fielded information, and it also has some brief descriptive text, but that's about it. There are also some public catalogs like these descriptive catalogs. And this is an entry for the same item. And it's much more narrative, as you might expect, from a published catalog like this. So different sort of structure of information here already. The other interesting thing about this, which is common to a lot of museum documentation, is it doesn't have any citation or referencing on it. It doesn't really tell you where this information came from. And that's despite the fact that the author Roland Penrose, just to drop a minute for a moment, took this almost word for word from an article that was published in 1901. It doesn't provide any citation or reference to it. So this is already kind of secondhand information being presented as truthful museum documentation. As you might say, in a Wikipedia sort of age, some sort of citation is needed. And that applies to a lot of museum documentation, I think. There are exhibition spaces where we're starting to get a lot more historical context and information that we may be used to in the past. This is part of a display from the American Museum of Natural History in New York. It's related to the expedition to the north and south polled by Lincoln Ellsworth. And as an archivist, I like the fact that it includes archival documentation like field notebooks, which gives you access to sort of primary source material about processes of these expeditions and collecting and personalities that you really get through a lot of other forms of material. Documentation systems kept on improving and changing. Things started to get automated. In Australia in 1976, the Science Museum of Victoria started automating its collection documentation. It used a library catalog mainframe from Swinburne University and started producing a microfish catalog every year. But that was mostly just for internal use. And then following the merger with the National Museum of Victoria, they started to move through this technology development of Titan in the early 1980s into TechSpress. Sorry about the S dropping down a line. TechSpress was interesting. It says that the documentation of the time, and some people at the museum have confirmed this, that users could start with a blank screen and essentially create their own data structure, which meant the number of databases or data structures in the museum just exploded. The process of getting all of that, these four are just from one department, which was indigenous cultures. I think they counted something like 120 different databases before they started the consolidation process. Speaking of data migration issues, it took them about 10 years to get into KEMU because of that. And some of the results of that are still visible in that system. And these sort of systems are the start of the pipeline that goes into online collections, like the Museums Victoria online collection site. Also, it has an API if people are interested with documentation online. And some people have said that this is just kind of the exposure of internal documentation for the public. But that's sort of oversimplifying things. A lot of work goes in, as many of us know in these institutions, to shaping that documentation, to choosing what goes online and what doesn't go online, to editing and enriching collection documentation. There's also a bunch of data that doesn't make it into these. I've talked in other forums about the splits that exist in systems between archival systems, library systems, museum documentation systems. And there's a whole bunch of knowledge, a whole sort of network of knowledge that exists in these institutions that curators have in personal files that exist in other small databases or other data sets that actually are about the knowledge that we have about these collections that don't make it into the collection database, let alone into our collections online sites. So we look at digital innovations in museums. And here's one from Museums Victoria. This is a immersive 360-degree collections browser that was developed by the museum in collaboration with Sarah Kendall and an iCinema team at University of New South Wales. This is a way of viewing about 80,000 collection items using data that's extracted from that collection database. And you can carry around an iPad in the center of it and use gestures on your iPad to expand and contract these different rings of material to expand it back with them forwards this way. But the starting point of this is to take XML data from the collection documentation system and then move forward from there. And I've been to a lot of conferences like museum necks conferences, museums in the web, digital humanities conferences. And a lot of these sort of projects seem to start from a sort of starting point that's the kind of Tim Berners-Lee raw data now starting point. The data is a kind of given. We'll take that data and start trying to do things with it. But as we know museum data isn't raw. It's not just a collection of facts. Lots of decisions have been made. It's always already cooked. It's got a recipe behind it. We've decided what goes in and what doesn't go in. In a lot of cases, those are kind of secret recipes. We're not very transparent about what we've decided to include and what we haven't. If you look at a collections database online and try and work out what's not there rather than what is there, it could often be incredibly hard. So to have a look at a couple of quick examples of collection documentation just to pick up a couple of points. So given Hutton was in the British Museum, an example from the British Museum, an object, we look at that and we don't know much about it. We get the collection data from the collection database. We know a little bit more about it, but we don't come away a lot wiser than we did before. There's not a lot of history here. There's not a lot of context. There's some physical description, a bit about registration and who it came from, but there's lots we need to follow up. And there are no links here to other sources, to publications, to archives, to other sort of material, which could provide pathways into other parts of the collection or of the institution, or potentially into other institutions. Museums Victoria has some jacks from the children's folklore collection. And in this case, there's quite a lot of extra data, but it's very text heavy, very narrative heavy. There is some fielded information under this, which I'm not showing, but it's also text information about sort of the history of knuckle bones rather than this specific object. There's lots of information about the collections themselves, but it's not particularly fielded, so it's difficult to really do anything with. So what I want to have a look at in the remainder of this presentation is some of the things that we can consider to start making collections data that we provide to