IIIF: Collaboration and Community Built Technological Innovations

This panel will look at the lessons learned from the experience of IIIF practitioners in terms of what it takes to institutionally adopt IIIF, and what it takes to collaboratively define the ever-evolving IIIF specifications. The panelists will share experiences from deploying IIIF as a shared standard across different collection types within an institution, such as different curatorial departments within a museum, or across several GLAMs collections within an institution, such as an academic campus. Questions to be explored will include: what are the strategic motivations behind such institutional deployments? What are the challenges of a collaborative approach to designing multidisciplinary collections interfaces? How are user research findings balanced with the temptation to roll out innovative interfaces? What are the implications for teaching and learning? A high-level description of the mechanisms governing the IIIF community will be also provided, including an overview of its international community groups and organizational structure, as well as its distributed model for the development of its technical specifications. The panelists will reflect on how the IIIF Consortium can engage more deeply with GLAMs, and museums specifically, especially on issues around technological requirements.


Unknown Speaker 00:00
All right, it's four o'clock. Welcome, everybody. Hi. Welcome to this panel on job. If so today we have three presenters and we have 30 minutes. We do hope to have time for questions afterward. But let me introduce my fellow panelists. On my left is David Newbery who is an enterprise software architect at the Jay Paul Getty trust. And David will introduce a triple if just to demystify it a little bit before the two talks here. And then we'll wrap up with what's next for TBF. On my right, is Thomas Roche, who is director of it at the Yale University Art Gallery, and Thomas will introduce a large collaborative project around trouble if at Yale, and then I'll follow up with at the Yale Center for British Art, and I'll follow up with a use case of triple AF adaption at the center of as an early adopter and what the challenges and opportunities come with that. So without further ado, let me turn it over to David.

Unknown Speaker 01:10
Hey, there, everybody, I'm, I'm here to give the five minute version of what triple F is we've manual, and I've been giving versions of this conversation for quite a while. And so I I'm hoping that this is a a sane version of what triple F is that we can do in a short period of time. And if there are people who want long versions, I'm always happy to talk far too long about this sort of technology. So when you think about triple AF, what you have to think about is that what it really is are a couple of standards. Triple F at its core is a community of people who get together in a room and argue for a long time about how to write a couple of web pages that describe the structure of a couple of JSON documents. But what that really does is it helps solve really standard problems that we have in the cultural heritage community around meeting normal use cases around images, because what we found is that a lot of what we do is dealing with images and how to deliver them up in software. And so the first part of triple F, and the easiest part to explain is the image API, which is really a standard way to get the right set of pixels out of a bigger image. And so if you look at this lovely picture of an iris on the left, there's a triple F API call, which is just a way to say there's a standard URL string and put into your browser that will give you this image. And by changing the little numbers in there, where it says full full zero, if you change it to say I want a particular rectangle inside that picture, and I wanted a particular size or a particular rotation, you can say I want that little rectangle there and get that different thing. And so there's lots of systems that do this work to generate a particular region of pixels to give you on the fly derivatives of images. Triple F said, let's just do a define a standard way to ask for those derivatives. So that if you asked for one from Getty, or from y CBA, or from Yale art gallery, you can do it the same way. So your software can get pixels from any system that implements Kripal, if using the same software in the same patterns. And so that's the first problem that triple F tries to solve. The problem that triple f the second problem that triple F's tries to solve is the triple A in the in the presentation API is a way to describe sequences of images. Because we know you often don't want one picture, you want three pictures in a row or five pictures in the row or you want the pages of a book. So what's the standard document to say this picture comes before this picture. This pictures are sort of book like. And also to give a little bit of context for those images to people. Triple F isn't a way to exchange metadata between systems, it isn't good at that at all. All it does is say I've got this pictures. And here's a bunch of texts that you could read, if you wanted to figure out what these are. And it helps preserve context around images. Because we know that's a really important thing to do. It also helps manage rights information and lets you put in standard rights agreements, credit lines, institutional logos, the things that we all know we have to do to meet our licensing agreements. They've talked to lots and lots of people in the field to help understand that. The last thing that presentation API does is it lets you annotate these images to a sort of a standard way let anybody make a comment on a particular image or a particular region of that. And also a standard way for me as a publisher to provide back some subset of those comments to you. Um, any one of you can write a comment on any of the Gettys images. Triple f does not mean that I automatically publish that for you, but I can choose one means that we provide or to provide access to other people's annotations.

