Anne McLaughlin 0:03 Excellent, thank you very much Mia andThank you really to all of our panellists for agreeing to be here. And for everyone in the audience would love it if you keep popping questions into the chat and we'll continue to monitor that. But, to kick us off we've got a question from Cathy Aster which is quite specific. For Davy, she's interested in hearing about the successes and challenges with importing and displaying IIIF images in an omeka s instance. But I think for all of our panellists that might be worth considering challenges around sort of displaying your research, presenting your research in whatever way you choose to go about that. So Davy, why don't we kick off with you, but then everyone else, feel free to chime in. Davy Verbeke 0:48 So good. I didn't see the question in the q&a, could you Anne McLaughlin 0:52 just repeat or attend to us interested in hearing a little more about successes and challenges with importing and displaying IIIF images in an omeka s context, or really in any context to present or display research Davy Verbeke 1:12 well with soy presented to project with the Omeka project we actually produce our own IIIF manifest so all the students that were working on the Omeka project, they uploaded their images and so we produced with a IIIF module, and each will have manifests. Now with the old Storiiies app, maybe some people will know that you had to give the IIIF app manifest, now it will be generated by the by the app. But, so we added two IIIF modules. Now the MADOC MADOC project so MADOC. The software is omeka based, but the. So, it makes for a full IIIF compatibility. So within omeka and this is actually as Omeka as base but with the second project so the Congolese comics project. We don't, we don't expect any issues with this communication because it's fully IIIF based, and so we will also be moving away from the omeka s for the for the users. So, I don't know if this partially answers the the question. But, again, for the, for the second project we will be moving away from this omeka environment Andrew Prescott 2:50 is very interesting. The things that come out from the discussion so far is the variety of use cases that there might be under different types of audiences. And I think that maybe for the more detailed, very niche research that I, or possibly peer would want to undertake. At this stage, it's difficult to envisage presenting that other of the via conventional scholarly means publication. And it's, it, there isn't the, you can see the possibility to IIIF to present that research, but you kind of either need to set up a very specific project with a particular piece of research by a warrant. Or you'd have to go down I think a conventional route, it's, it's kind of got not quite maybe there yet, that you could readily present a piece of research without setting up a very specific project. Anne McLaughlin 4:02 What would you need to in order to have that sort of display that you're looking for Andrew Prescott 4:08 In the particular case, the very small piece of research that I did, but it was interesting because it's actually reappraise the relationship with some manuscripts. The first thing I would need Abira's discussed the reasons for that would be to actually have all the manuscripts that I was interested in in a IIIF format that usually that's the main buffer that you hit, you can usually go along quite nicely, similar work with some tools or manuscripts, you can go along quite nicely with all three manuscripts that you find you want to relate it to a fourth, and you find that not available in IIIF form. It may be digitised but it's not very well IIIF a form that's a very frequent buffer I think, in my research, Joe Padfield 4:53 Andrew, you'll be glad to know that it's happening but it's happening in library time which is measured in years not months Andrew Prescott 4:59 It is understandable. What a huge undertaking is to retrospectively, convert the huge quantities of digital data that you and the library have regenerated, I think that everyone's aware of that, but it does limit what you can do from a research point of view at this stage Joe Padfield 5:15 I don't suppose that there'd be anyway of sort of setting up a wish list. Any way of organising the priorities for digitization. Mia Ridge 5:30 Maybe there's already and internal list. Joe Padfield 5:32 Well I'd assume there would be, who wants which which image to deal with first? Andrew Prescott 5:40 And the quantities are enormous and react to the resource I use most is one that has material from the National Archives on and that holds something like 15 million images that will ever be available in IIIF form. I think it harks back to what we were talking about, that Tom was talking about, the need for industrial scale and implementation as well. Anne McLaughlin 6:08 Tom would you like to follow that on. Yep, sure, I'm not a panellist. Tom Crane 6:18 Here is, it's the balance between small scale stuff and industrial stuff, but also I think I mean the biggest thing is tools. Now that, in that sense, IIIF is in its infancy. It's a project, it's an enterprise undertaken by organisations generally don't have a lot of funding to develop the kinds of complex software that people expect to be able to use you know if you were, if we were talking about making presentations for each other, but no one had invented PowerPoint or Google Slides and we were trying to cobble