20 May 2024

Fabien Benetou: Cool in little light. And now I don’t talk about this.

Frode Hegland: What don’t you talk about?

Fabien Benetou: I’d say just family matters. That I don’t feel comfortable sharing publicly.

Frode Hegland: No, absolutely. Sorry. I just heard the last words and got a bit you know, what did I miss?

Fabien Benetou: I would, you know. Well then. So for something.

Leon Van Kammen: Waiting for you.

Fabien Benetou: About it. It was. It was nothing. Yeah. Nothing critical about you. But I’m doing the promotion of the Lego box that I think I told you about, actually, before. And how for for to learn programing. I think it’s amazing. And it’s especially for people, including kids, but not only who think programing is not for them because right away you do something tangible. You can start to dream about robotics, but you do it from a way that is. That is not, I don’t want to say scary, like you don’t have like some weird syntax error or anything of the kind. So I think I really recommend it for literally everyone.

Leon Van Kammen: Yeah. Cool.

Frode Hegland: Just checking on people here. Sorry I’m late. I can only blame myself, which is horrible. I wish I could blame it on transportation, but it was actually quite okay.

Dene Grigar: You can do what they’re doing here in the States. You can blame your wife.

Frode Hegland: Yeah. No, I.

Leon Van Kammen: Politicians love to.

Dene Grigar: Blame their wives for upside down flags. The devil made me do it. It’s my wife.

Leon Van Kammen: Here’s a.

Frode Hegland: Funny.

Leon Van Kammen: Thing.

Frode Hegland: I saw this study on human behavior and psychopaths. Yeah, don’t ask how I came across it, but very strange thing is, women need to be more attuned to who could be a psychopath more than men.

Dene Grigar: Yeah, of course we do. Yeah. Right.

Leon Van Kammen: Exactly.

Frode Hegland: So one.

Leon Van Kammen: Behavior.

Frode Hegland: One behavior that women have adopted. And this is not some theory of me saying studied. That’s why they’re late on the first date, generally, because if it’s a normal guy, you apologize, they’re a little annoyed. It’s fine if it’s a psychopath, there’s a 1 in 5 chance he will get really angry. And that’s how you tell it to psychopath, because they have to be in control. Isn’t that.

Leon Van Kammen: Weird? I can tell you.

Dene Grigar: Having being a woman and the only one in the room that I know of so far, I was never late for that reason. But I will say that it was. I always would bring a book with me, and if they pulled the book out of my hand and angry, I knew I wouldn’t date them again. It was a book book test.

Frode Hegland: That’s a really good one. Talking of only women in the room, I saw my very first student on Friday, and she’s going to try to join us today.

Leon Van Kammen: Good.

Frode Hegland: She was traveling back to Paris, of all places. She’s originally from Lebanon, and when I was doing a little bit of teaching in London before I became a proper teacher she was in the class and she responded with an allegory related to Plato’s Cave. So I was like, if this is students nowadays, okay, I can do this. So anyway, that’s all good. So Fabian, you said that We got a message about our potential guest presenter today not being here. But let me say, what’s your message? What are you saying? Okay. No. That’s good. 1st of June. Yeah, we could probably do that. Okay. So let’s this has been discussed in confusing media. A bit of slack, a bit of Twitter. Fabian, could you post his profile in here and do a 15 second bio to see what the group thinks? Because you certainly had a potential. Yes, last week, which was all very interesting. And then it got a bit confusing.

Fabien Benetou: Right? So, so just a little bit of a clarification. Then I invited and actually, I think I used the wrong name last time, so I really don’t want to mess it up. This time Oleg Frolov to, to give us a presentation and he happily said yes. But I misunderstood, I guess with or I’m sure well, at least one of us misunderstood. And so I thought he would present to us today, but I think he had in mind next week only to realize he couldn’t next week. And I suggested today, but I guess that was a little bit short notice. So he would. He’ll probably present to us in two weeks, I guess. But he might also he’s based in London, so he might make it then. And for another bigger picture, especially since I see people with AI assistants who joined the chat on their behalf, I guess, to to get a view of our discussion even without being here. So Oleg works on interaction design. So what does the button look like? And that sounds a little bit trivial, but I guess everything that’s. I mean, when you start to dig a little bit in depth, what it actually means, it’s a little bit more trickier. What is like feedback? Do you have so thanks for put the link in the chat. But and what’s a button in in XR especially that.

Fabien Benetou: Now if you take a noble button like I’m trying to try to, it’s too far away from the camera. But let’s say if I take this button from the console, it’s actually a pretty emblematic set of buttons. You have this kind of buttons, the cross, and then those ones. But you also have vibrations. You also have those on top. And those are different kind of let’s say this is on and off. This might have some degrees of freedom to the sides. Some of them might be at the same time. But here in VR you can have a screen inside the button. You can have the button changing its own shape, like when you push on it. Maybe it’s a cube. Now it’s become a sphere. So it’s it’s it wasn’t trivial before. It’s even less trivial now. And he explored a lot of those, like what happens when you tilt the button and button, of course, is like a simplification. You have a lot of different shapes and behavior. And he did a lot of very interesting exploration which inspired me to explore also a bit last week and what I showed you with when you pinch in the air and you have a set of buttons and then you rearrange them, that was inspired by some of his work, but

Frode Hegland: Fabian, he’s just updated his main page on the second link. It looks different from a few days ago. It’s very cool. Yeah, we’re at that level of I don’t know if we should call it details or polish or implementation, but it’s really nice to look at you know, I’ve been brushing it off until now, almost saying, you know, pop up menu and then being very glad you’re all in different parts of the planet. But yeah, we need to look at this in specifics now. There is one thing that I really want us to do today. Well, there’s a few things we have to do. One of them is a little bit of a look at the papers that we still have some time on to make sure we’re all aligned on that, and then I really want to spend some time on what should be on the map. And I know that that’s an oversimplified question. There is two sides to this. What should be in the map in terms of what we have in terms of data metadata. The other side is why in the world would we want stuff there? As in, what’s the point? Right. These two things have to be together. So first of all especially for you, Danny, because of the discussions we had on slack about visual meta, I want to spend 30s to define everything of what visual meta is, because part of the problem with visual meta, and the reason it took me so long to get my PhD, is that visual meta is almost nothing.

Frode Hegland: And I’m not being self-effacing. That is the reality of it. It’s polished down to being quite literally at the end of a document instead of hiding. And I think, you know, all of this. I’m just saying there’s nothing more, really, at the end of a document and an appendix rather than a data resource fork. Say what the document is, say what it connects to, references and links and stuff. And put other things in there in a disposable list. That really is it. We’re using the BibTeX style as inspiration currently. And and the other important aspect of visual media is it has this introduction where it writes in human terms what it is. In case someone sees this and wonder what it is, that’s it. And the point of it is that a system like my own reader that can parse visual data can easily get all the metadata without errors. But because of the LMS and stuff, now they can use the visual meta to extract the metadata really cleanly because it says what it is and it reads it. That’s all visual meta is. There is, other than the basic notion of BibTeX style, you can put anything in there. So what it means, in effect, is that if we want to have rich hypertextual interactions, we’ve got to put things in this metadata to say that these are the elements of the document, that the reading software can then make cool and interactive. Any questions is that.

Dene Grigar: That makes sense. And we’ve been doing visual meta from that perspective in the next, because every exhibition space for the 2571 works has its metadata right on the space for each work, going as far as the sensory modalities and sensory sensitivity issues. So.

Leon Van Kammen: Yes.

Frode Hegland: There is a related issue that Peter is working working on for the paper. And that is I’m for the PhD. I’m entirely document centric. I live in a PDF, so it is in the PDF. If the metadata is on the same web page, so to speak, that is also good, useful and worthwhile. What Peter is working on is explicitly defining stuff that is extra appended. And what to me is so exciting is that Fabian is taking this another step. You know, I’m all about the metadata being on the same level as the content. He is all about the code being on the same level as the content. So in order to make a really not just richly amazing environment here, but to allow other people to do it, it’s really interesting synergy that’s happening.

Dene Grigar: Did you see I posted earlier? Mark reposted it that Google declares that the end of the World Wide Web is here. Very interesting article.

Fabien Benetou: And I think it’s because they are fucking it up to be very I think it’s their fucking it up for money. I just listened to another podcast on I yesterday, I think or, or this morning even. And the argument is basically they have control of the platform being the web itself or Android, which a lot of people use it to consume the web from, or Chrome, etc. and they basically hijack this like, oh, you don’t need to access, let’s say, Denny’s website. We’re just going to give up the answer straight from it. Which for your website, I guess it doesn’t matter much because as an academic, you have another way to publish your work. But for a lot of people like journalists, if people don’t go to the website, there is no monetization, there is no subscription. So basically they’re going to starve to death. And if the large language models like hijack the normal flow of the web and the web is hidden away from the user, then yeah, the user web is not going to exist anymore. But it’s not it’s not a mistake. Arguably, it’s like a business model to say, oh, we don’t ever need this anymore. We need to provide this. So. So I think I’m. Yeah, but I don’t think it’s a good thing I agree.

Frode Hegland: I see Mark commenting suspect much for visual media going forward. It being data will be JSON or Json-ld as you know Mark of course I completely agree. I don’t care what’s in the visual meta. Fabian, I also saw a tweet related to this end of the web thing pointing out that over so many periods, you know, so many web sites are missing. There was something today, I think I retweeted it. Yeah. The the web is brittle, but it’s the best we have.

Dene Grigar: But can I mention why? I mean, I didn’t respond to you because it would take for so long to write it, but there’s a lot of reasons why those sites are gone. Many of them were made with drag and drop technologies like Iweb, muse, Dreamweaver, you know, crack code that becomes errant. A lot of it had Java applets Iframe’s that were ill prepared. Many of them were made for a specific browser like Netscape that didn’t exist. It doesn’t exist anymore. I mean, this is what we deal with in the lab every day, is Andrew can shake his head and agree with this is it’s not the web it’s causing. It’s not the websites are gone the way the web sites were prepared, made it go gone, you know. So it’s interesting. But the good news is that the Wayback Machine is our best friend and we can find things. So we live in the Wayback Machine in the in the lab.

Leon Van Kammen: Yeah, we’re dealing with.

Dene Grigar: The piece right now. We’re saving a piece right now called Red Moana. Beautiful piece created in late 90s. Gorgeous. And it was not using flash or any proprietary software. It used arcane coding languages, you know, and you know, just this really bad code. And so now we’re rebuilding it in HTML. Css and JavaScript can be beautiful when we’re done. But sincerely, it wasn’t a flash piece. So.

Frode Hegland: Open standards. Important. Clean code is important. Absolutely. Yeah. Leon, please.

Leon Van Kammen: Please, please.

Frode Hegland: I also have something to add. Can you hear me?

Leon Van Kammen: Unfortunately, yes.

Frode Hegland: Sorry.

Leon Van Kammen: So yeah, I was also actually.

Frode Hegland: Thinking about.

Leon Van Kammen: This that these kind of articles, I see them more and more and they always leave out things like the Wayback Machine or archive.org or like IPFs and the same in the same way also. For example, Elon Musk is, is never is always highlighting that. Yeah, I should I should sort of save the people’s information this market town hall. But yeah. Fabian was also mentioning that this domain was now changed to X.com. And he was actually wondering, like, is this going to be a permanent redirect for the end of times until the end of times? Or are all these links going to be broken? And yeah, that’s a good question. And then so I was thinking like, why is he not talking about this Elon Musk, about the Wayback Machine or about IPFs, where you sort of have like forever, it will live forever, which in a way can also be an incredible problem because, for example, the EU started this whole thing about the right to be forgotten. So in a sense, a broken link is also a feature. Some sometimes maybe you want to break it as an author. And last but but least I’m, I’m also surprised. Why isn’t there a archive.org dump being shot into space, you know, as a sort of you know, backup for because, like, if Elon Musk would be really concerned about saving the, the, the human knowledge and all that, then he should shoot an archive.org dump into space ASAP, Leon I agree.

Dene Grigar: Hey, it’s called lox. Lots of copies kept safe.

Frode Hegland: So. Yeah. On this. I know this is no news to any of you, but, you know, look at that. Look at the numbers. 256gb and something that is smaller than my fingernail. Right. That is our reality right now, and I’m perfectly happy for the web and the internet to do transportations. But I don’t understand how in the history of humanity we’ve ever wanted to say, oh, you want the knowledge? Don’t keep the knowledge yourself. It should be over there. Knowledge should not be on servers. It should be distributed as much as possible. You know, we should all have all the data we need to do all our work. It’s just, you know, and then we have synchronization in the background to just trust a few places for it. I think it’s absolutely insanity. Danny.

Dene Grigar: Yeah, I was going to say one other thing. We’re doing this. This answer is responds to what you just said there was. There’s an art form called net prov. So it’s improvizational performance on social media sites, particularly Twitter. And the king of net prov is a guy named Rob Wittig. And so he’s produced 36, 37, 37 net problems over his career. An incredible works, right. But they all are taking place on mostly Twitter and then Facebook. Right. Little LinkedIn. Not much. Now we’re trying to build a net profit engine in the lab that’s going to allow us to document the net profits because we’re concerned. That Elon Musk is going to, like, wipe out the net profits in the future. You know, and this very, very hard to get the API working. And how much are we going to have to pay for API in the future? So we’re trying to get our hands on all to get a major dump of all of his net profits taking place on Twitter, so that we can recreate it in this engine for future people. So it’s the first time anyone’s tried a net profit engine. So it’s well, I’ll let you know how it’s going. But the bottom line is he is the enemy. Elon Musk and his Twitter is an enemy of all art form. I mean, he’s killed the performance art in Twitter, which is awful. Just awful.

Frode Hegland: Yeah. So.