Unknown Speaker 05:05
There's a couple of other API's that are there, but less commonly used. One helps you search within those annotations so that if you had a book and you were looking for a particular words in that book, how do you go from those annotations to the page of that book? And the other one is, if you have some of these that are rights restricted that you want to authenticate, what's the way to discover what the rules are about that annotation. And so, at a core, that's what triple f is, it's a way to get pixels, it's a way to talk about sequences. It's a way to annotate images, a way to find particular pieces of content within it, and a way to help talk about the rights of it. And why should you care about this, because there's lots of tools that do this kind of work. But triple F really is, is it's a standard, there's not a vendor tool that you're using, there's a whole ecosystem of tools. And it's a system that works together, all of these pieces are designed by the same people to work together pretty seamlessly. So you can get all of these pieces out of the box the same way. But the most important thing is that it's a community of people. It's designed to help these systems work together. And also to let us build on each other's expertise. So I don't have to write an image viewer or an image server or an annotation server, or an interface, because other people in the triple F community have already done that. And it means also, not only are the software tools there, but the experience and the learning from each other is really, really important. And so at that point, I'm going to hand this over. And we can talk about how this has worked in some of the institutions at Yale.

Unknown Speaker 06:42
So we're gonna take a little more, we're going to take a huge divergence and talk strategy for a little bit. The talks description had the phrase institutionally adopt AAA if. So what I'm going to do, and I'm going to try and do it as quickly as possible, is act as a manual sidekick, to talk about the Yale environment and paint a picture of where we've come from and where we are. So when she talks about triple if what we're working towards is a campus solution, even though we are very many individual silos on the university. So that being said, we have to go back 10 years, the cultural institutions, specifically BAC and US started working together to form a dam when it came time to purchase the Peabody Natural History joined us. So we have a long history of working together with the dam, we've created services on top of the dam. So we've had various working groups throughout the years, and, and lots of partnerships within the cultural Heritage's at Yale working together digitally. And in the physical spaces, as you can see listed there. We've been working together on a variety of levels in the physical space as well for teaching research. And then the last item I have, they're fairly new to Yale, we're working in the conservation area, all of the cultural areas have Conservation Departments, there's now a central, cultural, excuse me, preservation and conservation department, we are still trying to figure out how we all work together. Again, we're individual silos. But there is that that physical space that we're also collaborating on, we jump forward to today, and the library has joined as a strategic partner, we now have our second generation of the dam, we have gone with net X and preserve occa as a sidecar to us to send items for preservation off the side. Our first dam if you're interested was open tax, we have moved on to net x. So we're continuing. Now we have that strategic partner of the library. With this working group over the past two years, it's roughly two years. And I yesterday, we had our phone call, and we're working on our second fiscal budget and the request for that. So with the library in this new working group, we're in about our second year. With that we're working on the cross collection discovery. It's a personal dream of mine, one stop shopping for a search engine that crosses across all of our individual areas. That's something that we're working on. It's funded and we're making great progress on it. And then there at the bottom, we're gonna circle back to triple f we are working on Triple AF as a service. This will be something that is managed centrally that we can all take participate in on it goes hand in hand with the cross collection search. So you find an object, you're going to be able to then be able to look at that object with triple AF. How do we get along there? Very quickly how that all came about is a new CIO came in to campus worked with the provost. And the six silos were created as functional silos across the University. One that you'll see there is cultural heritage. So we now are at the level of the provost. And we have funding that comes from there, if it's approved. So, in this working environment, we've had all day think group sessions, where partners from across the university, and this is how we came across the cross collection discovery, as our priorities and how the fields were defined for the cross collection discovery. We've done focused reports, we are doing education across the university, which I know Emanuelle will talk about. And the important part is we continue to have senior leadership bought by in the director of each of the museum's meets the director of the iron library meets, we have monthly meetings, and this is an ongoing priority for the university. What's next for Yale, we continue with our various projects, linked open data, aaa, F, these are all continuing these, we were working on grant development, we are waiting on pins and needles to find out if a grant is going to be improved. And then like I said yesterday, we're actually in phone calls working on our next cycle of budgets. And then the physical presence we can't ignore, we have a huge physical space at Yale that we are all collaborating on not only on existing physical space, but we're planning out future physical space. Speaking for the art gallery,