users, to the public, to digital projects more generous than it is at the moment. And I'm using the term generous here with a nod to Mitchell White Law, who's done work on things like the idea of generous interfaces. This is the sort of generous data side of that, of we can make those interfaces even more generous if we're backing it up with a more generous supply of data from inside our institutions. So if the British Museum is potentially at sort of one end of the scale of providing some quite traditional fielded data, I don't know places like Tupapa are doing a lot of work in providing linked information to other resources and other contexts that we can have a look at. And that's kind of the first step in this. Museum of Victoria did it in a very narrative way. Here's an illustration from a very recent blog post by Seymour Rao, the blog's called Brilliant Ideas, where it shows the sorts of things that can be wrapped around an object to start providing more of that context and information, more of that history that someone like Hutton was looking for to tell us about the object, its creator, the period it comes from, its original context that can change through different contexts. There's a whole museum context that exists here. What might things start to look like when we wrap that information around it? But there are also limitations of this in that this as an illustration is actually quite truthful in terms of what we see in a lot of museum documentation systems, that it's all that stuff wrapped around a single object. But a lot of this stuff is shared across multiple objects in multiple contexts. These things can actually be pathways to other parts of the collection rather than just being a wrapper that goes around these discrete things that we then retrieve using search or filter. Whereas from a theoretical perspective, and I won't spend long dwelling on this, but there's a whole range of theoretical material. My PhD will be out sometime next year if anyone wants to read more on that. But sort of starting from the 60s, this whole kind of movement in a whole bunch of fields from sociology, anthropology, archaeology, material culture studies, history of science that is all talking about the fact that objects don't exist as discrete things with a bunch of inherent meanings, that they're part of these broad networks. It's about entanglement. It's about moving away from kind of formalist interpretation towards more contextual interpretation. And the fact that these objects in these sort of spaces don't necessarily have fixed attributes. They have attributes that they inherit through their relations into these broader networks or mesh works of content. And museums do explore this in some of their spaces. I think it was the 13th of October at the Australian Museum. They opened the Westpac Long Gallery, which has been under renovation for quite a long time. They explicitly used the notion of entanglement as part of this and entangled relationships that exist between things. And from a visual display point of view, they bring in artifacts from their indigenous cultures and anthropological collections. They bring in rare publications. There are archival manuscripts in there. There are natural history specimens. There are objects related to the history of collecting. They're all brought together. And the intent behind that is fairly clear. It's that these things can't be understood completely in isolation. They exist as part of these complex kind of networks. So what might it look like if we start to capture that sort of stuff as part of our collection documentation and provide it to our users? So I'm going to quickly throw through a single object, which some of you may have heard me talk about before. But it's a dugong hunting rope from the north of Queensland, with harpoons attached to the end, also known as a corpois in the local language. So up here in the far north of Queensland, on Cape York Peninsula, collected by Donald Thompson, who's one of the most significant ethnographic collectors in the first half of the 20th century in Australia, a really significant collection that's managed at Museums Victoria, but it's actually still owned by the University of Melbourne. He collected other material from this sort of space. He collected other harpoons and other type of harpoon equipment. He took photographs and went out on boats with the men who were hunting dugong and collected bits of material to do with their canoes. He wrote extensively in the field. These fieldmates are actually from Arnhem Land, but he wrote continuously. And there's now around 10,000 transcribed pages with rich cross-references and annotations that were developed over about 15 years after his collection came into the institution. His specimen tags and the tags on his artifacts can contain language and cultural and anthropological information. The early documentation, including this microfish catalog, they didn't have search then where you could retrieve things by search, so they used cross-references. And we've got cross-references to photographs, to publications, to field notebooks, the TFN to botanical specimens. And that was in that documentation from 1987. And in the gallery, because this rope is on display in Melbourne Museum at the moment, we get these digital displays that include a bunch of cultural information, including about the origins of the techniques used to create these hibiscus fiber hunting ropes. And the reason for the picture of the gull is because there's a story of a being called Katara who was involved in the origins of this sort of technique and got turned into a silver gull. Though this exhibition space was developed in collaboration with a lot of communities, this particular story comes straight out of Thompson's work in an article that he wrote in 1934. I don't think it's a direct quote, but it's very clearly based on it. And that article contains diagrams of the canoe, the rope in the top, shows what they used dugong skeletons for as part of burial practices. All of this sort of information, popular press, one of the most Australian covers that you could possibly get. I always sort of looked at that and thought, there's a photographer there, so am I going, just add more koalas, it'll be fine. So he wrote about dugong hunters up in this same sort of area for the popular press, as well as wrote about a lot of his other work. And as someone who started off as a natural history, as an ornithologist, he also continued to collect natural history specimens through this period, including this beautiful dugong skull that is in museum Victoria today with one of his tags attached. So all of that sort of material wraps around these and I didn't discover any of these connections for the first time. People in the museum know about these, this is in existing documentation, it's in old cataloging systems, it's on the tags that Thompson created, it's in the transcribed notes. But