Fabien Benetou: I won’t even touch on Musk because it just triggers me. It just. I think he’s a big idiot. What? I just wanted to share a rich one and successful one, but it’s totally different. What I just wanted to share, though, in the chat is my my offline prep folder, so I’m not exactly a prepper. I don’t expect the world to crash, but as fraud shown us with the 250 gig micro SD, it’s like data is not free, but it’s. If you live in the West, it’s basically newly free. So. But sometimes, yes, you can expect the internet not to work or whatever. Not to work. And sometimes you just don’t want to be online. And I think it’s a good habit to at least test yourself in term of resilience and in mental also resilience. And going offline, I think, is good and doesn’t mean being cut off from actual knowledge. And the the little directory example that I shared there in the chat has Wikipedia has StackOverflow has a couple of other sources like OpenStreetMap also, so it doesn’t have the web’s entire knowledge. Obviously not, but it has some of the places that I know are of quality or helpful to my thinking, basically process. And that honestly, I pretty much never use that. But to know I have it there, I feel safer, I guess I feel better and it really doesn’t cost me pretty much anything. It’s like a small percentage of more hard disk. Took me a couple of hours to download it. And they were stored in a format that if I double click it or if I use a little Raspberry Pi when I’m traveling, I can share it with anybody on my Wi-Fi. So I think there are still so ways to get that knowledge accessible through some terms with a nice interface that somebody who is not technical can access. So I think it’s, it’s it’s a good habit from time to time to, to try those things, and I recommend it.

Frode Hegland: Yeah, absolutely. And you.

Fabien Benetou: Need. Sorry. Just to get you there is the Gutenberg project in it. So it’s like not just Wikipedia or StackOverflow for technical knowledge. It’s like, even if you were to I mean, yeah, it’s it’s in itself, I think quite precious.

Frode Hegland: Mark, just really briefly because I say Rob put in the comments there, which is Pacific Gas and Electric cut off his powers. He’s having trouble getting connected. It’s kind of a in this technological age, living in the middle of Silicon Valley, you can still have problems even basically getting connected. So thank you, Rob, for that very relevant cut off. I hope you will see the transcript. Anyway, Mark.

Mark Anderson: I. Yeah, I was just yeah, I was just going to check. And this isn’t in any way counter to the thing of doing federated things. But I one of the things sort of we’re mulling on in the paper I’m doing with a mentor, guys for that demo is, is, you know, the sort of legitimate issues about the fact that how do you sort of deal with content control and redaction? And this isn’t about this isn’t just pure money making. It is about the fact that if there’s some, you know, once something escapes into into broadly federated system, it can it can be potentially very difficult to stop its further progress. And this, this goes beyond notions of good or bad, but just as a sort of an interesting systemic thing to deal with, because it is the flip side of being able to have everything all the time, wherever we want it, because we want it. And so, I mean, I don’t have an answer on that, but I just sort of I just popped that in the, in the conversation or something for us to sort of bear in mind as we, as we look at this. Thanks, Peter.

Peter Wasilko: Okay. I just wanted to report a few experimental results. I tried using json.parse, the HTML versions of the papers that Andrew has been using for his demos, and that was unsuccessful. Apparently, the JSON parser module that they’ve provided in the NPM system has some glitches in it, and as a result, it’s sort of attempting to extract the text from the files, but it’s leaving sprinkled bits of HTML tags in the middle of the text, making it unusable. Then I started trying a few different HTML5 parser. Libraries and they were having problems with the ACM generated HTML. So at this point, I’m going to write my own parsing expression grammar to actually pull the content out of the files. Then I tried a system called Dot, dropping the link in the sidebar which is a local minstrel 7 billion parameter LM that can read files off of your hard drive without uploading them to the internet and allow you to talk to them. So I tried that with the papers from Denise blue Sky selections. The result there was interesting. The LM, first of all, seems to have no clue to how human names work because I asked it what’s Martial’s first name? And it thought that Cathy Marshall was a man, so it totally misgendered her and assumed that when it was given a last name, that that was actually someone’s first name.

Peter Wasilko: Then it managed to hallucinate some non-existent papers from Marc Bernstein. I don’t understand what it is with LM, but every time I get a hallucinated citation, it’s always attributed to Marc Bernstein with really ridiculous paper names. I don’t know what it is. There’s this, like, mental block in all the I’s connected to his work, and that was utterly bizarre. But that’s sort of where we’re standing there now. I think if we’re going to have LM generated results in visual meta, it would behoove us to include the prompt, the LM model that was used, the model’s parameters at the time to provide some element of reproducibility when working with it. So someone might want to vary the prompt to see if they get different results. So I think that information is needed. And if we’re using just like conventional extraction things like Porter stemming algorithm or something, it’d be sufficient to just indicate what the conventional algorithm was that was used to generate the results without going into details there. So that’s where. My work is at the moment. And of course, I’m working on the paper with fraud to address the external visual meta.

Frode Hegland: Yeah. I suggested when I read the paper that he just open it here that he submitted with me as a second author. So the paper is detachable visual meta, where he talks about monotonic visual meta and non-monotonic meaning deletable and not deletable, which I think is a very useful addition. So once he’s finished his his draft, then if you could please have a look at it, that would be really, really great. See, what do you think about his suitability and everything?

Dene Grigar: Why do I need I’m sorry? Why do I need to look at it?

Frode Hegland: Because you’re the chief academic, brilliant person who might notice things. We might not.

Dene Grigar: Okay, thanks.

Frode Hegland: It’s also good that we kind of, since we’re talking about the same source, that it kind of connects a little bit. This is for human. Yeah. Giving Klaus a good selection of things for him to choose to choose between. Right. So Oh, yes. Leon.

Leon Van Kammen: Yeah. If before we go into the serious mode, I have I have a small update which might be interesting for the Sloan project or even other projects because it, it basically addresses the problem of data storage, which is you know, basically the problem that all the researchers, academic professors, they all have their own way of storing their data. Maybe it’s a university drive, maybe it’s a cloud drive, maybe it’s a Dropbox. So before you know it, there is a huge discussion on, you know, where should we store all this, this, this library of documents. And I will quickly share my screen to show you that there is actually a pretty simple solution for it. And it’s called I don’t know if you can see my screen. Yes.

Frode Hegland: Can do. Okay.

Leon Van Kammen: So it’s basically called Rclone. It’s an open source app. And what it it’s just a command, actually. It works on Mac, Linux, everything. And I’m just going to quickly show you what it can do, like you have like, all these cloud providers like Dropbox you have protocols like WebDAV. Fabian was also working with it. You have gcloud, you have iCloud, all these things. And with Rclone, you can basically create a proxy drive, which is basically a bundle of that of these drives. So in the command line on a server or on your laptop, you can basically serve for example, a web dev drive where, where your webXR app, for example, can connect to but under the hood this will basically expose a drive let’s say a DNI, which is a Dropbox and maybe fraud, which is I don’t know Amazon S3 or or Mark, he has some kind of SftP drive on a university server somewhere. So you can basically create a proxy drive, almost a meta drive, so to speak, where all the PDFs are basically scannable as a whole. So this is a pretty nice, unifying free technology, which already existed for a while. But yeah, it’s very easy to not know about it. So that’s my small, small update.

Frode Hegland: Thank you for that.

Dene Grigar: Thank you for that.

Fabien Benetou: So I put a link to a demo I’ve done using a clone couple of years ago. I recommend it, but to be a little bit upfront on on this. I think for our audience, if we imagine an academic who quote unquote, just want to publish a paper and they don’t know what the command line is, I think it’s a little bit too many steps.

Leon Van Kammen: Yeah, I, I agree, I think it’s only worth it the moment you want a multi-party session where you could also basically see other people’s documents. And I think in such a context, there should be some one technical person and then then it will be usable. But yeah, it’s not it’s not a sort of It. Yeah, it wouldn’t even make sense for one person. It’s only makes sense if you want to do this on a, let’s say, topic scale or university scale or basically bigger scale, because, well, that’s a bit unsure also and unclear to me. Like what is the intended scale of the sort of is the scale a mono library locked, or is it sort of should it be multi library, you know, connections and links going from one PDF to somebody else’s library PDF. I don’t know what the big picture is in the end.

Frode Hegland: So that leads perfectly well into what we’re looking at next, because one of the discussions we’ve had for a while is when you’re in the library view using the term loosely. What do we mean by library? And one more in a proceedings. What do we mean by that? In a sense, as Andrews pointed out, I believe they’re all lists. So what is the scope for what you’re interacting with in human scale? We have to decide as users and designers what where we should end the list. Right. So the thing we’re going to be looking at for a little bit now, please, is the most prosaic thing of once we’re in XR and we’ve chosen to go into a map view, what should be there and why. Now here is my top level thing that I really need, particularly Udeni being the prime academic to make sure we’re all on the same page. Use case one. Is reading with the aim of writing something. So that’s the DNA use case number one that we’re focused on. Below that in brackets, so to speak. I would say as a crap academic who took forever to get my PhD, that one of the real use cases is going through new papers in a proceedings or not to understand what’s interesting, what’s worth it, but also so that in the future, when I’m writing a paper, I can, to be honest, easily find something to back up my claims. Right. If I’ve read something about this, that and the other and I need some citation, that’s really, really worthwhile, because then what I have read becomes the scaffolding I need for the academic paper. So Deani am I being a bit cheating by saying that, or do you think it’s fair enough to say that reading to somehow organize it so you can easily find it, to cite it as an important part?

Dene Grigar: I don’t think it’s a either or. I mean, I think it’s both and. Right, I mean, we sometimes I come up with a hypothesis and I’m looking specifically for other people that might have thought about it, too. And if I can scaffold my thinking with someone that’s already thought about it, to see if I’m taking it in a different direction, if it’s new virgin territory, or if I’m just plodding old ground. Right. That’s not not very interesting. So I look to see if there’s people that have already written about this or thought about it at the same time I read, because I’m interested in ideas and I want to see like, what’s out there. I didn’t know much about spatial hypertext until I began reading around it in about a year ago, a year or two years ago. There’s a whole body of work in this field, you know? You know, Klaus introduced me to it. And so now I’m reading about it. But what’s interesting about reading about spatial hypertext is Mark pointed out this morning is that there’s the engineering view of it, and then there’s the other view of it. And Mark Bernstein is the other view of it, which I think is much more useful for the way I think. Right. And so there’s different ways of thinking about spatial hypertext. And then that becomes a synthesis activity. Where do I fit? How do I organize this information and where do I fit within that information. So if you follow Bloom’s Taxonomy, synthesis is a high level thinking before you start your evaluation. So there’s. I don’t think it’s a I don’t think it’s one or the other road. I think it’s both. But what I try not to do is go look specifically for things that are going to back me up, because I’m probably going to miss something. Because there’s opposing views. As John Barbour often says, you know, questions, viewpoints, opposing viewpoints, you know, questions, answers, opposing viewpoints like what is what? What’s the opposite view of what I’ve got?

Frode Hegland: Yeah. No. Absolutely. Absolutely. So yeah that’s that’s very good and reassuring. So I’m going to share a screen a little bit. You can see it right. Yeah. You can see. Yes.

Dene Grigar: Can you make it larger?

Frode Hegland: But I can, yes. Boom! So it’s small for a reason. You don’t actually need to read it. I’m going to spend 20s on this because I think you’ve all seen it. But this is what I extracted from last year’s proceedings. So what we have is a title list of authors. And then the first sentence of the abstract. It’s just arbitrarily for this. The title here is. A defined term. All that means is obviously, you know, but for context, for this, all it means is, is something that has expressed metadata stuck to it. That’s all it means, right? So it happens to have an AI summary which doesn’t need to be there, but it does. It does have the authors names, which is crucial, I think. Then it also has keywords that can be AI, human or mix whatever. So the point of it is when I go to the map view, we have this. These things. Connect. I have a little bit of an issue here, so it doesn’t always connect as it should, but for instance, lower left hand corner here are people that I work with. So this is for me valuable, so I can see where they’re from. Up here are people who are referenced in the text, but they are not. They haven’t written anything this year. These are people who have authored stuff. So that’s just what I happen to have done in author. That is not what we’re trying to recreate, but what we are trying to recreate is some sort of manner whereby you can load the proceedings, hopefully this year’s proceedings in a useful way. And choose to have different views, different layouts, different interactions, so that you can find out what you want to read. Annotate. This is important. Get rid of this and also to help you find it in the future.

Frode Hegland: So those are the elements that I’ve emailed you guys as a starting position. And I’ve gotten quite a few comments back. Thank you. So now I would like us to look at. Or getting something more on that list. However, I think it is really, really important that we have a discussion first without being constrained by this as to when you have this magical, massive reading space. And have all the papers from the proceedings. What are the things that you guys want to see, both in terms of nodes and actions? Hey, Brando. Good timing. We’ve gone through things, and we haven’t done Fabian demo yet, so that’s good. You’re for that. What we’re talking about now is Andrew is building a graph, a knowledge space, a map, whatever to call it, as we talked about when we’re together last time. So now we’re discussing, considering that our default set of information is one proceedings slash one journal all the papers in it. What is it we want to see on this screen and why? What will help us decide what is important or not? The final thing before I give the mic to to Leon, that is in discussion with some of you. One thing that has become clear to me is that the fact that they are all from the same proceedings is very important, but only to a point. So once you’re in a map view, you probably want to have the proceedings just be another tag, so to speak, that you can show and hide, and you should be able to bring in more proceedings and other documents and so on. So what we’re talking about with proceedings for our own use case here is only really important for the initial confined data set, right, Leon?

Leon Van Kammen: Yeah. So I was thinking about your question. What do we want to see when we enter this incredible big space? And that is quite quite a big responsibility because as you know, that you can only read one sentence at a time. So it’s very easy to sort of yeah, get too much stimuli. And if I think about it, about our computers, if we boot it, we start with a clean slate. There is just a desktop, and there’s basically nothing. It’s it’s more the your personal journey which decides what is going to be where. So you start with one window. So what I can imagine is that you also start with a clean slate. And then if you open a document, then maybe if you. Maybe that would be the starting point, that you can maybe pull out the same document out of the document to have sort of two views, where you can maybe leave the page on the page on, on, on this side, and then you can start maybe scrolling or whatever, I don’t know, but basically that you basically build out your journey on the fly and maybe later some presets. But basically cloning and splitting up windows is, I guess, which will be the most understandable from the user perspective. Maybe I’m also completely wrong, because maybe with certain tasks people want a sort of, you know, opinionated layout where things automatically snap to a place or so. I don’t know about that, but yeah, those are my thoughts.

Frode Hegland: Thank you. I just want to when you’re talking about starting with a blank slate and all of that, that’s cool. But for the person sitting down. Put on the headset. The first thing they will be shown is all the papers listed for this year’s HDMI projects. That’s our goal. And the goal is that when they take the headset off, they desperately want to find a headset for themselves to go back in because it will, in actuality, help them decide what to read from the very conference they are at that that is our number one chief goal. Mr. Yellow Hands, near Southampton.