Unknown Speaker 11:29
we have a lot of catching up to do. Our our gallery has not stepped foot into the triple if space. So we're looking forward to the new services that are coming forward that we're developing as partnerships with across the university. Along with that, we have to clean up our own data. We call that the fields of dreams. We've we've labeled that to work with our curators and emphasize to them that we need to get our data act together. And then of course, as we begin to work with our new website, we're rolling in these features that we're seeing coming down the pipeline of triple if linked open data. And with that, I give it to a manual.

Unknown Speaker 12:05
Yeah, I think I'll use this microphone. How's that sound? Okay. So as I've said, the way CBA has deployed to buy a few years ago, so I'm going to take a look at the current project from the perspective of an early adopter. And so what does the Jabra have history looks like? What does it look like for the ICBA? So actually, Yael as a whole became core founding member of Tobiah in 2015, which means that actually, yeah, university paid are still paying dues every year to support the development of the API's and the community in general. So back in 2015, there were 15 core founding members. We are up to 53. Today. Does that mean that everybody needs to pay to come into trouble? If no, that's absolutely not a requirement. But it is important for the development of the community, essentially, the David has talked about how the community has grown. And actually, we're about to hire a third full time employee for the consortium, who will be an Events Coordinator. So the the consortium is really growing. That same year, the way CBA then became the first unit at Yale to deploy trouble if two API's the image and the presentation IP API's. And the way we did it, is that we flip the switch for our online collections catalog. And that's one of the options that other museums have done. But it's not it's not, of course, the only option. The VNA has been doing a lot of very interesting experiments with using Tobiah, for exhibitions, online exhibitions. So you know, more focused, had ad hoc kind of projects, which are very, very interesting, I encourage you to look that up. But today, what does it mean for the way CBA? Well, we have about a little bit more than at that at 1000, triple F images for about 35,000 objects that are in the public domain. So we do all that rights vetting work before making those images to buy if compliant. And that project essentially, releasing two pi F to online collections catalog totally meets yells open access policy. So fast forward to 2017. Thomas has highlighted and we've we've now have a cultural heritage it pillar at Yale, which really enables us to have deep conversations about collaborations and deep structural changes on how we will support teaching and learning at Yale using in technology, and Tobias is definitely a core element of that. And of course, the big announcement here is that Yale as a whole will be deploying Trump if probably in early 2020. There was a nice typo there. Sorry. But essentially triple AF. So the content delivery is a service that will be triple if enabled at Yale will be the first tangible element of the year wide cross collection Discovery Service, which is quite a, quite a feast. So I'm going to run through what Tobiah looks like for the white CBA to kind of give you a tangible idea of what what it what it looks like. So the way CBA uses Domina, door open source viewer downloads of yours you can choose from and at the end of the talk, we have a link to a lot of very good documentation. And you can have a look at the other viewers. On the right there you see a panel that essentially describes the the object and that metadata comes from our collections management system. And that part of the triple F ecosystem, if you will, is called a manifest, which is essentially a resource where you you you have data to describe the objects and the images in in the viewer. Mirador is very versatile. You can have as many frames in that viewer as you'd like. It allows you to pan and zoom and do all sorts of things, you'll have five images, I could have more, I could have just one. Here, I'm comparing abroad in images from the Getty, and the cobia, museums, both early adopters, of course of the triple AF standard.

