 Yes, so sorry about confusion. If you looked at the schedule about two days ago, then you'd see that Archon was supposed to go before me, but Archon and I were talking and we realized we were really talking about the same project. So better that I start with the introduction and then hand over to Archon afterwards. So hi everyone, my name is Jonas Erberg. I'm a Shuttleworth Foundation Fellow and I'm here to talk to you about contextualizing creative works. Something quite different from the context that Peter was talking about about 40 minutes ago. And I will allow you, a few seconds, to bask in the glory of this comic before I ask you a question on it. Now, the question on this one. Okay, you're still basking. With a show of hands please, who can tell which comic series this is from? Yeah, that pretty much sums it up. XKCD, so again, show of hands, who can name the author of this series? Okay, not as many, but still, we're getting there. All right, so XKCD by Randall Monroe. He has a peculiar style of drawing, he has a particular sense of humor, so it's quite easy to detect here. Now, let me show you another one. This one, this image here. Can someone tell me where this is from? Show of hands. One, two, not more. So for those of you who doesn't know, okay, you can take down your hands. Would you be surprised if I told you that this image here is in fact XKCD comic number seven? It is also by Randall Monroe. Now, I would like you to take a moment to reflect on how you felt about this image before I told you that and after. And I'm quite sure that many of you would agree that your relationship with this image has changed, knowing that this is part of the XKCD series, that this is XKCD comic number seven. It's not from anyone's sketchbook, it's from the sketchbook of Randall Monroe. This changes the meaning of the image. It changes the value of the image, having this information. I can almost guarantee you that I would find, well, maybe I'll find a few people, but very, very few who would buy a print of this image, not knowing that it was made by Randall Monroe and XKCD comic number seven. I would find a significant amount of people who would buy a print of this if they knew that this was XKCD comic number seven, because that's a print that you would buy. You'd not buy a print from someone's sketchbook of a girl laying down. But you would buy a print of XKCD comic number seven. What I'm hinting at here is that understanding the context of a digital work, like this image, it changes our perception of it. It gives it meaning and it gives it some value. Now, this context that we're talking about here, it can be who took this image or who created it. It can be the license of this, in this case it's Randall Monroe, he's XKCD, so it's a Creative Commons license on this image. It can be that this is part of the XKCD comic series. The context can also be where was Randall when he was drawing this? Where is the sketchbook where this is posted? Is it in Randall Monroe's drawer in his basement, or is it on exhibit at the Museum of Modern Art? It should be. This all relates to the context of a creative work. It helps us understand where it fits into the world around it. And if we had this information, imagine if you could go anywhere online, if you could find any lolcat and you would be able to hover over it and have this information expand from it. The context, the complete context of any creative work online to see who made it, where, why, how, and importantly as well, how it relates to any other work. Is this a remix or is it an original work? Has anyone applied any of the gimmick filters to it or not? Now, having that information readily available to you would mean that you could for instance attribute things accurately and ideally automatically. So it would have a practical relevance. We could take our tools and we can tell them, well, hello, computers, you know this, you know who took this image. So just put that pesky attribution there for us and don't bother us so much with it. But it can have, of course, as I hinted at, can have a monetary relevance. If you have the context, if you can find the author, then you might be inclined to tip them a little bit or post something under forum say, hey, great work, I love that image. Or buy a commissioned work. Now, the context can also help to understand where in a tree, okay, that's not a tree, that's a circuit board, I'm sorry about that. It can help you understand where in a tree a remix is, this actually fits. What came before it, what happened afterwards? Were there any remixes of it? Were there any alternatives of it? Oh, sorry. Let's go back there. This all starts with having an accurate attribution. Having the creator and the source named and available when you're looking at images, when you're browsing for works to remix, for instance. But unfortunately, as I hinted on, ideally, maybe we should have our computers do that for us, have the attribution happen automatically. And so if we want to do that, then it's not enough to just have a correct attribution. We need, come roll, machine-readable information and attribution. We need something that our computers can act on and understand. We've all seen image by Wikipedia or photo from internet as a means of attributing an image. So how could we possibly go beyond that if the problem is that people just don't know how to attribute? How could we even get them to not only attribute correctly but also have machine-readable information available? About a year and a half ago, I founded Commas Machinery. The aim of Commas Machinery is to make attribution easier by automating the process of attribution. And the realization here that I came up with or the thought that I had at one point was that, well, if I have one image that I need to attribute correctly and I look at all the how-to's of how to do that and the best practice guides and so on and so forth, I might be able to do that quite fairly easily for one image. If I have 10 images, okay, I can manage that myself. It's manual process, but I can do that. If I have 100 images or even 20, and I start mixing and matching, I remove one image, I insert another one as I was creating this presentation, for instance, it gets really tricky to keep track of this information. And I'm quite sure that there's no one here that will be able to recognize or to remember the context of 20 different works, the license where