Mark Anderson: Yeah, thinking of this paper had me thinking a lot, actually. But one of the things that’s sort of coming to the surface for me is that rather than think about what what things do we want in the paper is to think is to separate, have a really good separation between the objects on the map. And then the data we attach to them. So this blob here might be the third sentence in the fourth paragraph of a paper. This might be a whole proceedings, and it might be pertinent to have them in the same view. Another time you might want to decompose a paper, you might want to decompose in a view, or you might want to decompose to a different view. So what this also says to me is this sort of speaks to a sort of view specs. So these are just presentations of the same underlying data or a subset of the broad underlying data. And so what you see is part of in essence, the initialization of the view. So in other words, rather than opening it with something in there, it’s conceptually slightly different. You take some stuff and say, right, give me a view of this. So you pre-decided before you go, I want to see all the papers that were in hypertext 22, or I want to see this ad hoc list of mine that I’ve made up because I want to review it, which is similar but slightly different to the way that I think it got presented.

Mark Anderson: But I think that was useful in bringing out this nugget because it gives you a much cleaner separation. We don’t have to worry about so much defining as to, in a sense, what a paper is or something we have we we can have objects, and as long as the object knows where it is in relation to a reference point in the view, it doesn’t need to store lots of data about where everything else is. If it knows the origin, whether it’s in a 2D or a 3D space, if it knows where it is in relation to that, it can compute where everything else is in that space. So that cuts out having to store, I think, a lot of data. So therefore we have an object in our nice malleable VR space. Then in essence we can clothe with anything and dependent on your style. If you’re someone who very much likes to see things that are sort of quite literal, you might want to see a paper represented as a card or a white rectangle or something like that. Whereas another person looking at essentially the same data might just need to see a blue blob, because they know that blue blobs aren’t green blobs.

Mark Anderson: And I think that makes it an awful lot easier. It’s sort of it it’s to sort of control this and get the interaction in and out. So we can therefore, for any given level of abstraction, you know, we can know that proceeding is broadly a binding of a of a number of papers. We know that a paper will have some metadata that we think we can access about it in terms of its internals. And so there’s a degree of sort of nesting and association there. And we can traverse that. So in other words, we haven’t got to say that I need to have these things in the view before I start and then find, oh crap. Actually, there was the thing I needed that I haven’t got. We ought to be able to traverse to it in, in this, in this view. And I think we can do if we, if we separate the data with the data and the visuals. So whether it’s shown as a card or whether it’s shown as a rectangle or whether it’s shown as a pink flamingo. Doesn’t really matter, because that’s just what we do when we set up the view spec. It’s not a property of the object.

Mark Anderson: It it gains this, it gains this. The styling. By the dint of what we want to do with it in the space. The data about it, the factual things like the title of the paper, it’s a paper is available by the metadata. If we want to see the title, we can tell the object, we want to see the title. And we could probably do that via classes and prototypes and things like that. But I think that seems a much, a much cleaner way to get into this. And that, I think, mirrors 20 odd years of working with knowledge maps in tinderbox. Half of which I spent really getting it wrong. And the main lesson out of it is less is more. So don’t start with a ton of stuff and try and hide it. You know, start with a best guess of what you want and and. Annotate or style the stuff as you need it, you know? In other words, don’t don’t make the computer tell you what it is you tell you tell your map what it is because that way you actually get you get practical use and you get information out of it. Otherwise you’re essentially looking at an infographic, which might be useful in some circumstances, but I don’t think wants to be your starting aid. Thanks.

Frode Hegland: Well, just briefly for framing before

Frode Hegland: For you for a couple of assumptions here we need to agree on one of them is the default view here will be the titles of the articles, because that is our key working thing. Should we be able to do the things you said? Mark, I completely agree with you. And then you mentioned blobs and paragraphs. That’s not really what I’ve been thinking so far. So I think that’s very, very useful that conceptually at least should be able to extract things and put it in this view. Yeah. So just just.

Mark Anderson: To clarify, the fact that you start with a title is just the style you’ve chosen. In other words, it’s not a property of the thing. It’s just that you say, I’m going to initialize this view with a set of papers, and the styling I wish to have is I wish you to go to the metadata, get the title field and draw it onto the object. Yes, the slightly different thing, but I think it gives you a lot more power and control.

Frode Hegland: We absolutely need to have different themes, so to speak, and an abilities for the user to be able to see things visually in different ways. Absolutely no question what we can do that. But we do need to have a default situation. Someone puts their headset on and they see the papers in the proceedings, and there is further data and interaction there, right? Okay.

Mark Anderson: So I think you misunderstand what I’m saying. I’m not saying you can’t have that, but what I’m saying is that would be the styling. You would, you would set up. So you, you basically would call a view. With with saying which what styles you wanted for the objects.

Frode Hegland: Oh, yeah.

Mark Anderson: You’ll see a view. So I just you circle back. Said it was important. Then you circle back in a way that implied that it was something different. But I think it’s entirely doable. So if we if we want our if we want our view of a group of papers to be a number of objects in our space showing that title, that seems perfectly doable.

Frode Hegland: Yeah. And what I’m really trying to establish today with Andrew sitting here with a massively open mind, is the initial view and the options the user should have for changing the view and both the interaction and the opportunities. And Fabian, is it okay if I ask a question before you? Okay. Thank you. So, Dean, in writing here, you’re talking about the attributes of connections, which is of course very, very important. The initial view is. Of our of the papers in the proceedings. So they will be citing earlier papers. Are you in this context also referring to conceptual links within the papers, or are you referring to how they link to external papers?

Dene Grigar: Probably both, because in your article to me, your essay to me, you asked for responses. Right? And you describe you pull that up here. You describe. The attributes of connections as thinking about color and thinking about when they appear. I also think we need to think about why they’re there. Oh, absolutely. How they’re functioning. Right. So not just those two, but adding that third element, which I think is more important than color and weight. It’s like why we’re using them. Why are we making this link from this object to this object? Why are we making that connection? Yeah. And so Jeff Parker did a really good job years ago, kind of defining how links function, dividing them into two types. And one of them is functional. Right. And so I’m just suggesting we’re talking about the function as a third option. And under that category he had blatant as one of the options. So. Yeah. Oh, so blatant, blatant links. But, you know, functional links that are blatant. They’re there for a purpose.

Frode Hegland: By the way, Ingrid says hi. If she is partly here, she’s getting ready to go from one country to another. Ingrid, when you feel like it, please raise your hand for an introduction. Yeah. Danny, that is really, really crucial to the kind of link the connections I was talking about is when you highlight some node, something appears to others. So it’s exactly the kind of thing we need to discuss. So I’m just wondering, in terms of our discussion how we should layer this because we need to think about what kind of things should be on screen and what kind of interactions, to see how they connect. So yes, let’s keep drilling on that. But let’s go over to Fabian, whose hand must be getting really tired.

Dene Grigar: So let me just clarify one more thing. So you’re saying that we will add a third category called function.

Frode Hegland: Yeah. But I think we need to discuss more what we mean by function. And let’s do that immediately after Fabian. I don’t think we should do that later, because I think for Fabian, we’ve also got like programmatic links and stuff like that. So yeah, we should absolutely do that. Yeah. A phobia.

Fabien Benetou: And so to to the question of what should appear first, in my opinion, it’s an onboarding step because the person they’re, they’re they are not familiar with probably anything yet. They’re familiar with their own goal, like how they expect to get there somehow, but they’re not familiar with the interface. They’re probably familiar with the document. They’re not familiar with interaction. So it’s I imagine it as an onboarding session. For example, I want to share too much about my personal life, but I do enjoy video games. And I think what is and I think adults should play video games and I think adults should be playful. I don’t want to backtrack too much, but that’s why I also I showed the the gamepad earlier. I think games are fun. And one thing games do really well. At least good games do really well. Besides, making you have fun is a kind of sense of empowerment. Like you can do things you would otherwise never be able to do, like being a superhero, basically. And how they do this, besides the narrative is a first level, like a session of couple of minutes where you basically learn the control like the A button when I press it does this. So, for example, they throw at you a gentle enemy, somebody that is not too hard to beat up, basically a bad guy that you need to beat up.

Fabien Benetou: And and those very first moment in the game, in my opinion, are crucial. Like if they were bad even if the game is good, like, people are just not going to play it. And I think we can basically learn a lot from video games, those very first moments in a game. Because again, you’re literally thrown in a virtual world that you’re not familiar with. It’s a weird moment, let’s say. So I think doing an onboarding session makes sense. The question then becomes is for whom? And I was thinking so I ask Dini, Marc and others on the slack to basically tell me about the job, like, how do they do the jobs? And that was very helpful and precious. The thing is, we basically have, I think, to be at the intersection of all those roughly and let’s say somebody that would be a colleague of Marc and fraud, who is going to that conference so that if we make something specifically for Dini, it’s great for her. It’s honestly great for Marc. It’s not necessarily great for somebody else. So it has to be kind of the intersection of all those needs. So I imagine that what we see first is an onboarding for somebody who is at the intersection of Denise needs, Marc’s needs, other academics that come to that conference.

Frode Hegland: We absolutely need to do that. But I really feel the approach we should take towards that is make the initial view spike very, very clear, obvious and simple and provide very clear, obvious and simple initial interactions that that the user can do. You know, I’ve been imagining this for a while now, and I could imagine, you know, we’re all in a group putting on headsets. And I can imagine Deeney for some reason. I visualize you, Deeney putting the headset on someone and I’m just listening and all you say to them is you will see this when you put it on. You will have a control here and here. Good luck. You know, almost that simple. And then when you get in there it’s like, oh, I do this and this. So the experience itself should be revealing as well. Leon.

Leon Van Kammen: Yeah. I want to talk a bit more about this revealing and as well as Fabian’s very good point on you know, this, this sort of initial onboarding. Because I noticed, too, that a lot of people you know, for a lot of people already, just seeing something or experiencing something without using controllers is already yeah. Is something what they are comfortable with? The moment they have controllers, it starts to feel a bit more like, oh, I have to what what should I do? Like this? This whole question of what should I do can already be a bit harder for for certain people, but I was wondering. About the lowest common denominator of what Denny would like to do when she puts on the headset and, you know Mark, you anybody there at this conference, like, would it be something what what would be the primary user story? Would it be like showing hiding things without with their fingers? Would that already be a sort of like something what would impress all different types of users? Seeing the metadata or or is it actually the editing what you want to show as a as a primary use case? Sorry for the hard question.

Dene Grigar: Shall I respond to that or do you want? Shall I wait till you’re done? Okay, so I think there’s two answers to that question. The first one is, I mean, I want to do something I can’t do in real life. Right? And that’s have enough space to work in, to have enough space and enough places to put things and to think. And for me, space is thinking, I need lots of space. I can’t be confined. The second thing I like about the VR environment for the work I do is the idea. I can share it in situ, you know, so the spatial persona where I can bring somebody in and we work on something together in that space. Right now we’re working on papers and Frodo does not like Google Docs. Right. So we’re he’s passing a PDF to me. I’m making comments. You’re making comments. He’s having to to take all that information and put it in one space. That to me is not useful, right? It takes time and it’s wasting his time, which is precious. It’d be better if we could all enter the space together and read through the document, read together through the document, and work on the document in situ at the same time, and at the same time. We have our own spaces in front of us, right? So we have the shared space and the and the space that belongs to us, and we can work simultaneously in those spaces. So that kind of experience I think is important. We’re not there yet. But we will be one day. And that’s that’s the hope. That’s why I’m interested in this project right where we’re headed, not where we are. The dream. And for you, Frodo, in the chat about the functional links.

Frode Hegland: Yeah, no. That’s great. Thank you. So I don’t like Google Docs. That’s absolutely right. You know, in some cases it may be useful, but I came across an interesting article on why writers generally don’t like Google Docs. Because in Google Docs, you don’t know if someone else is writing. At the same time, you don’t really have control. There are different kinds of things. The only reason I share a PDF is that so we could easily cite each other. The model that I would really like to see is I send a PDF, and you guys copy from that into a new document as a citation. So when I read your comments it cites back to that paper which is, you know, kind of a simple model. Just commenting on a PDF, I agree with you is is suboptimal. But of course now we’re getting into authoring issues, which is next year though of course mentally we need to prepare for it. So the notion of these links. And thank you, Danny. I. I’m not sure where we’re going now because. I. These are links in the click to go somewhere links, as far as I understand, what I was referring to was visually connecting lines on the screen. But so far, the initial view of the proceedings. Dany, do you want us to go from one place to another? And if so, who would be making these links?

Dene Grigar: Well, I think it would be shared, right? I mean, I think the answer to that is once again, both and I mean the author, the thing I love about Open spaces is that the author can insert linkages that they see as important. But I think what we’re trying to do as well is leave it open for the reader to also make linkages or change linkages that the author made. So we’re constantly refining the document.

Frode Hegland: Yeah. But the in terms of our very specific initial use case my understanding is that we are simply presenting the user with the proceedings as they are and giving them a wide workspace to look at them in different ways to decide what is relevant to them or not. So we definitely need ways to escape that to go into different spaces, but I think it would be really good if we primarily provide the, the user with a, almost just initially a list of titles, and then they have controls in some way that we have to design both how they interact and what they do, where they can specify. I’m interested in this or that. And, you know, it’s like Disney’s Fantasia, you know, that they can paint with the space and, and move things around with the elements that are within this space to make it really concrete. Well, in.

Dene Grigar: Year one, we’re, we’re providing them with reading experiences. And year two, we’re going to provide them with writing experiences. So I think while we’re giving them in this paper, in this example, is what we have laid out for them with the promise that in the next iteration of our project, when we go back to wherever the ACM hypertext is, which I hope is London. They will have the ability to manipulate those links and change them and define them themselves. So I think the answer is pretty, pretty simple. Right now we’re we’re offering this for them. We’re making the demo. We’re showing them what’s possible because they can’t they don’t know. Right. We’ve got to teach them. It’s the breadcrumb theory. Again. You got to show them what’s possible so that they can take it a step later.

Frode Hegland: Are you saying that we should go into the papers and add these types of links within the authors papers? But now I’m a little confused. No, no, no.

Dene Grigar: The author’s already got the paper. No, we’re just. We’re just highlighting what’s already there.

Frode Hegland: Because. So.