when we look at what's available in collections online, and it feels like I'm picking on museum Victoria here, but I'm not, I just happen to know them better than a lot of other institutions. We get some information and some beautiful images of the rope. We get a very brief summary about dugong hunting, a physical description, some tagged fields with the links that are then fire up searches, based on that information, and a bunch more tagged fields, and that's the entry. So we don't get the cross references that we got in the MicroFish catalog, we don't get the references to the publication that tells you about dugong hunting or the source material, or into the archive, or to his photographs and things like that. We get more in the gallery space, we get more in some of the old documentation. What would our documentation start to look like if we started to capture those sorts of links by working with authority records, by linking our archival material and our publications into our objects and our other collections databases? What would happen if, as users, that information that we already knew as institutions about the source material for this was provided to them as part of that documentation, or was at least available so we could provide it to digital projects that were doing things like collection browsers or via APIs so people could use it for other projects. It's moving these systems towards more of knowledge management systems where you're trying to manage the knowledge that you know about the collection and the networks that it sits in rather than just trying to capture information about those discrete objects. And it would be interesting to think about the effect that this might have on what we could then do in terms of creating apps or in gallery displays or online exhibitions. So in pursuing some of these ideas, I'm not suggesting that we should stop the digital innovation that we're doing and kind of take steps backwards and say, no, we have to fix up our collections data first. That's not realistic. We're also doing some really great digital innovation at the moment and we are providing access to more and more data and more and more digitized content. But at the same time, I think we need to not take those projects as starting with the existing collections data and moving forward without also at the same time going back in and starting to do some of that revision work and thinking what should the next collection documentation systems look like? Should we start moving towards documentation systems that are getting away from kind of digital or automated hierarchical cataloging systems towards things that can manage this sort of complexity more effectively? And what sort of standards support that sort of work? And some things that would have already been mentioned like EAC, CPF have relational content in them. The CDOC, CRM has relational content in it. We start getting our systems supporting those sort of standards effectively and then conceptualizing our documentation in this way we could start to transform the work that we do. I'm just gonna have a quick look at a site that some people may be familiar with which is the tape from the UK who have done a lot of work in this space. This is an entry for Francis Bacon, an entry about the artist rather than this specific artwork that's just used as one of the illustrations here. There's biographical information that's managed by the institution but they also draw in Wikipedia entry information and link you up to that so they're tying you in. I think they use wiki data and harvest it in automatically but don't quote me on that. They have his artworks as you would expect attached to this. They have works where the artist is the subject, some of which he painted and some of which other artists painted so immediately you have links into the works of other artists and potentially their biographies and their items in the collection. You have film and audio material that refers to Bacon. This is all on a single webpage. You have narratives about art history and art historical terms and topics. You have digitized material from the archive including letters that Francis Bacon wrote. A bunch of related art terms so you can still do the keyword thing. You can still move through related material and discover other things that isn't, other things that aren't explicitly linked. And Amazon style, here's some other stuff that you might like that then takes you on pathways into other parts of the collection. And interestingly, and this makes me excited as an archivist, you click across and see the archival material but not only do they provide you the letter, this isn't just a reference sitting off the bottom of an entry for Bacon. They have the details for it and the archival context for it so you can get into the archival description and the artist in the artwork actually functions as a gateway into the archival collection as well as into other parts of the art collection. And this archival collection is not the records of Francis Bacon that's not as simple as that. It's the records of a gallery that he was involved in so these are letters that he wrote to that gallery and so you get into the gallery collection that takes you into a bunch of other artists. Now this project isn't ideal, it's very nice. I'm sure there are people who would like one of their own but it was also quite labor intensive, quite resource intensive, not necessarily extensible over really large collections, particularly over things like natural history collections. There's a lot of item level digitizing, page level description and it was funded by heritage lottery funding which unfortunately we don't have access to here in the same sort of way. But it is starting to build, we also get some suggestions there, it's starting to build collection description as networks and creating pathways that people can follow. It's creating collection description where we can navigate through different parts of the collections or into different aspects of the collection like the archive, like publications, like audio, visual collections and we can provide those pathways to users. I think that as we consider what we need to do in the digital space and the sort of data that we need access to and the sort of experience we wanna give to our users so that they're not stuck on single objects that we need to go further down this road and start capturing more of this information. And there are a bunch of different ways we could talk about doing that which I don't have time to go into here around things like crowdsourcing tools, around things like rethinking what happens when items get registered in collections. I'm not suggesting we back date this sort of documentation across our entire collections but at the same time, as people are doing research for exhibitions, for example, they're uncovering all sorts of collections as part of that research and if they captured it