Unknown Speaker 17:02
And that's a very nice feature. Obviously, that's essentially the core feature of Tobiah. F is this interoperability layer that allows you to bring images from other museums, or other institutions, libraries, or archives into your viewer. And that's a round trip, of course, to get you could bring our images into their viewer as well, even if it's not me, I don't actually the manifest there on the right, again, as David mentioned earlier, has our logo, but it also has nice links to our machine readable data. There's a link to our XML and another one to our RDF, which means that essentially now the manifests become kind of a gateway for external services to come and scoop up all data, right, they don't actually need to knock on the door, or call or even come to our website, if they have a whole other manifests, they can have a hole, they can come and scoop up machine readable data. And here, this is a quick slide just to illustrate what David alluded to. Just before, he was showing you a change in the URL here to zoom in on a region of the image. And I've just changed to perimeter of the image API to flip the image upside down. That's one ad there in the URL, and it changed the other parameter to gray, I could have done bitone or something else. And this is without downloading the images at all. I'm just changing the parameters that are in the URL. So fast forward to 20, June 2019, then yeah, where are we? You can tell we've done a lot of work on campus to kind of get buy in. And this was a really nice day where as Thomas said, the directors of the various collecting units came together to a day of presentations by yell staff members, as well as external, triple F experts, really to announce Gannett, this large change on campus, revolving around image interoperability essentially. So a little word about collaboration. And this is a slide that dates a little bit actually, this was produced by OCLC probably around 2011 2010. But just to reflect for a minute about what collaboration is, right, and we're very lucky right now. Yeah, we were in a good phase where we both have a mandate a specific mandate from the Vice Provost for collections and scholarly publications, as well. So as highlighted, right, we have funding, we're really happy to work with our colleagues in central ATS. And we also have a problem to solve that actually addresses a long standing frustration of the various collecting units on campus, which was around this inability to exchange images properly. So this is this is great, but what happens when the cycles of deep collaboration don't happen? You know, and I want to argue that actually, it's, even though there might be those low times might be a little bit frustrating, a whole lot frustrating, it's also a good time to get to, you know, go check out your colleagues across campus and get to build those human connections, and so that when funding and skills and you know, all of that gels or gels together, you're ready to go. So what does the triple AF, content delivery service infrastructure looks like? So, we're very much in the final stages of defining this specifically. So I'm not going to go into any super details about this. But what I can say is that, of course, we're building on a previous instantiation of a content delivery service. And that's maintained by your central ATS. The that CDs will be shared by the museums, the library system will develop its own proprietary implementation. And even though that might sound as a little bit of a sad note, I actually expect that that's not going to be a big problem. And the reason is that there is very strong collaboration around developing a structure and practice for building manifests between the libraries, archives, and museums at Yale.