Dene Grigar: Because we’re not doing any hyperlinking. Besides, like, here’s the title of the work, here’s the abstract. We’re not we’re not creating any. Choose your own adventure. Links. We’re showing them the connections among that I call I call it the iterations of the document, right? There’s the iteration of the of the the abstract, the iteration of the reference, the iteration of how it connects to another article. We’re just showing that as opposed to, hey, you’re going to read this and go this direction. Oh, then you’re going to change your mind and go this direction. We’re not we’re not doing that.

Frode Hegland: Right. Okay. That’s fair enough. Okay. Cool. That makes sense.

Dene Grigar: Does that help?

Frode Hegland: Yeah. Yeah, yeah, that that makes sense. So what? So when you when you’re referring to the functional links, you are referring to, not not something you want to specifically to implement now. Right.

Dene Grigar: Well, we already are implementing them by making those connections we’re showing. Like, here’s the references. Oh, by the way, when you click on this one you can then link to the abstract. So we are already doing those connections for them. We’re not changing the text in any way. We’re just pulling out data for the user, the reader to access. Those are functional connections.

Frode Hegland: Okay I, we have slightly different language, so that was very useful. Thank you. Yeah. Okay. Mark.

Mark Anderson: Yes. I think one possible confusion is I, I certainly the picture you show of the, the the the author map is, is not something I’d start with because it has way too much information and less is definitely more in this thing. So if you’re looking at a set, I mean, if if our start point is to look at the at a binding that is that happens to be proceedings. So proceedings will be a collection of, say, 50, 70 papers plus front and back matter. That’s going to give you a 70 objects alone. They actually don’t really have any explicit links apart from the fact that they’re members of the binding. And if you’re looking at the papers, actually the thing you will need to do is, is basically, well, the one, the one thing you’ll have outset will be a title. And on that, you’ll probably decide whether you want to drill further. And two things I put my hand up for to, to, to circle back a second. In terms of comments, I think there are sort of two sorts. There’s almost comments, as in almost a conversation, a common space, and comments that are more to do with sort of understanding and argumentation. So the the comments I was putting the PDF the other night were really I do understand this. I don’t understand this. Or there’s an ambiguity here or there’s a clarity there. And that’s not something I think I’d be doing in a sort of common space, because there it’s slightly more of a conversation because you’re, you’re essentially speaking by text to the other people in, in, in, in the space. And in terms of links we definitely don’t want to start showing lots of lines because that just that’s horrid. Links. One of the problems with the link is the moment you two things on the page have a line between them.

Mark Anderson: They’re almost privileged. Your eye can’t step away from that. And something is very nice in the implementation of your mapping author, is the fact that the links aren’t shown until, in a sense, you hover over something. I think that’s sort of visualization is quite a nice thing. The it’s the degree of detail downwards as to which and what links get shown. And indeed, whether you’re visualizing explicit links or implied links but or inferred links, the, the difficulty is that and again I go back to experience of working with okay, Mark Bernstein’s maps but they I, I think, I think the, the closest relevant experience I can think of for this is that it’s a gentle play between the display space, especially the machine and and you. So the problem is the wrong the sort of the wrong, the wrong question to ask is what should we start with? It’s very. What do you know? What are the starting materials. And then from you’ve got to choose everything from there on in, because two different people will read through that set of proceedings in a different way. Someone will probably go and start with the first item and look at the title, and then they’ll go and look at the second title. Somebody else may just scan their eyes around, dive in somewhere and want to go straight into a paper. Somebody else may want to consume the whole. So we have to we have to be we have to accept that sort of wideness of experience. Because if we just if we just force people one down, one method, what will probably happen is a few people will love it and the rest of them will never touch it again.

Frode Hegland: Right. So we I’m going to share a screen for a second. I think five more minutes on this and then we go to Fabian. I thank you for the comments you made about the lines and so on. But there’s another thing that I think is really important to highlight on the way that I’ve been working on this so far. One of them is about the lines that only appear when you select. Thank you for highlighting that. It is actually useful at times like now, you can easily see that most of the people I work with are from the UK. You know where they are. That is useful, but it is as useful or rather important to be able to hide. And reveal, right? And don’t forget this is a 27 inch monitor. I’m not saying we should have. We should start with all this stuff. That is not what I’m saying. But what I’m saying is we as a community, we need to decide on an initial view, obviously testing and modifying as we go along. And this interface is not very good. So whatever means we’re going to give the user to select what is on the screen and how it should behave, I think will be a major part of our work. But first, we really do need to decide on an addition to the documents which have to be on there at least initially, because that is our core data sets. You know, how can we help the user who is given this stuff make sense of it? And it most certainly isn’t what you’re seeing on the screen now. So of course we all agree on that. So I see Denise hand is up. We’re going to do a few more minutes after Denise before we go to Fabian. But if you guys could think along the lines of what stuff should be on the screen that’s useful, that would be incredibly useful. Thank you. I agree with that.

Dene Grigar: So just to respond to what was said earlier Marc helped me here because what I thought was useful about story space. Was its different views the outline view, the tree view and the story view. And so readers could, could toggle among the different types of views so that they can get the information they needed in the way they needed it. So you could look at a hypertext story, look at it as box nodes and paths. Right? Story view. Or then you could say, okay, I want to see the outline. I want to see how it’s structured. Click on Outline View. And then you would click on the tree view to see the hierarchy. Very very different ways of seeing things. And this is what you were talking about Mark about view specs. View view view specs. Is that the right.

Mark Anderson: Yeah well I was I was channeling sort of Doug Doug’s term which never stuck. But I think it expressed it beautifully because with with story space or indeed with tinderbox or, and you know, other tools following that model, you have one set of data underneath it so happens in the apps I’ve mentioned, it’s an outline, but it doesn’t have to be. And then there are you can have multiple views on that now on a 2D, on a 2D screen, on a 2D space, broadly, you’re only going to see one at a time, but in a XR space, you could possibly have those as, as separate discrete views within your overall visible space. You may choose to make it large and immersive, but it’s the same data underneath. This is why I was I was pleading for a very strong separation between the objects and then the data you clothe onto them, because otherwise it gets messy. If we go down, if we if we, if we talk to directly to literally in terms of labels, what we’re basically doing is we’re reverse designing from the view back to what’s underneath.

Mark Anderson: And that’s hard and messy and difficult. But I, I. I think I mean, so one of the things that we’re we have a slight issue. We’re going round, round circles in the sense that. So if if our starting thing we need to show is looking at a set of proceedings, then the map process has been showing us is not that that’s just a whole lot of metadata slammed onto a screen. It’s all relevant. That’s a bit that you want accessible, but you sure as heck don’t want to see it. That’s way too much. That’s all I mean. And Peter has some experience with these maps too, and can probably weigh in here, but the the experience from a group of people who’ve used this is that broadly against the expectation that that that more is less. So you want you want to start with not very much. And what you what you actually want to do is instead of showing and hiding stuff, it’s more that you want to style, for want of a better word.

Frode Hegland: Write what you have.

Mark Anderson: So what you want to do is you want to illustrate, illustrate or interrogate the relationships by, in a sense, altering the visualization of the points that you have. So this is why it’s key that when you start you say, okay, I want to look at these 20 papers, or I want to look at these 400 papers or these.

Frode Hegland: But, but Mark.

Mark Anderson: 600 people or whatever. Because that’s your starting point. And then you start, then you start interrogating and you, for instance, links. Link. As long as you know about a link, whether it’s visible or not is literally just a matter of make this visible or don’t show this link type, but show this link type, because you might want a common thing that’s quite useful is to have one or a few link types visible, and the rest are implicit and done in in in the users, the viewers mind. Okay, because that’s a bit they still have it.

Frode Hegland: I’m not sure why you’re highlighting this. Again, we agree I agree.

Mark Anderson: Well that’s not what you you keep showing us a map that’s fundamentally different to that. Okay? That’s why that’s why I have a concern that we’re not on the same page.

Frode Hegland: Okay. We’re talking about being on the same page. This is kind of important because the last thing I showed you was this, and I said, I thought very clearly, this is not what we want people to see. Well, let’s.

Mark Anderson: Not look at that. Can you show us something? That is what? Because that’s what we need to look at next. Otherwise we’ll we won’t move ahead. So let’s not look at that. Let’s look at the map that we might look at. And I’m sorry, I don’t mean that unkindly, but I mean.

Frode Hegland: No, it’s.

Mark Anderson: Useful stuff, but not, not not in the context of what we’re trying to bottom out here. Okay.

Frode Hegland: So the thing to me is that when someone gets into this, we they’re going to have a larger workspace than we have on our monitor. That is a very important part of what we’re working on. Right. And whether we show them just a list and they click a button to see one of the views. And I strongly agree with you, Danny, that it shouldn’t be one view we’re providing them. They should be able to toggle, as in the Mark and Adam citation demo where you literally hit the tab key. Strongly agree that is super important. But in order to do that, there’s got to be some. At least one of our views has to be. I am now sitting with an absolutely massive workspace, and we as designers are providing the means through which that is useful, right? So we have to show, at least in the beginning, all the papers, right, that are highlighted, there have to be here and then we have to provide them the means to, you know, hide a few, make a few special. All of that. Right now, we may choose to do something like this to give them special keywords on the side or whatever it is. This is what I think we’re designing now, but I don’t think that I mean, okay, it might very well be that we only show the papers without anything else, and we have a list similar to this where they can now say turn on this attribute or turn on that attribute. Yeah.

Dene Grigar: So we start with just that simple list. And then we have a menu that allows them to see, oh let’s look the tags. Oh let’s look at the the abstract. Let’s look at the you know and you have a choice. And then that kind of branching occurs. And it’s the branching that expression of the connections between this document and this piece of text that helps to describe it, define it. That’s going to be the playfulness of this piece. And it will not confuse the reader. Right. It makes logical sense. Yeah.

Mark Anderson: But I mean, the other thing is the reason I was sort of talking about the list is if the if the use case is just looking at the proceedings, that’s what you look at. And in fact, what you’re doing, what you’re really doing, if it’s a new set of works, is you’re doing a triage, you’re looking at, you’re basically asking, is it a next to no interest to me? Is it of an immediate interest, or is it something I want to circle back around to. And so keywords author stuff really isn’t relevant because that’s not what you’re trying to do. It’s relevant when you’re looking at your library. You might say oh that’s that’s by Cathy Marshall. Yeah. Didn’t she write one of those other what’s the other stuff that but that’s a that’s a totally different use case to I’m going to I’ve been given you know I’ve just had this stuff released. Because I don’t.

Frode Hegland: I think the.

Mark Anderson: Keywords the other thing is we know from research is the keywords are generally not any good. They don’t actually tell you very much. The main thing they tell you is something like the name of a system, and the name of the system is probably in the title of the paper. Okay. So that’s that. So there are lots of things that you can show, but it goes back to this fact that if it’s not actually helpful, it’s more noise.

Frode Hegland: Okay. Okay. Mark, in the interest of time, we have a few other things to go through. I have understood the point that too much is too much, right? We’ve gone through that many times and I strongly agree on that. Right. But we still have to decide together what the initial things will be. Right? We just we just have to have a few things on the screen.

Dene Grigar: So papers is the papers. Papers just the list.

Frode Hegland: And then what.

Dene Grigar: You get to click on it? Here’s here’s what I want. You’re asking me what I want. Okay, you click on one and you have a choice. You get a menu. The menu says you get to do these things. You can look at the tags, you can look at the abstract, you can look at, you know, the way it connects to others. You know, maybe the key words connecting with other, other items in the list. But we start with one thing, right? And then let people explore.

Frode Hegland: Right? I completely agree with that. I don’t think there’s any disagreement, but what we need to decide what’s on those lists and what are the options?

Mark Anderson: Well, you’ve just been told I’m confused at what you’re being told.

Frode Hegland: Okay. Since I’ve been told, can you do me the politeness of telling me again? Well, I’ll.

Mark Anderson: Just. Okay. To restate, Denise, this is you have a list of papers. So. So you basically don’t want to look at the paper again or you want to know more about it, in which case you want to drill down to the next level, which is essentially the abstract possibly the references.

Frode Hegland: So the first option is to reveal more per paper. Yeah.

Mark Anderson: Because because all the other richness that you’ve shown, which is in all sorts of other context useful, is not useful here. Not not when you’re trying to do this triage across papers.

Frode Hegland: I’d like to know what this.

Mark Anderson: Is, where the confusion has crept in.

Dene Grigar: Sorry. I mean I think yeah. So I mean, it’d be good to know. Okay, here’s a here’s the list. I bring the list back up. Just the list, nothing else.

Frode Hegland: The list is just a list of papers.

Dene Grigar: I know, but just humor me for a minute. Because you love me. Okay. Get rid of any linkages. Get rid of all the keywords. Just the list. Okay. That’s what I want to see. I want to start with that. And I want to be able to, let’s say go to the first one, Hypertextuality and virtual reality, which is my paper. Okay. Now I click on it and I get a menu, and the menu says abstract. Okay. Just like that. Right. Abstract. Author. Country. You know, the bibliographical information. You know also full paper. Can I see the full paper. Maybe I want to get that full paper. And then we go from there, and I can then move it out of the way and go to the next one and pull up a second paper and have them side by side. At that point, we can talk about making linkages within those texts in the future. Not today, not this year.

Frode Hegland: Why not? What do you mean by link linkages within those texts? Because okay, let’s say that you move this one over here. Then you go to this one and you have a look at more information. You read the the abstract and so on. If you don’t like it, you should be able to click a hide button. Fair enough. Or ignore or whatever. But if you like it, what kind of interaction would you now like between these two papers?

Dene Grigar: Well, put it to the side. Pull it up on the other side of my document. Okay, now you’ve picked Serge Boucheron’s work, which is about Hyperfiction. He’s a he’s an artist. And then I’m talking about Rob. Rob Swigart’s portal project. They make sense to talk to each other. Now sitting in that list, we don’t know that. But the keywords allow us to make the connection now. There we go. And that should also connect to mine. There we go. That’s what I want to see there. Yeah, right.

Frode Hegland: That’s that’s totally great. That is exactly what I’m trying to establish.

Dene Grigar: Yeah, but you’re asking what information we want the user to define when they arrive. And it is just the list.

Frode Hegland: And I understand that I’ve agreed to that quite a quite a long time. So I’m not sure what you and Mark want to do different. I completely am happy with showing a list, but we still need to decide through whatever means the user can have access to what then will appear, right? Well, that’s a different question.