in their documentation systems rather than in a set of files that sits alongside them, this data would gradually get enriched over time. So there's also opportunities here for thinking about systems that can feed back into documentation systems. So API calls with people harvesting stuff out but also ground tripping it back in. There are people like Deb Verhoeven in Australia with the Humanities Networked Infrastructure Project where researchers can capture relationships between things and describe those relationships. What would our documentation start to look like if we start harvesting those sort of links back in? And in doing this, we start to provide these important connections and pathways that people can follow so that unlike Hutton who was kind of grieving for the lack of information that he had, we can start getting these vital collections into our documentation. Thank you. And if anyone's interested, I'm gonna post a version of this talk on my blog so sometime this afternoon and I'll send out a link on Twitter so people can have a look and follow through any references, if you like. Thank you so much, Mike. Does anyone have any questions? Any questions at all? Yeah, just wait and then we'll do you, Adrienne. Mike, when you're designing, I guess the data, like you got a choice between, you showed like the text heavy stuff before. Are you kind of aiming to kind of go through that and to pick out the elements of it that can become data elements and then use that to create your pathways? Is it sort of what you... I think in some cases that needs to be done. For example, that museum, Victoria example, with the text heavy description of the child's jacks, the toys, part of that text is actually collection level description about the Dorothy Howard collection that that material is in and at the moment in the museum, Victoria system, that's text that's replicated across multiple items that all come out of that collection, which creates a real sustainability problem apart from anything else but also means that you can't use the aggregate description as a kind of entity in its own right within these broader networks. So I think that sort of information, yeah, should be extracted out, recorded somewhere once. Museum, Victoria uses emu so it could be a short narrative, for example, as a starting point and then relating things to that so that you can navigate through that more as a network than a slide with text. People don't want that information. Something that we get a lot internally is that you're putting too much on the page and you're overwhelming the user and they're not interested in this kind of thing and it's a really fine line to balance. Yes. So thoughts on that? I completely agree. I think that the issue that I have with this sort of information is that it's not available to people who do want it and so it's a matter of working it out from the other direction and that includes internally within institutions. There's an example curator at Museums, Victoria, who started about two or three years after her position had last been filled. So she was finding a way through the collection and the knowledge about it without a handover process, without a kind of, this is what I've found out, this is where all my stuff is. And in instances like that, the amount of work that she had to repeat and the amount of time that she had just been just finding out about this herself was really detrimental to her ability to kind of hit the ground running in the job that she was doing. So I think the starting point for some of this is probably to deal with it as an internal knowledge management type of issue. Talk about it as the risk of knowledge loss of if staff leave when curators move on to other things. Are these sort of vital connections between parts of collections that are part of our knowledge about them actually documented somewhere effectively? And if not, then that creates internal resourcing issues. And then it does become a question of how do you get this into interfaces in a useful kind of way without overwhelming the user? And some of it might be as simple as almost hiding it except for people who are really digging and really want to find their way through. But I think I have heard that used before as a reason not to put this information up and accessible online. And that's just as frustrating for the people who do want it. There are a lot of experts who don't work in museums who want to know this stuff. So you talked a little bit about the different kind of standards for libraries, museums, and archives for describing this material and making it available. Yeah. Further thoughts? I'm sort of... I don't really know the answer to that, but I think that the, this is a kind of cop-out answer, but I'll give it anyway. One of the main things with this sort of work that I've talked to museums, Victoria, about a lot is that there are legacy systems and legacy practices that mean that places need to try and implement this to a large extent at the moment with what they've got. The solution to this isn't, well, you're going to need to check out your system and we're in a whole new standard and a whole new system to do this with. It's, if you conceptualize things in this way and then start looking at your documentation, what can you do differently to what you're doing now to start moving in this direction? And if it starts to prove itself over periods of time, then there become points of, are we using the best standard or are there parts of the standard that we could be using more effectively than we're using at the moment? There are also standards around objects and things like that, but when you get into standards around things like people and corporate bodies and those sort of areas that archives have done work in, is that as effective at the moment in museum documentation or standards is an open kind of question. What I want to do at Museum Victoria with something like the Thomson Collection is, for example, I showed those different databases that came out of text breasts that all got combined. But despite the fact that that happened, there's something like 15 authority records for Thomson in EMU because there's the one that came from ornithology, there's the one that came from marine vertebrates, there's the one that came from somewhere else and somewhere else. The first step in this is starting to consolidate some of that stuff and then we can talk about what standards it goes into and what happens. There's some basic data structure things, I think, that are the starting point to this. That's the kind of first step. And when you work out what data structure you want, then you can look at the standard and evaluate it more effectively. Well, thank you so much, Maya. Thank you. Thank you to all the other speakers as well. We now have afternoon tea in the Oceania and that is being sponsored by the Ministry for Culture and Heritage. So thank you so much. Another round.