it all together ourselves every time we gave a presentation, you kind of want to remove all the technical. No one talks about the technical underpinnings or file formats when they're making a PowerPoint presentation. Eventually IIIF should be the same size to where, therefore I can use it in this tool. This tool. This tool, pick whatever when you want, and it's just a very early stages of that. It needs, needs time and money spent on tooling. Joe Padfield 7:16 I think there's a broader issue as well in terms of thinking in terms of like service design style user journeys where sometimes I know an institution has done something with IIIF, but it's difficult to find those collections. And I know from the British Library's catalogue. People miss the cues in the labels and the faceted aspects of the catalogue that let them know that items are available. The digitised items are available outside the reading room. So it really takes a commitment across the organisation to not only do all that infrastructure work to get your collection digitised items into IIIF to have some kind of UI to update your catalogue, but also to think in terms of how do you let everyone know that your collections are available in this amazing format that can be used in so many creative ways. Tom Crane 8:13 Maybe maybe the index international collections. Possibly. Ben Bakelaar 8:22 I can comment a little bit about that in terms of user journeys and visual cues in the IIIF community as part of the Discovery for Humans Working Group. We're in the process of conducting original user research, and that will be presented actually in three weeks at the annual IIIF conference, so everyone should tune into that. But it's absolutely true that users themselves need cues, and, and even a working mental model of what IIIF is and what it can do. And a lot of times the current generation of implementations has, it's an afterthought, at most, right. And so we're starting to see some emerging patterns that are more familiar to users. And I think there's a lot of room for improvement in that usability arena. Andrew Prescott 9:21 What that says that in my experience as a user, this is more often a question of smaller archives, who got a quite limited. Implementation of IIIF rather than, you can usually. If you're determined enough wrestle it out with what's going on with a big collection like the British Library, or one, as my says one gets used to the cues, but in smaller collections, it can be very frustrating to you can think you're in a fully IIIF environment then you try and do something that you'd expect. And you find you can't, and you dig down and you can see that the implementation is peculiar, for some reason. And that's often in much more very small corrections, and institutions. Anne McLaughlin 10:09 So is that possibly an argument for some sort of standard way of displaying IIIF content, whether that be at a national level, towards national collection, or even Andrew Prescott 10:23 outside, and it would be interesting to the view of the other panellists, how far outside the sort of large institutions that we've got represented here. How far was a genuine commitment to the open aspects of IIIF, and whether there were more institutions that are just doing it for internal housekeeping reasons who don't necessarily want to see the image shared in the ways of proceeding in IIIF Joe Padfield 10:55 Just picking up on a question in the q&a about the relationship between open access and IIIF. Because IIIF was built with the understanding that sometimes you need to put limitations on access, it actually enables more access than another tool might because people feel confident that they can put parameters on where they need to either in terms of I know Wellcome you need to login for certain materials so that they can give you Terms and Conditions about material related to living memory or resolutions of images available. So I think that because IIIF presents an understanding of the institutional context and institutional fears about what happens if we're too open it can actually enable more than things which are just like everything is open, it's, it allows for more shades of grey and that can be really necessary. also facilitates institutions to be able to keep track of how the images are used. This speaks to justification and resources as well. Because if you say well we spent all this money on presenting these in IIIF and then you can actually point to see all the resources are being used, they're being used for research, used for engagement, right down to which bits of images being searched for can be potentially quite useful. Ben Bakelaar 12:30 One additional comment I can make is that as we're thinking about standards or guidelines in the IIIF community for how to present the resources it. It's an open source community with many, you know it's a consortium of institutions, right so it's very hard to disseminate, any, anything close to a rule. Whereas something like a national collection can unify and standardise and present in a single interface right, just like Google does a certain way of getting at those resources, and so that's maybe one of the benefits of building a national collection, a national interface is that you can provide that standard, even when you can't necessarily get all the institutions to adopt the same standards. Mary-Ann Constantine 13:22 Of course that depends what you mean by a national