Mark Anderson: That wasn’t the one that was asked. That’s why we’re confused.

Frode Hegland: So is the question that was asked. But the question that was no.

Mark Anderson: So what what does he what do you want to see when you start and what you want to see when you start? Is the list the supplementary? I think I think right to try and stratify this. What you’re saying is okay, so what do you subsequently what affordances or what other textual objects or objects do you wish to be able to show in that space? After after the initial reveal. But the initial reveal is the list of the papers in this use case, because that’s what you’re doing.

Frode Hegland: You’re going through with all loving respect. Mark, the very first sentence of the document I sent to you guys is this document is internal to us to help us list what should be visible in a map view in XR, and what metadata should be available about the listed elements and what interactions might be useful. Right. So I’m not sure what became contentious. Anyway. I guess we’re not going to be able to decide on any elements today, but Ingrid is here to introduce herself in the middle of our discussion. And then Fabian has demo. Ingrid, do you have a minute?

Ingrid Hage: Yes. Hi. Can everyone hear me?

Frode Hegland: Yeah, but we can’t see you.

Ingrid Hage: Well, I wouldn’t recommend that because I’m a bit under the weather, but fair enough. I mean, it is a disclaimer. If you guys are all fine with it, then so be it.

Frode Hegland: Hello. Yes, yes.

Ingrid Hage: Here it goes. Hi. I’m Ingrid, I’m a UX designer. And to be honest, I’m sorry I haven’t been completely following the whole conversation because I joined a bit mid-way. And I have a some some baggage to to pack before. Before I head back to Paris. So. Yes. What should the focus be on? Fraud.

Frode Hegland: Well, why don’t you mention. Well, first of all, I’ve told everyone you are my first student. Okay, great. So that’s nice. You were studying advertising at the time, but now you’re working on interaction design, which is highly relevant to this group.

Ingrid Hage: Yes, it’s more focused on user experience design. I love the interaction part, but it really depends on the project that I work on now. From from the bits and pieces I picked up because it is a complex project. It is a complex topic. It seems like you guys are debating on what to show an audience during during a conference, is that correct?

Frode Hegland: Yeah. The conference paper is wrapped in a proceedings is exactly. You know, someone puts on their headset, they get a list, and they can interrogate that list to understand what’s relevant to them and what’s not relevant to them.

Ingrid Hage: Okay. So they do get a list. So you do ask them what’s relevant or not to show.

Frode Hegland: Well we don’t ask them. We allow them to go through this list and we have to decide on what interaction should be available, what metadata should be surfaced and how. How should they be able to organize this information? How should they be able to see the connections between the different documents, authors, locations, institutions, and so on? So it’s the whole thing of what’s this all about without just reading the abstract, which may give you a good idea of the paper, but maybe not where it sits in the wider field.

Ingrid Hage: Okay, I it’s safe to assume that you guys have a very clear vision and objective of what you want to show exactly to the public and why they’re here.

Frode Hegland: No, that was the discussion.

Dene Grigar: That’s a problem. We don’t have a clear vision. We’re trying. It’s a struggle.

Ingrid Hage: Okay.

Frode Hegland: So why don’t you just solve it all for us, Ingrid, over the next minute.

Ingrid Hage: When is it due?

Frode Hegland: We will presenting that will be presenting this at the ACM hypertext conference in September in Poland. And Andrew, sitting there with a very bizarre grin at the moment, is the one who has to realize this to make it real. We are the ones who are trying to figure out what it should be. So it’s hang on, I think I mentioned this to you. I’m currently reading this book that is really quite lovely on this is it’s more on conflict resolution, but in a good way. And the legacy of the person who gave us our money for this. Alfred P Sloan. He did not like agreements. Famously, in one meeting, he walked out of the meeting because he was mad at the people there, because they all agreed. And he said, that clearly means you haven’t thought it through. So I think today we can say we’ve all thought it through because we have strong disagreements. We will get some fire and get some results. So it’s remarkable, though, when you drill down how different aspects of these things come up. So it’s intellectually it’s it’s fascinating.

Ingrid Hage: Well in these kinds of situations, I don’t know if in case this can help in these kinds of situations, what we tend to do at my agency and at the agency I work in, in Paris is workshops. So workshops to jot down ideas. They could be like standard workshops on figjam, for example. I don’t know if this sounds familiar to any of you. So it’s like a tool for brainstorming, and the idea is to jot down all of your thoughts and challenges and then start making sense out of them. So I think one way to kind of, you know, a spatial rearrangement of your thoughts might help if you guys are heavily debating on one approach or the other, and then they are prioritized. These are usually followed by prioritization methods. So the first approach is to set is to really agree. And this is super important. It’s really agree on the problem, the problem that you want to solve. And then the second stage is usually to kind of jot down the ideas and go into ideation mode where all ideas are welcome. And the third stage is to prioritize and think about impact versus effort. Maybe this might help in reaching consensus. Probably not agreement, but consensus.

Frode Hegland: Yeah I know, thanks for that. We’ve been going we have been going through various versions of this. And it’s also wonderful in this community that people have very different perspectives, some very technical, some well all over the place in different ways. So yeah, we will keep going through it and, and brainstorm in different ways as this is also an example of. Yeah. And then Peter and then we got a quick question.

Dene Grigar: Peter. Peter. Frodo, do you have a copy of your paper that I mean, where’s your paper sitting? For hypertext.

Frode Hegland: I don’t know. Yeah, I have a copy of it, and there is I’m doing some major changes to it at the moment.

Dene Grigar: Okay. So when do you plan to get that to us to see out of curiosity?

Frode Hegland: I can make an effort to have it to you. By Wednesday, I was planning to do a different thing. And then with Peter. And then I think Peter’s paper that he gave to me as a section was good enough to be its own paper. So it’s going through a bit of changes. And also the demo paper seems to be very much what you are helping head. So. It’s been a few things there. Sorry.

Dene Grigar: Okay, so so let me just clarify then. The demo paper was something that a small group of us was doing. Peter and you were doing this other paper and you were going to lead the team paper as a team paper. Still going to happen.

Frode Hegland: Yeah. The team. The team paper is the one I’m talking about. That is what I’m working on. Yes, but the demo paper is also a team paper, right. Because it is demonstrating what we’re doing.

Dene Grigar: Yeah, but I guess yours is the longer yours is, the bigger paper. I’m just wondering where we are with that. I haven’t seen it.

Frode Hegland: Yeah, we still have a couple of weeks, so we we can spend the week on this. I think. Should be good. Also, I’ve been thinking more about the elements that we should have in the view, because I believe that will really help inform what’s actually going to be written about in the paper, because it will help inform what we’re actually doing.

Dene Grigar: Oh, so just a just a quick thing because the deadline was the 26th of May. Of May, which is just coming up. Right. Which is great. The original deadline, my goal was to have the demo paper done before then, because the next thing I’m doing is the hypertext exhibition in Victoria. And once I put this paper into the ACM hypertext upload, I’m going to turn my attention to that, because I have a lot of logistics of moving a bunch of people, a bunch of equipment to Victoria for a week. So I’d like to see the paper and be able to comment on it before I get sucked into that.

Frode Hegland: Yeah, yeah, absolutely.

Dene Grigar: Around. And I’m not going to be with you guys in London. So that’s the other thing.

Frode Hegland: No, I understand that. But my priority really has been to try to figure out what to put in the screen, which I really. Oh, yeah. Ingrid, good to see you again. Have a safe trip. Please. Ping me when you’ve arrived safely in the land for coming.

Ingrid Hage: Thank you.

Peter Wasilko: Thanks for joining us.

Frode Hegland: Yeah. See you later. Bye. Yeah. No, Danny, I just think the discussion we kind of had today to to find out the initial things really need to be nailed down. That really needs to be our priority, because whether it’s what they say before they click a button or not, we still need to know what things to have on screen. Without that, Andrew can’t build anything useful for this particular view. And I really think that a map or graph or whatever thing is the one view that is, let’s say a native to XR, because it can be big, especially in the cylinder view. It’s just absolutely crucial for us to get these potentials down. Yeah. Mark.

Mark Anderson: You keep checking back to see what’s on the screen. I’m getting very confused because I think. But what you then go on to describe is, for instance, so you said, you know, if you’re looking at the papers, you want to see a list. That is what is on screen. I think what’s still in play is what additional information you may show. Well, a start point before you even get to that is what we may reasonably think we know about the objects that are papers. I mean, it’s great that you can put keywords up, but the keywords that you’ll have at the time, especially if it was a at a conference, will only be the authors keywords which we know from research are next to useless, so.

Frode Hegland: They don’t have to be. It doesn’t have to be just that.

Mark Anderson: No, no, I that wasn’t an exclusive list. I just using that as an example. But so. So we know what needs to be on screen. The papers. That’s the answer to that question. I think the separate question you’re asking is then what further things would you like to how would you like to enhance the screen which, which which is a slightly different thing. And I think that’s why why there’s been this sort of loop. Because I think we’ve answered the first thing in terms of you’ve asked the question. A number of people have, I think, consistently given you an answer to it. Yeah.

Frode Hegland: And I’ve consistently agreed with it. Mark, I’m very happy that the initial thing is a list of the documents, and I absolutely agree with you. The next thing we have to do is decide on what things should then be available to the user.

Mark Anderson: Okay, but that wasn’t okay. Well, that wasn’t how it was stated earlier, which is why it was confusing because you said yes, I agree that. But then we need to decide what’s going to be on screen. But what is the.

Frode Hegland: List on screen doesn’t necessarily mean visible at all times, but something you could toggle on and off. And we have to decide that’s okay.

Mark Anderson: That’s that. You didn’t explain. That’s a very critical.

Frode Hegland: Thing with.

Dene Grigar: The stuff that’s already in the menu that that you’ve got. I mean, what’s wrong with what’s there. Can we just review that and make a decision of what’s necessary and what’s not right away? That would be an easy, easy approach. Show us what’s in that menu I did.

Frode Hegland: That’s the document I sent. That is what that is. And the thing is.

Dene Grigar: Okay, that wasn’t clear to me either. That was not okay. Thank you for clarifying. I’ll go back and look at it now. This is finding things.

Frode Hegland: Sorry guys, but just for absolute clarification. You know, I don’t think we should just copy author. Right? That is why I experiment with author to learn from it. Just like Mark is an expert in Tinder box. Right. So I really think it’s really worthwhile for all of us to write down what they think for themselves would be the best kinds of things to be able to be. Right. So that’s what I was pushing for. I wrote a list of these are the elements we have in the metadata that should be made available. But also there are things like tools for doing layouts and views and things we need to talk about. So we’re getting closer to understanding each other. And this is one of those, weirdly, as Mark just said, close to but not the same. Right, Peter?

Peter Wasilko: Okay. When I’m looking at a whole big collection sometimes I think it’d be more useful just to have a structural sense of what the papers have in them. And what I’m picturing here is, like a pie chart. And one wedge of the pie chart would represent how much of the paper is devoted to a background historical survey. Another wedge would be how much is devoted to. Proposing new concepts. Another wedge could be. References to people and places. Another way.

Frode Hegland: What do you mean by wedge? I don’t mean to interrupt you just for clarification, so I understand.

Peter Wasilko: Okay. Just a rough sense of how much of the paper. Addresses that topic now. Imagine you had a user interface to help people provide this amount of data. It could consist of a set of questions and one question. This paper discusses past work. Strongly agree. Strongly disagree. Somewhat agree. Each of those metrics could then determine the size of that wedge, and it could be represented as a pie chart. And then I could have a pie chart based controller where I could resize the wedges, and that would control the papers that are being filtered on the display. So I’m looking for papers that are about experimental results. I could sort of stretch the wedge on the controller to be really wide, and as I made that wedge, bigger papers would drop off until I was left with a number of papers that seemed like a reasonable number for me to take a look at in that context.

Frode Hegland: Okay. Peter, that’s interesting. Please feel free to write it down. There’s a couple of things. Number one with that approach, it may not really use the space. We’re not really looking at what is uniquely XR that you wouldn’t want to do on a laptop, so to speak. And also a definition of how the the wedge data would come from would be useful as well. Andrew. And then we are rushing over to Fabian.

Andrew Thompson: Yeah. I’m fast, don’t worry. Just because I know the whole discussion for what does the view look like has been pushed today? Because that’s what I’m starting development on today as well. I’m not sure if that’s the same reason or coincidence. I don’t think we need to worry about it too much. Because of the time. For now, I’m just going to kind of put it in a blank space, just get all the the keywords in there and we can worry about design later. Just because I know the discussion went on for a bit longer than we planned. So hopefully no stress there for everyone and we can refine as we go.

Frode Hegland: Yeah. Thank you for that, Andrew. But the thing is, today was really tough. But it is useful because we look at it from very different ways, and we’re really hammering away at this massive rock to see the shape of it. So yeah, let’s continue on Wednesday. And I contend Right? Yeah, I see what you’re saying, Danny. But the thing is, this is all fantasy right now in the good sense, right? So if we all write down our fantasy options, it’s really quite useful. I know that a lot of what I’m writing down is really boring and dry. And let’s talk about it more on Wednesday. And Fabian quick before someone else. Can’t hear you. Can’t hear you. Oh, you’re not speaking. That’s what that’s. That’s a good reason why. But did he disappear now because he tried to share. Right. Okay, well. Well, he’s trying to come back in. Oh, yeah. You’re back. Hello? We can hear you. Yes.

Fabien Benetou: Yes. So I’ll. I did not share my screen earlier because. Can only do it once, so please let me know if you can see my screen.

Frode Hegland: Yes.

Fabien Benetou: Yeah, super. So I actually posted it on slack the other day. And you can see this list of papers or the titles rather of papers. What I hope is interesting is that I don’t actually even know now, I know, of course, what that list is, but I did not know initially what it was. No. Cared. The thing I knew is that Andrew saved it. So I was able to show it. I did not think too hard about the layout. I mean, I started from the bottom, I think, and I went up so there you can see the layout is a little bit of a mess. It’s just to highlight that I can do my usual gesture, namely pinch on the first letter with my right hand and move it around. And then I could do the rest. Let’s say when you drop it, something else happened. And one thing that happens when you drop it. So if you pick that let’s say the title of exploring the feasibility of crowd power, blah blah, blah what it does, it saves it right away to this URL. So what’s interesting is this URL is the same thing that Andrew gave me or shared with us a couple of days ago. Plus on the only on the title moved, there is a new field. So in that paper, let’s say number six with custom data field, or rather number seven there is a new field called position with x, y, z coordinate. So if I refresh that page like here, then I have those two papers in the right position, basically not the right orientation but position. So that means if Andrew loads back this URL and takes into account this position, then we can start to go back and forth between the two system and see the impact of some of the things we would do.