Unknown Speaker 22:08
So building blocks, we said, we've recently migrated to a new dams. And we are now using NET X. And that's a system that will send metadata technical image, the technical metadata to RCDs on another building block, that has not changed and will probably not change for some time. I'll the respective Museum's collections management systems, right. That is a very stable layer. The Art Gallery and the center are using GMs. But our natural history museum actually uses emu. Challenges. So perceived challenges, I should have said, Actually, we were a little bit worried because we hadn't implemented Trump if at the museum for a few years. So we're a little bit worried that that might mean deep changes in workflow for us, we'll see that it's not exactly true. What it does change is the how do we explain to users how to handle trouble? If right, well, they are not limited to asking for specific sizes anymore, they can ask for sizes up to a certain limit, right? So there's a there's a variable here. fairly big change actually is the the change in the source image format, we use to use JPEG 2000. And we're going to change to pyramidal tips, I want to point out to a nice survey that the Getty did on the benefits of the of pyramidal tips, if you can look it up. Another big change is putting in place activity streams. So that's totally new for you. And that's that's going to change the way we interact with aggregators and the respective museums are actually going to only you know, express that the changes in their metadata or images or their deletion or new creation to the CDs service, so that that service does not have to harvest everything every single time. So that's that's a that's a that's a large piece. Another challenge is this compromise on data around triple AF manifest and dissenter had experimented a little bit with triple AF and layers in trouble if and that's that's not going to be maintained for the next phase of the project. But we will enjoy experimenting so I think it's it's probably fine at this point. So just to highlight a few opportunities, this change in our perceived change in workflow is actually not so much a big deal we're still very much in control of. And that's true for the gallery as well, how we assign access levels to images in the dams, that's going to stay the same. What the conversation around triple F has brought, though, is is nice investigation of a workflow, business workflow tool, we're looking at gobby. But it might be something it'll be different. But that's going to probably going to be a shared tool. And the center, as well as the other museums are still very much in control of the metadata they will send to CDs. So you know, it might be different data, it might be different levels of of metadata. And that's up to each museum to decide which is very nice. In terms of user education, well, yeah, that's, that's a piece we have to solve, have talked about the performance of p&l tiffs. And we're extremely happy actually to have activity streams coming our way that's going to make interacting with aggregators, so much easier. A big piece of the opportunities here is a conversation around the adoption of REITs standardized licenses and true, and I want to go a little bit in depth about about that for a minute. And then two slides away from the end, you'll see here, the top three, three elements highlighted in bold. And that's basically the required the required elements for the yellow manifest. So it's really, really light.

Unknown Speaker 27:03
What the way CBA is suggesting to our partners is really is also an additional element of license, which requires a your eye link. And that's the one field that actually has brought on this wonderful conversation about harmonizing how we handle rights licenses, and, and we're looking at Creative Commons licenses and rights statements. And that's a huge leap forward for the your collection with each collection. So far, express, you know, our own kind of local terms for use, which can be a little bit problematic for users, right. So I want to end with this slide of people talking and working together, because I think collaboration is a big, big part of this, of this story. And leave you with the thought that really triple if has still is a lot of work, but really is a nice opportunity to bring us to a nice best practice level in terms of how we disseminate our digital assets at Yale and beyond.

Unknown Speaker 28:19
And just to quickly wrap up and to go, sort of backup to the high level about what what is coming next for AAA if as a community as a whole. I think the biggest news that's happening is that the newest version of AAA, if is, if things go well coming out this week at the triple F conference. It's happening in Arbor right now. And the biggest difference is that triple F is is now aware of time because one of the things that we've heard over and over again is, wouldn't it be nice if the sort of interoperability benefits that we're getting for images also worked for audio and visual video. And so over the past four years, the community group has been working on understanding what it would mean to express that. And we're very happy to say that we now have a version of triple if that will work for this along with new viewers and new tools that will help with that. Another major thing that is coming is this discovery API, which is built on the activity streams that Amanda was talking about. It is really the tool that we need to support aggregation, which is useful both for large consuming things like DPLA, or your piano, but also for this cross institutional collaboration or even cross system collaboration. It's a really, really useful tool that will help with that process. And as the manual said, there are all of these new viewers coming out. And they often aren't. What we're seeing is the first wave of viewers where we want something good for books, we want something that's good for comparing images. What we're seeing now our viewers that are are much more specific. One of my favorite is from an organization that's here called gap that the AAA F viewer that's designed to do sort of slideshows through through art, and give narrative experiences, new ways to start using triple F now that we have so many resources this way. And I think that's the most intuitive thing that I'm seeing coming out of the triple F community right now, is that over the past six years, we've put a huge emphasis on publishing and on standards. But we're seeing now are more and more people who are using the images in interesting and different ways to tell stories, to do transcriptions, to teach science to, to create art. And I think that's what's so interesting is that we've put this infrastructure in place, and we're now using it to do things that are completely out of the realm of what we were thinking with. When we started with AAA, I started to do books, and now we're doing art installations. So thank you all very much. Here are some useful links to help you get deeper into the AAA if and were around for questions.