Anne McLaughlin 0:03 We've got a question typed in from Barbara Williams: When you do crowdsourcing, is there a verification system. Suppose Jason, that might be a kickoff for you. Jason Evans 0:16 Yeah, I could start with that one. With the work that we're doing with using Wikidata to crowdsource through the lockdown. There is, there is no verification system. It's, it's more out of kind of necessity with a lockdown that we've used it and it's not ideal. It's what it was kind of set up as a way of kind of experimenting and showing what was possible, ahead of launching our own crowdsourcing platform that does this. With our own crowdsourcing platform, there will be vericut verification from what I understand, multiple people can tag the same thing, and then you can look at whether everyone tags the same thing or whether there are discrepancies in what's been tagged. And obviously if any of this data was to be added to our own authoritative data, there would need to be additional checks, we did that. Joe Padfield 1:19 Okay, thank you. We've had another question in the chat from Vera. Have you used computer vision, also to automatically detect details in your images. So with any of the panellists like to respond to that question. Or I can respond. Various people are doing this Vera. There's the Visual Geometries Group in Oxford have been doing, trying to type some of their tools directly to IIIF manifests. Yes, that's the one. We have been working with them here at the National Gallery to try a few issues, and looking to do some more work with them in future projects. But at least there are a number of different research groups looking at computer vision, and how we can do direct ingest through IIIF manifests. So I think I suspect so we talked I think there'll be a lot of development in that area as more trip less manifests become available. And we're guessing. Connect all of that up. So, yes, but a lot more stuff to do. Are there any other people want to respond to that, or any other questions, Anne McLaughlin 2:28 You've had a question come in via Slack, actually, for basically all of our panellists today. They note the difference between a Abira's approach so using GitHub and going very organically, to the ways in which the Royal Collection, and the National Library of Wales have been able to mobilise more, I would guess more institutional support, might there be a way for larger institutions to collaborate more effectively with individual independent researchers or individual projects, Roger Hyam 3:12 think so I think the key thing is open source, I think the more that we can adopt open source technologies and solutions, and use larger institutions kind of thing, enhance budgets and abilities to develop those solutions and make those tools more readily available. I think there's much greater channels for, if not even formal collaboration, but kind of everyone benefiting from from that expenditure. But I think one of the things that IIIF particularly is making those do more, is look, is look outward, whereas previously we've looked. We've always looked inward for solutions this is, this is teaching us to look outward, we still got a very long way to go to learn how to do that but I think that's something that's inherent in in utilising IIIF. Abira Hussein 4:09 I would say that it would be. Yeah, I think it would be great to be able to work in partnership with institutions, where there's a possibility to either work within institutions that have connections that relate to that particular community, so I've been helping them to think about, as they're developing the technology for representing IIIF in thinking about in terms of contexts in how to improve multiple stories. So thinking both the content and actually working on how to work on the representation aspect, as well as being able to support either financially in thinking about how communities can either exist outside of institutions, I think. Yeah, I guess it's developing that trust where either they can work where the institution, supports them to actually providing the technology to host their collections or support them outside of the institution. Joe Padfield 5:05 Thank you Abira, I think there was one more question, Anne, were you going to answer that one live in the q&a. Anne McLaughlin 5:13 I was going to say, we could just ask the question I was asking all of our panellists respond. And so, Tom has mentioned the very vital issue of agreement on schemers across organisations and Jason's presented some of the benefits of using Wikidata alongside IIIF. What are the speaker's thoughts on integration between or with IIIF and the semantic web, for example linked up and data in general and endorsements for applications like Wikidata are their ambitions within the National Collection project to move towards linked up and data in ways that benefit IIIF integration as well. Tom as, you're the first one mentioned I would like to kick off with your thoughts. Tom Crane 5:56 Yeah, well, IIIF is itself, linked open data it's JSON-LD, but it tries to hide that from users as much as possible as It's unnecessarily complex. So I would say, like, add, add one more bullet point to that list, they're linked open data is nice but usability is more important. So, all other things being equal, usability Trumps Semantic Web concerns, but I didn't know, I mean I would assume that if there were some simple scheme, schema.org like thing, that IIIF manifests could link to so that aggregators such as a potential national collection could aggregate them, then all other things being equal, that should be open data and rather than just some, something else. But yeah, as long as it's usable and and friendly and simple for people to make an understand. Joe Padfield 6:55 I would say just to talk about the wider programme Towards a National Collection, the IIIF project is the foundational projects is one of the smaller projects. There's a call open at the moment for much larger projects, which you should hear about in the next couple of months. And some of those might be exploring more semantic information, I mean the IIIF project was relatively small in comparison. We are within the National Gallery have involved in another project called SHOCC which is the social science humanities open cloud project, and we are trying to make some more FAIR datasets, sort of semantic datasets available talking about heritage collections within there we have IIIF resources, so we're looking at the modelling of how you'd use a heritage ontology, such as the CIDOC-CRM to include references to IIIF manifests and linking them to what they're showing the objects they're showing and sort of how collections are manifest, a group of works by a given artist and these sorts of things as well, because you can potentially look at the notion of dynamically generating manifests based on queries to link to open data. So, if you've got a nice set of heritage linked data in some semantic structure, how IIIF is used in relation to that can be the result of a search output. In addition to anything which is a pre-packaged collection or story being told, so I think work is happening in this realm, but again with, as mentioned in the previous topic, as more and more institutions have robust IIIF of solutions, it will become easier to then base projects and experiments on them. So at the moment experimentation is happening at the same time, so.