Andrew Thompson: Fabian. That’s awesome. You got that implementation very quickly. It was. I think I gave that to you like the end of last week and you already have it working, which is.

Fabien Benetou: Thank you. Well, I kind of promised it, so of course I had to make it work.

Andrew Thompson: Do you think it would be useful to rename position as a as a variable to, like, title position or something? Since we could also be moving the tags around.

Fabien Benetou: It’s whatever you want. What I suggested also before is as long as you start to make change, then we start. We need to actually agree. For now, it’s it’s more of let’s get the ball rolling. But whatever you prefer.

Andrew Thompson: Yeah. It’s not really a problem. Since you can name the position data, whichever you’d like. Since I, like I said, I use my data a little bit differently, but as long as we can both pass it back and forth. Yeah, that’s that’s really the start we need.

Fabien Benetou: And so I did not I don’t remember if I shared I don’t think I even committed that code, but honestly, it’s like 20 lines of code maximum. It’s very short. And the saving to WebDAV, I think, could be interesting for you. Again, because it’s part of those 20 lines. So you can even save it on my server so that whenever you have to save it, it’s it’s just experimental, like, it doesn’t take into account like username, password, these kind of things. But I think none of the information there is private. And it’s definitely not a personal information that somebody would not want to share. And with big red lines everywhere to say, hey, this is experimental, but I think if you, if it speeds up the way to do back and forth of exchange of data, I think that could be also a valuable shortcut.

Frode Hegland: In the in the discussion here between Fabian and Andrew to obviously connected with what we were just talking about the thing that. I want to make that clear is, first of all, strongly, no questions asked. Agree. We start with a darn list that is useful. And I really want all of you to fantasize about what can be next, because I can imagine Peter said all kinds of wild and crazy things, as he often does a short while ago. But let’s say that Peter programmatically managed to build a thing in this space, which helped do that. That is something that should be possible, right? I can imagine Dany, with her rich understanding of the history of this and electronic literature, which is not the same, but related. She could maybe even put together an LM that would be just look like one node in this environment. She touches it and it gives all kinds of information about these new papers in relation to what she’s already read. There’s a wealth of opportunities here. But as Marc and I agree, every single day overload the user. It becomes noise. You know I am a creative by birth, not an engineer, and I’m certainly not an organized person. I can think of a million things to be in there. They’ll mostly be rubbish. I think it is so valuable that all of you with pure minds, just write down what the hell you want. If you want a pixie to run around and do dust on these things because it’s useful, great. You know, if you want this, that and the other, I’m just begging to unlock these amazing minds we have in this room.

Mark Anderson: So we can we capture that as a use case, because that’s actually quite different to the use case we were discussing earlier, which is just trying to read a set of proceedings. I mean, they’re really interesting super ideas and I get them. But I mean, I don’t think they’re particularly germane in certainly as we stand this year in terms of how, how we reasonably use XR to look at, at understanding the nature of a set of proceedings. I stress that because I don’t I don’t want that task which we’ve taken upon ourselves to trample the other, the other much more creative thing. But that’s really blue sky by comparison. And that isn’t something that we can really make much, much movement on. At this, at this moment, in terms of the use case of showing someone the proceedings.

Frode Hegland: I have to apply to Denmark. Please. I’m so sorry. Yeah, sure, sure. Only because it’s time sensitive. Danny, I probably won’t be able to be there on Friday. Your morning. Because it’s my brother’s birthday weekend. We’re going away. But I’m very happy for you guys to meet separately if we don’t find a different time I’ve talked enough about. It’s just.

Dene Grigar: Logistics. I just need to get stuff.

Frode Hegland: Yeah, so if you guys can do it, great.

Dene Grigar: Okay. And also need to tell you that I need to get with you so I can finalize your. I’ve got your summer payments to get you. So I’ve got a lot of logistics to get done this week. Yeah.

Frode Hegland: No payments are great pace for programing. Yeah okay. Thank you. Thank you Danny.

Dene Grigar: Bye everybody. Thank you.

Frode Hegland: See you Wednesday. Thanks. Okay. I’m sorry. Fabian. Mark. Please continue. I just wanted to let you know.

Mark Anderson: That’s absolutely right. I mean, I think I’d close it out just to say what I don’t want. What I don’t want to do.

Frode Hegland: Andrew has to go to anything else. Oh, sorry.

Mark Anderson: By Andrew.

Frode Hegland: Any question or no. Okay. Wednesday. Good. All right.

Andrew Thompson: I think I’m good guys. Have a good one.

Frode Hegland: Yeah.

Mark Anderson: Because I think that’s another thing that’s been slightly sort of wobbling through this at the moment is we’ve got a quite prosaic use case which we’ve taken upon ourselves to do. And I think it just gets over complicated by trying to then take it into a very projected future, because it’s sort of. Not helpful because you’re going to have to explain each of these things that might be there in terms of a demo that’s actually going to be the overload. The key thing that I know is missing from the PDF that’s floating around at the moment is it doesn’t address at all any of the attributes of the paper, which is probably the key thing you want to know. And just one, and I’ll just throw one other thing into this that I think there’s a potential confusion in terms of what’s shown on might be shown is whether it’s shown as new nodes on a map on in the in the view spec, in in, in the space we see, or whether it’s basically data shown around the. Around the the paper because, I mean, one of the ways we deal with this is if you want to look into a paper, for instance, you could, as it were, drill into it, in which case you move into effectively, you replace the current view space with a different view space, which is the contents of that document, because that’s actually what you’re focused on in that moment.

Mark Anderson: That might be useful. I don’t say that. I don’t say that with any sense of whether this is a good thing or bad thing, but just the kind of thing that we can be done. But in the case of a new set of proceedings, most of the stuff that you have, most of the stuff we’ve been looking at and talking at, makes far more sense in, in the sense of a library, all these other authors, because I just because I’ve spent I have spent years working on this corpus for this. And the one thing I can tell you is that it most of it is, is just alien or single instance stuff. And so it’s quite difficult if we start pulling things up unless you’ve got filters said, you know, unless this author has more than five papers or something, you know, when we show the authors, we don’t see it. So it’s actually when you when you actually interact with the data, you realize, you realize some of the limitations that you face even before you get to the visualization. Yeah, I’ll leave it there.

Frode Hegland: Fabian, is it okay if I reply briefly to Marc? Thank you. So, Mark, that I believe is a design challenge. Yes. We need to have some sort of a focus or let me read this document properly without leaving everything behind. Absolutely crucial. That should be able to be done as a pretend PDF or paragraphs. This is very important for us to design and also what you’re working with Adam on timeline views to be able to integrate or leave the map in order to see historical data will be technically important. But even more important, I completely agree will be the interaction so you don’t get completely overloaded. But these are the things we need to really look at going through stuff. How does it work? So yeah, I think we all agree. Fabian, please. Otherwise, I let Leon’s chest talk.

Fabien Benetou: So I have. I have a minute left. But I just want to tease you all that. I think there is a very important paper published recently on VR which basically zaps your brain. So it’s used a little bit of electricity through your skull. And it does it just like on top, on the right way. And it does things like this where it makes your hand or legs move. So it looks a bit silly, but the thing is, if it does it just the right way, at the right time, for example, if you put your feet down and you put it up, you feel like you’re stomping on something. It’s a little bit crude and expensive and bulky at the moment, but I think it’s game changing in the sense that for now, so far, we had like haptic gloves and things like this. But if you start to imagine your entire body covered with haptic stuff like nobody, even if you’re super dedicated and want to try stuff, it’s just too much. So I think this I think it’s kind of a crazy thing. And I think the, the kind of dreams about VR, when you can have all your senses covered. To me, it’s yeah, I think it’s a game changer. And I that’s why I mentioned also on slack that if those people from Chicago University can come or I’d be tempted to go there if I can try this in Europe, of course I’d rather do that, to be honest. But but it’s it’s it looks crazy to me, but crazy in a good sense that I, it looks realistic also. It’s not I cannot imagine it’s bullshit. It might be bad in the sense like being super coarse and again, super bulky. But if we imagine if we project a little bit in the future. I find this fascinating. Not just look at stuff and hear stuff, but feel it.

Frode Hegland: Is this what Robert Scoble posted by Greg Mattison.

Fabien Benetou: Don’t think so. I remember the lead author was a Japanese student in University of Chicago. Can you give us a published for Chi 2024? So I think last week.

Frode Hegland: So maybe.

Fabien Benetou: Let me check.

Frode Hegland: So don’t don’t completely scramble our brains here, but Okay, so just to be clear, now that this just the four of us. We have a couple of agendas. Number one agenda is someone puts on the headset, does stuff, comes out of it thinking, wow, I can work in there. That’s our overriding agenda, right? Now we plan to do that is by giving them a limited set of information which is of interest to them. We talked initially about that being their own documents, which will be technically possible, but in order to make an effective demo, we’re doing a demo about the location and event that they’re at. This is something that I learned that I’ll never forget. When I showed I did a globe of the world, showed it to my mother in law, and all she wanted to do show me where she grew up. You know, when the data matters, it matters. So we want them to be in there and not look at a list of papers that are whatever it has to be, something that they’re already thinking about. So. Oh, there’s my friend. Or oh what’s that. Right. So that’s so we have completely agreed. We start with a list. No question. And I completely agree with Danny that we should have accessible different view specs. Not the term she used, but that’s what it is right. So we have to design that and it will be very very important. That we are able to do that in a way that connects with what Adam Mark is doing, what Fabian is doing, what Leon hopefully will be doing if he has the time. But the default thing really has to show off the native resolution of XR, which is of course 360 rights, to use a strange term. So that really means that it needs to be simple.

Frode Hegland: But as they move things about, it needs to get richer and richer. Now, I think Dini and I use the term link very differently. Today she’s more talking about click through to go somewhere links. When I talked about connections in the document, I meant either by selecting and seeing lines or computationally or by criteria. Lines appear and we should support that. We should maybe redefine spatial hypertext in this environment. Maybe not. Maybe it’s not the right kind of thing, but that’s why it’s so important. Mark, you and me are in this sense very different that we both have ideas for how we want to go through it and see how those ideas work together. And to me, I half jokingly say it should be possible to have an LLM sitting in the corner commenting on the corpus you have in front of you. That is kind of code for saying Fabian World, right? With Fabian’s computational world that can be Lego together. To have that in and out of this space will be, I think, very powerful and absolutely no idea how it will be useful. One more thing on this. I believe that we’re going to be using AI a lot, but in an incredibly constrained way, where basically we’re going to go through all the documents and extract keywords and us as humans, before we release this, we’ll go through and edit those keywords. If something looks bullshit, we just delete them. But for now, they can help us build things. Because things like if you want to look at papers with history, they may not be labeled history, but they may prefer to Ted Nelson and therefore it’s a useful thing. Before Mark finishes this, anybody have any comments on just that bit?

Fabien Benetou: Oh yeah, I’ve already done a couple of demos where you pipe in some code snippets sent to LM, get the results, the results back. So that part. Is not hard. It’s been. I mean, it’s been done. I’ve done it.

Frode Hegland: I was replying to Leon here and text. Can you please repeat? Fabian, I’m sorry I was rude and I didn’t hear a word.

Fabien Benetou: No worries. No, it’s just that I’m just putting the example right now in the chat. But going from a code snippet in XOR to an LM and get the result back, I’ve done it before, so switching to another LM, use another prompt, whatever. Not a problem.

Frode Hegland: That’s cool. I’ve been using GPT four zero, and I found that it doesn’t have the intelligence of Claude at all. I can’t use them for a meeting summaries, but the speed is amazing. And to do basic things like extracting based on criteria and putting a specific ways, it’s almost real time. So it’s great that we have that opportunity to experiment, to see which one works for what.

Fabien Benetou: If it’s useful and if it can also prevent some of the recurring conversation, I will make it generalized so that you say this LM on input, this prompt on output or in input. Also actually this snippet like the source of text to send it. And for example your how do you say your token. Because maybe it might be an LM others don’t have access to. And I think then letting people play should be should be sufficient in most cases.

Frode Hegland: I imagine generalized is good. But it’s also important, as Peter says, to record what LM was used. What version? All of that stuff. I agree.

Fabien Benetou: I agree, I even interrupt you there because this specifically what I told the colleagues a couple of weeks ago is like, if there is no provenance, I can believe you’ve done it. But if I want to tinker with it and try to even reproduce and without even extending it, I can’t. So I 100%.

Frode Hegland: Cool. Mark, you get the final words. Believe it or not. Okay, well

Mark Anderson: I’ll start with the end. I mean, I really, actually, I don’t have anything at all against you use the sort of I in this sense, a especially in a sort of, in a close sense, but I am conscious that it rather it’s pulling away from the stuff we keep talking about. I, we don’t talk about the actual thing we’re doing. In fact, the one thing we haven’t talked about is reading, you know, is understanding papers. And I’m sorry, I don’t think you understand understand a paper by just getting it summarized by something else, because you can’t be bothered to read the paper, you need to interact with it. And we can’t get around that. It may be boring, but that’s the task we took upon ourselves to do, and I think we should just be mindful of that, because bear in mind that, you know, you’re going to be putting it on the heads of some academics and that we’ve got to think about what they like and want and not necessarily the excitement, which is why I said I see in two slightly different camps, the all the what, what ifs or what might be is that you’re mentioning, I think, a very exciting and really interesting, very pertinent to this display medium.

Mark Anderson: But I’d be very careful about trying to don’t try and create excitement where there doesn’t need to be any, because that is not going that is not going to attract the interest of the person you’re trying to show in the scenario. I think that’s the one fundamental mistake we can make. There are lots of other things that we can do and show. And as you said, if we can link out to the other things and for instance, some of the stuff that Fabienne’s done, some of the Leon has done thing so that’s all still there available. And I go back to my point. I just see them as different view specs on the same data. So if you keep that separation in mind that the, the data and the view, well, they’re almost three parts of it. There’s the thing you actually see the sort of view spec that describes what you see. And then the metadata the world is our oyster. But at the moment I think we keep wandering away from the thing we said we’d do, which is I’m no quite prosaic, but mainly involves actually looking at what’s in the paper, not what something else thinks is in the paper.