collection in the first place in a multi-nation entity or multi-country entity like the British Isles a lot of my material would want to link to something like the Paul Mellon in America, or to natural history collections in Europe. I think it's got to be about crossing borders as much as creating. Joe Padfield 13:57 within the wider TaNC or Towards a National Collection discussion, trying to discuss what national collection meant in the earlier stages went round and round and round. I think it was generally agreed that it was looking more about collections of national interest as opposed to collections of national things. So speaking of how collections connected to the wider world became very important in telling the story of the collections that may just happen to be in United Kingdom, Great Britain at the moment, so that that was recognised and highlighted as being a topic that would require further discussion, the issue being initially that Towards a National Collection within UK funded academic stream. So the focus didn't need to be at least primarily on the UK. Anne McLaughlin 15:09 All right. Um, so, before this question. Maybe most one checking the q&a very quickly. And does anyone else have anything they'd like to contribute about the idea about open access and IIIF how those interact, either their own research, or things that would like to see in the future. All right, seems like there's nothing else from any of our panellists that they'd like to contribute. So before the panel today we as a project, had a few questions that we sent out to our panellists for them to have some thoughts about, perhaps, prior to the discussion and one of the ones, what I'd like to pick up at this point is what role do you see to play in post-capture enrichment workflows, like machine learning or crowdsourcing, do you see further potential there? Davy Verbeke 16:10 Maybe I want to comment on that, something I've been thinking about that, so I primarily think from this educational perspective so I'm setting up student projects now if I would sum up all the projects that being done yearly at our university with for example the library with our own collection of our university library that's a huge amount of work being being done. So if you would have this feedback loop of for example student projects in coordination with for example the library, you would have these enriched collections so it can be an extra argument and our own university library is using IIIF but also for smaller collections when there, we've been talking about scattered collections about smaller institutions. I think Mary Anne was talking about the the narrative so that you can have objects from some small institution, museum, whatever, it can be an extra argument for them as well to know, and it's something Joseph said as well to know that their material is being worked with, if you have this feedback loop. So for example with MADOC we can we can produce the new manifest so the enriched manifests, which can be integrated in those collections that they adopt standard of course. But again, it makes visible that their collection is being worked with online and not just some paper that is written once and never read again. So I think this is really something that can and it's something that I don't know that is being done right now, but it's something that can really make for a new relationship between teaching practices and libraries and collections. Using IIIF and in my case, the MADOC platform. But maybe I'm being too optimistic and too ambitious and maybe the consultation won't be as as easy as I envisioned it. But, who knows. Andrew Prescott 18:15 No, no, you're being at all Davy. I really think that's part of the creating those new dialogue thoughts, and new ways of working is part of the excitement, the IIIF. The question is how we, we fit that to different audiences, and how that different works in Anne's question to ask is also raising deeper questions I think crowdsourcing, yes, there's a lot of potential there. But around machine learning at all there, there definitely has to be potential when we start to develop large scale corpora of the sort that Davy is describing the possibilities, dealing with approaches to collections are ways that we haven't yet done. And in a sense I wonder whether we actually imagined enough of that. Whether we need to we need to be thinking about how we might work with the collections that were different, in that, in the sort of context of those describe that Davy is describing for us. Joe Padfield 19:27 At the library. I was the first client about to play as content I used the manifest from the playbills collection which were digitised from microfilm, to build in the spotlight so that's kind of natively IIIF, in a way and we store the resulting transcriptions and marked up regions as W3C annotations, and using the IIIF manifests allowed us to use those high quality images without having to store them separately which would have added platform costs, so even sort of to that extent it made life easier. At the moment I'm trying to see if it's possible to import to use those IIIF manifest in Zooniverse because that would enable further work too. But I think there's a wider issue again in that, if we are doing things like crowdsourcing you'll have annotations created in different contexts to different levels of quality because that's always the first question that people ask about crowdsourcing, you