they came from and who created them. Not in your head, at least. But we do have something right in front of us. We have computers, these wonderful machines which are aimed and tasked with keeping track of large amounts of information. So why not get them to actually do the bidding for us? Why not get them to help us remember the context of images? Now, my first thought when I approached this was that, well, hey, you know, why don't we just embed this information within the images themselves? So make it a part of the image file as it gets passed around. And we've got all these wonderful standards. You'll recognize a few of them. This was an image done by Peter, my colleague down here, when he first started working for us at Como's Machinery, which gives an overview of some of the standards involved in expressing the context of a work. Now, Exit and XMP are perhaps the well-known here as well as maybe the IPTC IAM. So why don't we just use these? Why don't we just take these, you embed the attribution information, we embed the license, and Creative Commons even have some guidelines for how to do that. So why don't you just put everything in the file and then be done with it? Well, it turns out that reality is slightly more complex than that. This is an image or a snapshot from something called the Embedded Metadata Manifesto. And this shows you essentially a green button where a particular service retains the metadata when it publishes an image that you have uploaded. Red one is very, very bad. Metadata gets lost. So if you look at Facebook, for instance, it's just red. Whenever you embed metadata within your file, the moment you upload this to Facebook, lost, gone forever, never to be seen again. If you upload it to Google, well, Google is very good at keeping track of information. So they will actually retain this for you and for any other purpose that you might invest in. But this shows how sketchy the support for retention of metadata is. So even if we made the best attempts at embedding this information within images, the moment you start sharing this around, you start spreading this image on social media, you have no clue how long this would be retained and most likely not very long at all. In Commos Machinery, we're right now working on a project that we're actually not quite ready to talk about yet. And this is just a prototype logo type which we're still tweaking. But we have a name for it. It's a wonderful name called Elogio. Again, which Peter came up with at some point. Elogio will be your distributed attribution machine. It is your extended memory. This is a service that you can run on your own machine or rather will be a service that you can run on this all vapor-aware by the way. That you can run on your local machine, have it in the cloud or subscribe to service that interacts with the tools that you're using and helps you remember what you do. So when you take an image and you embed it in your presentation or you take an image and you put it into GIMP and you start remixing this, then it will remember what image did you take, where did you take it from and created that one, trying to capture as much of the context of that image as possible. So that after you've done all these hundred operations to the image, it will still remember that this was based on an image that you got from somewhere else and it will keep this information with the work. Ideally, as it gets spread around, and we have various techniques to keep things together, but essentially it's a database. It's a database task with keeping a record of the works that you interact with and the works that you use and the works that you create. And if we had this database or rather when we have it, then it lends itself quite naturally. So if this database keeps a record of what works you're using when you're creating a presentation, for instance, like this one, then you could use that database to get automatic attribution. You can ask it, what images have I used? And you'll get the list of that and then you base your attribution on that. If you remove an image from your collective work, then it will remember that, well, you used this image but then you removed it, so that's not relevant anymore. So it will remove it from the attribution. Now, we can also imagine reuse notification. So if someone uses a work that you have published before, then you could get a small notification saying, hey, your work has actually been useful. Someone here has remixed this. Could be facilitated by Elogio as well. Or if someone embeds your image in a presentation and you happen to be a flatter user, then potentially or conceivably, maybe the user who embeds your image can get a small hint, you know, why don't you flatter this person whose image you were just using? And this is just starting to hint at what could be made possible if we had access to the context of creative works and if we had this technology already in place. Now, Elogio, as we are designing it, is an open framework that we would very much like for others to extend further. Because Elogio itself, it's really just a database. It has some tweaks and peculiarities that make it suitable for working with contextual information for creative works. But in order to be really useful, we need tool integration. We need extensions for the tools that we're using to allow it to retain information as your importing works or pasting works into your application. Or when you're exporting information. In all those processes, we need to retain metadata and that requires some extension work. But we must also retain information within the application. So as we start mixing different sources within one application, we need to keep track of what is actually coming from where and how do we combine these works together to form a correct attribution in the end. Now, my colleague Artyom, he will now soon take over and he will talk about metadata in remix works. And if you're further interested in what we're doing or in general about metadata for creative works, then I would very much welcome you on Friday. We're holding a buff session at 10 past 12 with Peter and Artyom, where we can really dive into this topic more in detail. And I do hope that you will join us because I hope that together we can actually eradicate the image by Wikipedia attribution forever and forever in the future. And bring work, bring meaning and value to digital creations. So thank you.