Frode Hegland: Okay, I don’t agree. But but.

Mark Anderson: Okay. All I would say is you keep saying you’re not. Not an academic. And there are two people here who are who keep trying to say to you, this is what they’re after and you’re not listening.

Frode Hegland: Please say why I don’t agree. Okay. I have a PDF reader for Mac, iPad, iPhone and visionOS. Right. It is something that I think is really, really important. I’ve got a new user today who was a learning disability and he would like a negative view in PDF. See you later on. He has a learning disability, so he wants a negative view, which of course you can’t really do with a PDF. So I’m redesigning our floating view to support him. Right. I completely agree 100% that reading the document will be very important. I am not at all endorsing a summary based system. Not at all.

Mark Anderson: By reading, I don’t mean reading soup to nuts. I mean actually engaging with the text. And this is the one thing we haven’t really touched on at all. And that’s what concerns me in the.

Frode Hegland: Please let me finish. The thing is, in my experience so far, reading on an iPad, on a computer screen, or on paper is at least as good as reading in XR. Okay, it is nice. The thing is, a single document. If you are serious about it, you will read basically soup to nuts. And I completely agree with you that we should be able to deconstruct it. Your term. I think it’s perfect. The reason I am focused on the proceedings is because this is an XR project. So how can we best use that space? I don’t think that space can be best used to read one document that you really could have on your desk. It is to have all these documents together and see how they relate.

Mark Anderson: But I didn’t say that. You see, I didn’t say you’re going to read one document. What I was trying to explain is that, I mean, surely you’ve read, you’ve had you’ve gone through a corpus and you’ve looked at, you know, you’ve taken a view and you worked through there’s basically a triage process. A reading is perhaps a wrong term. It’s a sort of triage. And so what you’re actually doing is you’re looking you’re basically busy unpacking the richness that’s in the document, and you’re trying to assess its relevance. And it’s probably some way before you’re even making a judgment as to whether you like it or not. You’re basically trying to work out what it is. And, you know, is it is it what is it, what it says in the in the sort of broad title, it’s broad title probably sucks you in the first place. You then look at what’s inside and we haven’t addressed that yet. And, and all these other fun things we can do don’t actually address that either, because and so it’s not a matter of sort of pulling the thing up and reading it on a virtual screen. There’s emphatically no what I’m saying that it’s it’s about the ability to decompose in a way that you can’t do on paper or in a PDF because it’s it’s typeset into a fixed thing that you can pull together. So you can you can pull out a section of the paper or something, you can pull out a strand on something. And this is, for instance, where things like close scoped. I probably will help because if you think the paper is about subject X and subject Y, you might ask the AI to give you. It’s a sort of summarize which parts address subject X and which address subject y. So that sort of thing is really pertinent, but we still need to not lose sight of the fact that the task at hand is to be able to understand what’s in a proceedings and, and, and the nature of the paper. And that inherently involves a drill down of their structure. Now, when I say drill down, I don’t mean.

Frode Hegland: Okay, Mark, Mark, Mark. Okay, look, we can go on about this quite literally forever, okay? There are many issues here that I think are really will.

Mark Anderson: You will you will you please will you please put that into the mix because that that isn’t there at the moment. It’s not there. It’s not in your PDF document. It’s not in what you’ve been showing. And it is a concern because it’s what we said we’re going to do. But we’re, we’re we keep getting lost in the fun stuff, which is great and which I enjoy and I’m interested in as well. But I’m conscious that we set out, we said we’d do something. It might seem boring, but we should.

Frode Hegland: Do it well. Specifically. So I don’t misunderstand you.

Mark Anderson: Right. So. Well, for instance, I mean, one of the things I didn’t put in the PDF, sorry.

Frode Hegland: Just briefly to put what down to just list it, please.

Mark Anderson: So, so. Right. I’m sorry, when I was annotating the PDF, I didn’t put more stuff in because I frankly, I didn’t really have time, but I was just trying to address what was there. What was emphatically missing was there was nothing to do with the essentially the the reading of the right word, the contents of the structure of the paper. So there is some there are some internal elements of the paper which are absolutely key things. So if you’re asking, it’s going to be on screen.

Frode Hegland: I know what you mean I know what you mean. Thank you. Okay. Right. So the problem with our community, as I’ve said before, is people don’t write anything. Generally, you know, you have written things, some people have written things, but in general they don’t write anything. The. Requests that I’ve had now for a couple of weeks that I put in this document as well, is if you’re on a huge map in XR, you’re going through a proceeding. What are the kind of things you want to see him do? Right. I gave some ideas that I think should be there. The additional bit of looking inside the paper is kind of a focus view. It is on their focus, right?

Mark Anderson: No it isn’t, it isn’t. I didn’t get that from the doctor, I genuinely didn’t. Okay. But because the document’s not structured the document keeps bouncing around. So one of the problems with the document, it’s really hard to read because it doesn’t go into logic. Okay.

Frode Hegland: It’s hard for you okay. It’s hard for you to read. But you’re putting that the what I’m getting from you and Denny, you’re putting the entire onus on me to making this great document all of the right things. But you won’t tell me what you want the things to be. No.

Mark Anderson: Well, we have, we have. And you keep saying. You’re saying I agree, and then you go and talk about something else. So it’s this is why there’s confusion.

Speaker7: Okay, Mark.

Frode Hegland: Look, the one thing I’m interested in here right now is someone’s in a webXR view, and they have a list of papers in front of them. What are the things they should be allowed to do? Okay. If you want to write that down, that would be absolutely fantastic.

Speaker7: Okay. I mean.

Frode Hegland: We I’m not talking to to have a list view that I have. Actually, two weeks ago or one week ago, I can’t remember. I did show an outline mode where you could drill down and collapse and expand. I’ve done that so you could see the abstract. You could see more. I’ve gone through many, many iterations of that. It is, of course, important to access the content of the document. But one thing we need to do, so we’re not just doing something that will work on a 27 inch screen, is to use the space. To use the space means things in many different places. One thing that you suggested today that was important is to be able to take a paragraph and move it out of the document, put it in the space. That’s actually really quite difficult to do technically. I mean.

Mark Anderson: I understand that, but we can. The point is we we know it’s technically difficult to do, but but we know in a sense it is it is possible even if it involves, you know, structuring documents in a different way. But that actually advances the use case that we’ve set out to do. The other things are fun and exciting, but they don’t actually help, particularly. So you don’t.

Frode Hegland: Think having okay, what does not help.

Mark Anderson: But because. I mean, I just I get confused because, I mean, do do you have a real sense of what it’s like to go through a set of papers and, and sort of triage them and read them for relevance, bearing in mind that a proceedings is not of a type, that the main bounding thing is they’re all about hypertext. But even that’s very, very broad. So it may be that almost no two papers are on the same subject. So this idea of.

Speaker7: Instead of me.

Frode Hegland: Answering that question, I have a question for you. Yeah. What do you think we for this project should use a wide, expansive, webXR view to show the user.

Mark Anderson: Our exploration of the papers so I can go into the papers. I can I can start to basically find things in it that are of note. So rather than show things and sort of pre-decided things about it which which seem a bit rich because we don’t know what they are until we see them, especially with new papers. We know what the papers have said about themselves. We can certainly use that. So we could take the aggregate set of for instance, the author, author keywords for the papers. And potentially show those. But my, from the analysis I’ve done, what you’re going to see is you’re going to see a number of, of keywords that basically point to everything, and then a number of keywords that just point to one paper. So although we might be able to put them on screen, I, I’d almost wager you’ll find that it’ll be less actually, it’s frustrating. It ought not to be the case, but it is now, if you were dealing with a mature corpus where people had gone in and actually added metadata tags, whatever keywords, that’s a whole that’s a whole different ball game, but that’s not what you’re going to get with a set of new papers. A new set of papers are unknowns. And you basically have to you have to drill your way in from the

Speaker7: If you have.

Mark Anderson: One thing you could add.

Frode Hegland: To extractions done on them.

Speaker7: Sorry.

Frode Hegland: What if you have entity extractions done on them?

Mark Anderson: Well, what do you think? It’s what? So what? Give me an example of the sort of thing that’s going to tell you.

Speaker7: Panics are just one example.

Mark Anderson: Come on.

Frode Hegland: I can give you many an A list.

Mark Anderson: No, just give me an example. An actual real example of something.

Speaker7: I’ll give you a few real examples.

Frode Hegland: And this applies to our ACM hypertext last year. You know, easily see who from my university has written papers. There may be more relevant. And when it comes to keywords, keywords that I find are interesting are very often not included by the authors themselves. And for instance, I’m interested more in straight text interaction. So I’m interested in Doug Engelbart and Ted Nelson a lot more than of like things like the Bibliotecii Herzegovina did about just street maps and things. So if I could have my specific keywords of the people that I care about to see if they’ve been written about, that is a useful thing to me immediately, because they would show that that’s one example. Another example is to connect it with the previous citations that you guys have been working on, to see if are any of the authors listed this year, somebody who was entirely new to hypertext? That’s interesting. Maybe they’re a student. I want to spend a specific amount of time on that. If they have written many papers before that are highly cited, that is also a very interesting criteria for me. So there are many ways that me individually, who don’t want to go through all the papers that I believe have obscure titles, what I find useful.

Mark Anderson: Right. Well, that’s a lot more tractable than what’s offered the PDF. So the things that you’ve suggested there to me, if I say, for instance, that. So we’re having a presumption that we’ll have done an entity extraction on presumably on a number of thematic things I have, I have my doubts about some of the organization type ones, but, you know, names are fine. So another would be to run an analysis of. So you take the author list and you’re going to have to take best guess because of course people people and often if they’re non-English names, their names bounce around a lot because it’s down to the first author often how they list them. And so this is one of the things I’m making.

Speaker7: An assumption.

Frode Hegland: Like you, we are for this work, making an assumption that we have good metadata. This is a demonstration of what can be done if we have that. So we don’t need to go into the issues with names. Those are known understood issues.

Speaker7: Right.

Mark Anderson: So what what would be much more useful than us looking at, say, pictures of North map is to take some of these ideas that you have and actually show them, show them what they what that would look like with, with real examples because I think that’s that’s where it’s, it’s, it’s just flying past at the moment.

Frode Hegland: But Mark I’ve done that.

Speaker7: That’s what this is.

Mark Anderson: Well, I don’t see it. I don’t see it on the screen.

Frode Hegland: Okay.

Speaker7: It’s quite.

Frode Hegland: Literally this. Right now, I have a problem with some of the computational error that these links don’t always work in both directions. But here’s an example.

Speaker7: I want to know.

Frode Hegland: First of all, who’s interested in Doug considering this is ACM hypertext, the fact that it’s one paper.

Speaker7: It’s kind of fascinating. Well, what was the.

Mark Anderson: Okay, so the immediate question arises is what what what was the locus of citation of Doug in that paper? How do you how do I get to that from that link? From that?

Speaker7: Exactly.

Mark Anderson: His name alone is not meaningful.

Speaker7: Because I’m.

Frode Hegland: Trying to answer you. That is exactly what I think we should be working on. Write to find out. When you see a relationship, how can you drill down? How can you see the relationships? One of the things we have in our meetings. Yeah. We don’t need to go into that. One of the things we have in our meetings is. So our computer went funny is a list of extracted names from our meetings, and I’ve asked the LM to tell me who said it and what was the context. So, for instance, now in our summary by Claude, we’re going to get a sentence that says Doug Engelbart mentioned by Frode Hegland.

Speaker7: And here’s the sentence.

Mark Anderson: Have you tried that? Okay. That’s interesting. So what? So really what I’m getting to is that again, what I’ve, what I have experienced doing this in a sense the, you know, just manually, broadly because that’s what we have in the past is that Mention alone is not meaningful. So so the the the unintentional problem you have at the point of just entity extraction is you have a load of entities, but you don’t actually know their relevance. So the to have to be meaningful, you need to actually know the context in which they occurred. Yeah. So I could have just mentioned Frode Hegland in passing in a paper. And in fact Frode Hegland is not related to the paper at all.

Frode Hegland: Mark, this is what’s important. You’re completely right. I’m completely in agreement. When you interact with this system, we, you and me, need to. And everyone else who wants to need to deal with exactly that. When you come across Doug Engelbart and you see his listing of some things, one of the things should be able to see what was the sentence or the paragraph. Lift that up. Okay?

Speaker7: Right so.

Mark Anderson: Away that.

Speaker7: That is going to.

Mark Anderson: Say.

Speaker7: Yeah go ahead.

Frode Hegland: No, no it’s okay. Please.

Mark Anderson: No, I mean to try and help you. So what you’re why your PDF could have been much more helpful and got better comment was if you’d done something like that and said, right. So the kind of things that we could get hold of so and think things you know, you’ve done. So I can get hold of a list of the names. But I understand the names alone are not are not, are not essentially useful because I don’t know the context. So I don’t know if someone’s just being mentioned or whether they’re actually someone is asserting an interest or a use of that person’s work.

Speaker7: I’m going to stop.

Frode Hegland: Recording now unless there’s anything else you want on the record.

Speaker7: Anything else you want on.

Frode Hegland: The record before we.

Speaker7: Stop? No, no, no.

Mark Anderson: That’s fine. I mean, I’m.

Chat Log:

16:06:59 From Hrithik’s OtterPilot  To  Frode Hegland(privately) : Hey Frode Hegland, want to skip having to click ‘allow recording’ for every meeting? Connect Zoom to your Otter account: https://zoom.us/oauth/authorize?client_id=ULKDU8_8TLySSz8ZMyO9WQ&redirect_uri=https%3A%2F%2Fotter.ai%2Fsetting%2Fconnected_apps%3Fsource%3Dzoom&response_type=code&state=dXRtX3NvdXJjZT16b29tX2hvc3RfdXNlcg==

16:08:22 From Fabien Benetou : https://spike.legoeducation.com/essential/lobby/

16:08:45 From Fabien Benetou : and kit itself https://www.lego.com/en-us/product/lego-education-spike-essential-set-45345

16:09:52 From Peter Wasilko : Good morning from brunch surrounded by food and books to cite for the paper.