know how you define quality depends entirely on your goals. So thinking about systems that know about your collections data not only in terms of directing people to online versions of content, but also in terms of understanding where there are crowd sourced annotations whether machine learning, created annotations as well, because a lot of the machine learning image classifiers, for example will produce terms of just aren't appropriate for these sort of older and more culturally varied collections, they tend to have a very 20th century North American West Coast idea of what things are sorry 21st century. So I think there is a need for greater IIIF literacy and awareness in some of the systems that we use to really make the most of the content that we get because context clash could be quite, if not a worry but say Andrew's need as a scholar to compare different manuscripts, these annotations in that context, aren't necessarily of interest to others who are looking at different aspects of those collections and he might not want to share what he's working on. But having systems that really in a very embedded sustainable way. Understand what IIIF is and what the potential of it could be is kind of key to helping it succeed. Mary-Ann Constantine 22:05 I would just jump in here and say the work that the National Library of Wales has done with crowdsourcing has done two really wonderful things one is its produced it produces masses of data on things like place names, which is hugely valuable for the scholarly community and for big data crunching are more and more humanities projects need vast amounts of data in order to tackle research questions in a variety of ways. But then the other thing it actually does is it creates a community of volunteers who are local people, as Jason says working, often through their native language. And when I'm hoping to set up a crowdsourcing projects on the Pennant of manuscripts in the pen pictures to actually get local input into that building, and that river, that bridge, would for me, be a really exciting result, and to kind of feel things coming from the inside from a human point of view, as well as that massive, wonderful, exciting, big data, that you can do so much with as well so I think as Joseph just said you know one operation allows you to go in very very different directions from the local to the global. And I think some of the things that you've said this afternoon really bring that home to me. So yeah, thank you. Mia Ridge 23:41 I think Joe Padfield 23:42 the things like place names, another project I work on is living with machines which is broadly, data science meets digital history applied to digitise newspapers and other sources at scale. And some of the wonderful computational linguists on that are doing toponym resolution, linking to Wikidata, so really creating robust records. But those records exist for a tiny subset of the digitised books or digitised newspaper pages that we have. So, thinking about what that means in terms of the lumpiness of the experience were some items that have been looked at in a great amount of detail, either by scholars crowdsources computers in some combination will have really, really detailed information that the systems that we have can't support that level of granularity in terms of discovering that content. But we do need to think about how will work at. When we have material that has been annotated or looked at more heavily than than other material, because ultimately it will start to skew the user experience in some perhaps unexpected ways. You know we're aware of the canonization of the recanonization of things that are digitised and how that makes them more prominent and scholarship. So I think knowing that that might happen we could think about how to prevent that happening again. With that kind of IIIF-ified attention that might happen. Mary-Ann Constantine 25:19 Really good point, yeah. Andrews got a very good paper on how digitising a whole bunch of texts actually produces rather skewed results in searches for various things is really it's a really good example of exactly that. But yeah, thank you. Anne McLaughlin 25:39 Speaking of, kind of working at scale. We have a question in the chat from Yvonne Lewis. Asking us to what extent for existing commercial microfilming or digitization projects, restrict what could be made available in IIIF format. For example, there's been quite a bit of discussion in the research community about the quality of an 18th century text on the Early English Books Online. Ben, I know you work with a lot of film and microfilmed materials, is this something you've run into Ben Bakelaar 26:08 Are you asking me? I'm involved in the, the web presentation and the metadata preparation but not the actual digitization. So, Yeah, I couldn't comment well on that. Joe Padfield 26:25 I think a lot of that may be an issue of who funded the digitization in first place. I think we're not really going to get into it now, because we're gonna have separate discussions in the future but the whole issue of IPR and licences and all this kind of stuff around the use of IIIF is going to be a big issue for future discussion Anne McLaughlin 26:54 Pia, I think you have a great question. Would you like to ask it? Pia Eckhart 27:01 Yeah, I was just like, commenting on what Joseph wrote in the chat that IIIF you store once and use it many times but that is right as long as you just use the images, like, on the web, but if you annotate, and you are trying to, sorry, I got lost there. For me it was very attractive to not have to get the pictures because that is a lot of money that I just don't have. So that seemed great, but then you realise no, if I want to do annotation, and somebody moves their pictures in the institutions and I think we have heard today, how different approaches there are in institutions to IIIF and storage and preservation. Then my annotations are lost, broken because the canvas return is just not there anymore, so I still have to store the pictures locally, right now, and that is for me it's a huge problem. But they will not maybe somebody has an answer to that that would be great, Joe Padfield 28:13 is a few ways of thinking about that issue and one is. I've been involved in discussions about manuscript identifies and the use cases that people come up with are things like someone sells a manuscript to another institution or an institution merges or closes. So, the domain name, at which it is located changes. I can't think of any reasons why shelfmark or a sort of reference code might change but that's obviously possible, Pia Eckhart 28:42 or the URL. If the URL changes then it's, yeah, yeah. and then, Joe Padfield 28:48 that is, you know, it's a known that item has moved and there's ways of forwarding things, and particularly if there was a central system of identifiers. There's also an issue with cultural institutions aren't always brilliant at retaining their own memory, finding information about past exhibitions, can be very difficult past events. So, and past partnership projects as well because it can be difficult to sustain them over time, they might guarantee you, they'll be around for 10 years and if you annotate on the ninth year, and go to look at it again on the 11th year, you may be out of luck. So, again, I think we've seen the pendulum swing in terms of supporting or letting national infrastructure with a bit I'd hope that that would be when understanding the vital role of these items in a kind of national collection would mean that they are discoverable and sustainable, regardless of whether they've moved or the institution has changed in the long term, but that again takes work and an investment. Pia Eckhart 29:54 May I ask something about that, maybe I haven't understood well what the national collection is trying to do is it's something similar to the Handscriften Portal in Germany, like for manuscripts only but yeah, because that would still mean that the the actual images and the objects still lie with the institutions and I think that came up earlier that it's very difficult to get all of these institutions to work in the same way, or to commit to the same rules. But that's just an outsider perspective. Anne McLaughlin 30:33 I can't speak exactly towards what form the national collection will take. I don't think we know that yet. Joe Padfield 30:42 It's still 'towards' it, Anne Anne McLaughlin 30:43 The Handscriften portal In Germany is incredible. It's a phenomenal initiative, and incredibly useful. I think we could aspire to being as useful. So obviously there's additional challenges when you're thinking about thing that are not just manuscripts, how do you do this for art galleries, museums and libraries, as well as sort of cultural heritage institutions like Abira was talking about. Joe Padfield 31:12 I just, just wanted to add one thing here. Someone's sort of mentioned it alluded to it in the chat about when we PIDs for annotations, persistent identifiers so the focus of one of the other TaNC Foundation projects. And it may well be that IIIF is shown at the moment, as a way of presenting content to the world, how do we share it to everybody, but at the moment if you're going to carry out academic research you do this based on published suitable literature as well as books. So the notion of how we move from images presented in IIIF simply as being the current presentation from an institution, to being research assets that are citable with persistant identifiers is an interesting process. So, Pia, you mentioned the notion of Invenio RDM, so at the moment. If you upload an image into zenodo, it has some IIIF functionality. We did a little sort of experiment in the IIIF project where you can create a manifest for an image that's been uploaded to zenodo, and it works fine but you want to get about the first 20 tiles because then it's rate limiter kicks in and the zooming viewer doesn't work anymore, but technically speaking, you could then, you then have a DOI for your image, you have an asset that can be reused. So the notion of saying that if a national collection could act as the referencer for IIIF of assets that could be extremely useful. So if the image moves the namespace changes, don't care if your annotations are all based on a non-semantic persistent identifier for that asset, then the image underneath can move around, but this is quite a big infrastructure question, But I think, as, as more institutions with wanting to look at their images as research assets that issue of saying that how we manage them moves away from a digital media focus to actually more of how we look after our objects our paintings our books, we look after our digital objects, the same way that institutions may need to asset and look after, digital made digital works of art. We need to be able to reference them we need to be able