16:10:00 From Gökhan’s OtterPilot : Hi, I’m an AI assistant helping Gökhan Turhan take notes for this meeting. Follow along the transcript here:  https://otter.ai/u/v9LZUo8j0wF81rOCD-xRjwrJyjY?utm_source=va_chat_link_1

You’ll also be able to see screenshots of key moments, add highlights, comments, or action items to anything being said, and get an automatic summary after the meeting.

16:10:00 From Hrithik’s OtterPilot : Hi, I’m an AI assistant helping Hrithik Tiwari take notes for this meeting. Follow along the transcript here:  https://otter.ai/u/eVGbpGtAeCMFuoh-YvCGTs6-8Wg?utm_source=va_chat_link_1

You’ll also be able to see screenshots of key moments, add highlights, comments, or action items to anything being said, and get an automatic summary after the meeting.

16:10:30 From Andrew Thompson : Replying to “and kit itself https…”

I’ve been a bit out-of-date with LEGO tech, but this reminds me of the old Mindstorms sets. I loved those as a kid.

16:11:00 From Fabien Benetou : Replying to “and kit itself htt…”

It’s the same principle BUT with Scratch so much easier onboarding IMHO

16:11:22 From Andrew Thompson : Replying to “and kit itself https…”

No longer their custom script? That’s cool.

16:11:40 From Mark Anderson : Dene posted this before some joined … ‘Google declares the end of the World Wide Web’:  https://unherd.com/newsroom/google-declares-the-end-of-the-world-wide-web/

16:12:27 From Frode Hegland : Here he is: https://x.com/Volorf

16:13:46 From Frode Hegland : https://www.olegfrolov.design

16:15:10 From Fabien Benetou : Reacted to “https://www.olegfr…” with 👌

16:17:48 From Fabien Benetou : Replying to “and kit itself htt…”

You can still do something quite custom via e.g https://www.raspberrypi.com/news/raspberry-pi-build-hat-lego-education/ which led me to https://x.com/utopiah/status/1761311123712741613

16:18:06 From Fabien Benetou : Replying to “and kit itself htt…”

So Scratch, Web based is “JUST” onboarding, not a requirement

16:19:27 From Mark Anderson : Suspect much of VM going forward—being data—will be JSON or JSON-LD.  (Not that it is problematic!)

16:20:49 From Mark Anderson : Leon had a hand up!

16:21:36 From Mark Anderson : Wayback Machine FTW!

16:23:19 From Fabien Benetou : Can we just toss the dude to Mars?

16:23:49 From Frode Hegland : Reacted to “Can we just toss the…” with 😂

16:23:59 From Fabien Benetou : Let’s dump him ASAP.

16:24:44 From Rob Swigart : PGE cut off my power and I’m having trouble getting connected.

16:24:44 From Mark Anderson : ‘Locks’ Lots of copies kept safe

16:25:28 From Mark Anderson : Reacted to “Screenshot_20240520_172456.png” with 😀

16:28:27 From Peter Wasilko : Reacted to “Suspect much of VM g…” with 🚀

16:28:28 From Mark Anderson : Now thinking of Fahrenheit 451 – we all need to remember a number of websites.

16:29:05 From Dene Grigar : Did you know that F451 is a banned book in the US?

16:29:11 From Fabien Benetou : Reacted to “Did you know that …” with 🔥

16:29:27 From Fabien Benetou : Replying to “Did you know that …”

The entire US or just some states

16:29:33 From Rob Swigart : Right you are. It’s weird since I can’t get video at all. Trying again.

16:29:40 From Fabien Benetou : Reacted to “Right you are. It…” with 👋

16:30:55 From Mark Anderson : Reacted to “Did you know that F4…” with 😱

16:31:22 From Fabien Benetou : Reacted to “Did you know that …” with 😱

16:31:23 From Fabien Benetou : Removed a 😱 from “Did you know that …”

16:31:32 From Peter Wasilko : https://dotapp.uk

16:32:19 From Dene Grigar : Replying to “Did you know that F4…”

Various states:

16:32:48 From Mark Anderson : Is there any indications as to any particular aspect of the ACM HTML that is upsetting the parsers?

16:34:01 From Mark Anderson : Is this for HUMAN’24 or HT’24?

16:34:58 From Frode Hegland : Now: 1) What to put on a Map.  2) Fabien demo.

16:35:34 From Rob Swigart : Unable to connect sound or video, so listening. Battery backup sort of works.

16:36:01 From Fabien Benetou : https://rclone.org

16:37:08 From Frode Hegland : Reacted to “Unable to connect so…” with ❤️

16:37:19 From Frode Hegland : Grateful you are persistent Rob!

16:37:25 From Fabien Benetou : FWIW WebXR demo using rclone https://x.com/utopiah/status/1255141803621117953

16:39:11 From Fabien Benetou : also https://rclone.org/commands/rclone_rcd/

16:39:21 From Mark Anderson : I’ll be discussing rclone with my PSDI colleagues (PSDI = Physical Sciences Data Infrastructure. UK funding scope but discussing with similar efforts in De and US.

16:41:21 From Mark Anderson : Noting the fact that scanning new papers is as much triage  as reading.

16:42:04 From Leon van Kammen : oh that is so good

16:42:11 From Leon van Kammen : already in 2020

16:42:16 From Leon van Kammen : pioneer

16:42:34 From Leon van Kammen : (related to Fabien’s 3D filebrowser)

16:43:42 From Fabien Benetou : hard to believe https://github.com/Utopiah/vrify is nearly a decade old… damn :/ things move so fast and so slow at the same time!

16:43:43 From Mark Anderson : and Mark Bernstein’s still inovating. This view does even have a name yet : https://www.markbernstein.org/May24/ANewView.html (looks like a voronoi map, but is more than that)

16:45:27 From Mark Anderson : Wirth maps less (fewer notes) is more for sense/knowledge making. Something we should bear in mind.

Ergo how do we choose what goes into the view at start.

16:46:29 From Mark Anderson : I think splitting the nodes in the view/viz from the metadata is key to flexibility.

16:48:23 From Mark Anderson : How to navigate scope – expend in-view or move to a new view. E.g. having alighted on a paper in a set, wanting to look at the internals—abstract, title conclusions, etc.

16:48:25 From Peter Wasilko : I love the new TB view, has it landed in a beta yet?

16:49:35 From Mark Anderson : Replying to “I love the new TB vi…”

Not yet but MB said RSN.

16:53:39 From Dene Grigar : I mentioned to you, Frode, that we should describe the “attributes of connections” beyond color/weight and “when they appear” thinking instead how they function/why they are being used

16:54:09 From Frode Hegland : Replying to “I mentioned to you, …”

Yes, we need to discuss this now.

16:54:33 From Dene Grigar : Jeff Parker’s A Poetics of the Link gives us the category of “functional links”, one of which is “blatant link”

16:57:13 From Mark Anderson : Replying to “Jeff Parker’s A Poet…”

Yes!

16:58:03 From Mark Anderson : An important part of attributes of links is the possibility of link types.

16:58:33 From Ingrid : Hello everyone, happy to join to say hi, thanks for the invite Frode 🙂

16:59:11 From Peter Wasilko : Welcome to our community, Ingrid!

17:00:04 From Ingrid : Reacted to “Welcome to our commu…” with 👋

17:00:09 From Frode Hegland : To discuss now: Functional Links.

17:00:23 From Brandel Zachernuk : Sorry I have to drop – I _believe_ that I should have more time for this very soon! Good to see everyone if only briefly!

17:01:14 From Frode Hegland : Dene, can you define functional links here in the chat so we are ready to discuss?

17:01:53 From Dene Grigar : functional linksA Blatant LinkA Filler LinkA Random link.

17:03:09 From Dene Grigar : A Blatant Link tells the reader exactly what information will be revealed when activated.

17:04:44 From Dene Grigar : Filler Link:  the author wants to provide a functional link for the reader to the information necessary to contextualize the text at hand, to fill in the gap in linearity.

17:07:19 From Peter Wasilko : We used to do that with virtual pads in MediaMOO back in the day, albeit in purely text-based VR.

17:07:36 From Dene Grigar : yes, Peter. MOO tech is very superior

17:07:48 From Peter Wasilko : Reacted to “yes, Peter. MOO tech…” with 🚀

17:07:55 From Dene Grigar : Here is the URL to Parker’s essay: http://www.altx.com/ebr/ebr12/park/park.htm

17:08:26 From Dene Grigar : He worked on this project during Kate Hayley’s NEH Summer Seminar in 2001. It was his required seminar paper

17:13:24 From Leon van Kammen : when Dene talked about links with color/weight..I was thinking..is there such thing as an anti-link? (this is absolutely the opposite of that…Freud vs Jung..are connected no?)

17:13:32 From Dene Grigar : Ingrid, will you introduce yourself? We’d love to know more about you

17:13:56 From Fabien Benetou : Reacted to “Ingrid, will you i…” with 👍

17:14:16 From Dene Grigar : Leon, Parker talks about a “complicating link”

17:14:42 From Dene Grigar : He says “It should begin to raise concerns and tension”

17:14:50 From Leon van Kammen : thanks, will read up on it

17:14:56 From Fabien Benetou : Replying to “when Dene talked a…”

not sure if that’s what you mean but links can have values and it can be a negative

17:14:56 From Peter Wasilko : Replying to “I love the new TB vi…”

I might call it Cellular View

17:15:17 From Fabien Benetou : Replying to “when Dene talked a…”

(in graph theory sense)

17:15:37 From Dene Grigar : He also invites us to imagine other types. One thing that many of us in hypertext agree about is that there is still much left to tease out in terms of linking.

17:16:33 From Leon van Kammen : which Parker? (there are so many called Parker I’m learning right now)

17:17:01 From Dene Grigar : Jeff Parker

17:17:12 From Dene Grigar : http://www.altx.com/ebr/ebr12/park/park.htm

17:17:15 From Leon van Kammen : ❤️👍

17:19:09 From Frode Hegland : https://dougengelbart.org/content/view/218/

17:20:18 From Peter Wasilko : Reacted to “https://dougengelbar…” with ❤️

17:20:49 From Dene Grigar : this is something we can test for

17:21:21 From Dene Grigar : We don’t have to provide a final environment at ACM HT, we can use the conference to gather information

17:23:01 From Dene Grigar : I just want to see the primary list of references

17:23:23 From Dene Grigar : Then I want to explore the various expressions that can be derived of it

17:26:25 From Andrew Thompson : The current ACM tags are not very useful, yes, but I think we plan to generate our own tags through GPT or something similar (through reader right now) that will be more usable

17:27:08 From Ingrid : Reacted to “Ingrid, will you int…” with 👍

17:30:39 From Peter Wasilko  To  Frode Hegland(privately) : Lets let Ingrid talk next, she has her hand up.

17:37:44 From Peter Wasilko : Replying to “The current ACM tags…”

It seems that they are either way too common or nonces.

17:38:10 From Mark Anderson : FWIW, the HT’24 and Human’24 have the same deadlines.

17:38:16 From Andrew Thompson : Replying to “The current ACM tags…”

Yes, that was my impression as well.

17:38:20 From Mark Anderson : Is the team paper on Overleaf?

17:38:32 From Peter Wasilko : Replying to “Is the team paper on…”

In June, right?

17:39:07 From Peter Wasilko : Replying to “Is the team paper on…”

(For Human that is)

17:39:21 From Ingrid : I’m sorry I have to drop to catch my train. Hope you all have a lovely day/evening!

17:39:42 From Mark Anderson : Yes, but note the June date is an extension – there won’t be another one!

17:43:32 From Rob Swigart : I will be MIA until June 12, though available in email. Now have another appointment. Key discussion today.

17:43:43 From Frode Hegland : Reacted to “I will be MIA until …” with ❤️

17:43:54 From Dene Grigar : Replying to “I will be MIA until …”

bye, sail well

17:46:39 From Dene Grigar : Now that I know that the pdf you sent out contains the very items in the menu, I can go back through it and offer a suggestion.

17:47:11 From Andrew Thompson : Essentially: I can still work on the project, take your time as a group to develop the view solution.

17:47:32 From Dene Grigar : Reacted to “Essentially: I can s…” with 👍

17:47:36 From Leon van Kammen : Fabien..you’re on

17:49:49 From Frode Hegland : Reacted to “Essentially: I can s…” with 🔥

17:51:19 From Frode Hegland : Fire!

17:52:02 From Leon van Kammen : @Andrew Thompson rclone is a quick way to test with webdav locally too

17:52:02 From Leon van Kammen : [This is an encrypted message]

17:52:32 From Andrew Thompson : Reacted to “@Andrew Thompson rcl…” with 🙌

17:53:07 From Dene Grigar : I’d like to set up a meeting with you, Fabien. Are you able to visit with Frode and me on Friday?

17:53:30 From Fabien Benetou : Happy to, just need to find the moment that works with timezones

17:53:30 From Dene Grigar : I need to get to my lab’s meeting. I will be leaving at 9:55

17:53:45 From Frode Hegland : We need to dream and fantasise on this: even put LLMs in there?

17:54:06 From Dene Grigar : Would 8 am PDT work, as today’s meeting starts?

17:54:35 From Fabien Benetou : yes

17:55:26 From Peter Wasilko : https://www.linkedin.com/posts/zuza-sliwinska-661076130_geospatial-googlemaps-googleio-ugcPost-7198310907058278400-ZPvi

17:55:26 From Andrew Thompson : I also need to run, take care everyone

17:55:34 From Peter Wasilko : I have to head out too

17:55:48 From Frode Hegland : Bye Peter

18:00:12 From Mark Anderson : Interesting!

18:01:14 From Frode Hegland : https://x.com/Scobleizer/status/1792000171653689710

18:03:33 From Fabien Benetou : paper https://lab.plopes.org/#source-effector with TMS

18:04:32 From Leon van Kammen : The sushi is waiting for me, I have to go now.

18:04:51 From Frode Hegland : Will the sushi get warm?!

18:04:53 From Frode Hegland : 🙂

18:05:01 From Frode Hegland : See you later Leon. It’s exciting

18:05:12 From Leon van Kammen : hehe

18:05:12 From Fabien Benetou : https://x.com/utopiah/status/1621760472461594624

18:05:18 From Leon van Kammen : bye!

18:05:23 From Mark Anderson : Bye