to rely on it and build on this assets I think that's a very valid cause for concern. Anne McLaughlin 33:32 In our last few minutes, we've got one final question that popped up in the chat asking with regards to Pia's project, but I think it's something that we might have more comments on here, was there a tension, as to the choice between TEI and IIIF for annotating materials that you used. I know in my previous role at the Parker Library we used old TEI and scripted it into MODS for a See Also field, which is permitted within IIIF. Pia, how did you address that in your case. Pia Eckhart 34:09 I didn't, no. I thought about it, and right now I think I will do without TEI. Because of the like the, keep it simple approach, so if I, if I choose the annotation in a way that gets all the visual information that I need for one phenomenon, then I can live with not really directing the text on different parts of this annotation. Did I make myself clear. I don't know, But if I come to the conclusion that I need this direct connection then I think I have to go with, TEI, and I think I would, that maybe that is a naive notion I don't know, can also help with, like finding good tag names, like you could use that vocabulary for your tags. But that's just an idea. Joe Padfield 35:18 Okay, we are drawing close to six o'clock. Are there any other questions that have prompted by the discussion, or anything else that the panel would like to raise. We do have a few more minutes, Anne McLaughlin 35:38 Ben I was interested by your most recent comment in the chats. You mentioned that IIIF and the web large are essentially decentralised in nature. So permanence of inherent challenge, I think that that's certainly true. I know I've had a few IIIF projects that have been broken. Would you like to comment on perhaps, how do you see that challenge being overcome or if that's even possible or desirable. Ben Bakelaar 36:05 Yeah, it is something I've thought about a little bit in terms of kind of an exportable or downloadable IIIF object, a local copy. But I think the main, the main thing I can bring up it's useful as the Internet Archive, archive.org And I don't know how widely it's used outside of the US but it is, you know, it's the closest thing to some sort of permanent record right for for websites. And so if we think about exporting that model to IIIF I think right now. Yeah, IIIF is not going to be archivable in that way. Like a work file or ARK file may not fully capture everything about it because there's functionality and JavaScript involved. So, hopefully somewhere out there, through open source innovation there becomes a way to store a IIIF object, and then hopefully somebody has the willingness to provide the resources to have, you know, centralised storage, maybe it's Wikipedia or Wikidata. But right now the challenge is still technical and that you can't really encapsulate a IIIF object without all of its hyper textual references. So, archive.org is my, my reference model for them. Andrew Prescott 37:38 Is that something Mia I wonder whether legal deposit teams in Britain are addressing or interested in Mia Ridge 37:46 which paticular aspects sorry Andrew Prescott 37:49 The long term preservation of IIIF objects on the web. Is that the sort of thing that legal deposit would cover. Joe Padfield 37:58 If it's British web content, then, yes, I mean the digital preservation and web archiving community tend to share sort of working stuff together so hopefully they'll be picking up on the results of things but one way to test it would be to submit a IIIF enabled website to the web archive and see how it works out. I do think they use it for some of their training IIIF training is related to assets. The issue is is that I think those sets of archives the speed and responsiveness of I need the images, shared with drop drastically. That might still be okay for checking your annotations and stuff. Ben Bakelaar 38:48 You can create a new digital object or a copy of a digital object and then you can get the IIIF manifest out of it. But you couldn't ingest an existing IIIF Anne McLaughlin 39:00 Well and then you run into some issues around copyright as well but I don't I prefer us to attempt to tackle or even broach the subject of today, that needs some more legal minds around it. Joe Padfield 39:14 I think broadly there is a point where if you're referencing, you know, we know that people will use the online version of an item, and then write the citation, as if they've gone into the library because it looks fancy. But I think that, particularly as you're using collection items that aren't sort of as well known and might need to be around in some form, there should be a way of sort of saving it to the Zotero or something, or saving it to Web Archivejust so you've got that backup, and evidence for your own scholarship. Because if I didn't if you're doing a PhD and your examiner says asks you about an item and it's no longer available online, there should be some way that you can say well this is what I saw at the time that I looked, you know luckily I work in a collection that doesn't go into the 20th century really. But increasingly, a lot of my research does work with web archives. And I've relied on that archive countless times and if people can't save items in some way it does make it harder for them to be embedded in that sort of ecosystem and scholarship. But again, copyright lots of other issues,