24 April 2024
Frode Hegland: Hello. How was your analog holiday? Well, you’re still beautifully muted. Come sit on my lap.
Mark Anderson: For some reason. My my hello. My my mouse had lost itself.
Frode Hegland: Well, we were just talking about computer mice. Yeah. Had a good break away from it all.
Mark Anderson: Lovely. Yep. Beautiful. Beautiful weather. Basically we’re out. We were out walking every day, so I’ve been delightfully detached for about a week, which is all to the good. Yes.
Frode Hegland: Absolutely. Where are you going?
Mark Anderson: I did come back, I came back, I’ve updated my headset. I’ve read. I have read down. I think through the slack. I’m probably up to date. I saw that Andrew has posted something new.
Frode Hegland: Yeah, it’s linked as well. Now. The So we now have a new series of recordings that I put into a PDF with Claude. Summary. Absolutely astounding. I mean, it’s not rocket science. It’s simple stuff.
Mark Anderson: I saw the overview. I saw the overview you posted. I mean, I don’t know. The thing is, I can’t remember that deeply. I mean, the thing I find hard about it is it’s so, so anodyne. Lee. Sort of. It’s sort of smooth, but it’s so generic. I don’t know if that’s actually really the true reflection or not. I mean, it, it it reads as perfectly good English whether it’s accurate or not. I have no idea.
Frode Hegland: But that’s the whole point of the PDF. So I sent it out to everyone. So if you after every meeting, I will update it. New document and maybe even Wednesday meetings because they’re a bit different. I will see. We’ll all have to agree or not. But the point is, if you see something not reflecting what you saw happen, tell me. You know, this doesn’t have to be pure. I, I’m very, very happy to edit it. For instance, Fabienne’s last name is variously misspelled for some reason. But what I found quite shocking was the summary of all the meetings, and it said, yes, the team has progressed in this and that way. And I found that was completely correct, that that was a bit beyond.
Mark Anderson: I suppose, what I’m saying and perhaps and I don’t and I don’t mean this over cynically, is that I think I could have written that without possibly attending the meetings, in the sense that it’s so generic that.
Frode Hegland: It’s not, though. Hang on a second. I hang on because it says things like shifting from broad discussion, discussion to implementation details the technical challenges in XR. Right?
Mark Anderson: That’s what I mean. It’s actually it’s sort of I was I was just reminded of this or something I was reading about this problem. Summarization is that, I mean. The difficulty is that it too readily gets jammed into a sense of, you know, whether one likes or doesn’t. I think that’s the wrong look at it. I was just just thinking on the fact that what’s hard to do is unless you’re actually really taking quite detailed notes or reviewing the whole thing again, actually remembering. I find that just for a second, I find the very generic summaries not actually very informative, because they’re so generic that I don’t actually know. It doesn’t really tell me anything.
Frode Hegland: Now I understand that perspective, but I don’t think that that generic because it depends on what we ask it to do. Right. So what I’ve done here is I’m going to paste this in because this is quite central to our work as well. One second. It’s so annoying. The chat window is hidden if the window isn’t huge. Right. So in the chat now I’ve pasted in the prompts. Oh, it was too long so I can. Right. So this is the first part of the prompt, kind of the intro. But what was kind of, I think amazing is that. What the hell? Sorry. It’s just so I’m not liking. Yeah. What I find kind of amazing is I can do these multiple tasks in one go. And that saves me, obviously, a ton of time. But I also find it interesting that for instance, what was discussed regarding visual media a couple of times. He said there wasn’t really any discussion on visual media at this meeting. That in itself is useful.
Mark Anderson: And is that your. Is that your recollection? Again, don’t don’t mistake this as a sort of a transcript.
Frode Hegland: I have the transcript. Yeah. Based on the transcripts, I can search the transcript and we have the video. So if that’s what I’m saying, if each one of us, if we feel something was missing or something is wrong from your memory or from your notes, just tell me and I’ll edit it in so far from my memory. Yes. Okay.
Mark Anderson: That’s what I’m. I mean, the point is, because otherwise, if we’re having to go, if we’re having to go back and mark something that’s being done for us automatically, it’s not really being done automatically because we’re having to do the work twice over. That’s really what’s driving it.
Frode Hegland: Yeah. No, I know you have a different level of enthusiasm. But I mean, no, no.
Mark Anderson: You see that that misrepresents. It’s not. I just I just have reasonable doubt over accuracy because I do understand that it unintentionally. It’s it’s doing something that sort of tricks us. So if we get something that’s very readable, we tend to assume it’s sort of correct because we’re looking for errors. And the the thing is, if it’s well, confected, we slide right past the things we half remember. And none of this particularly matters.
Frode Hegland: But Mark, I don’t I don’t agree. Because if you look at the most recent meeting, for instance. Right. First thing, it lists who was there. That’s a simple thing. And it’s easy. And it’s done nicely. General summary may be good or not, but then we discussed a year or two ago using the word important and so on. It does a pretty good sentiment analysis. I think for these pretty good. And if someone feels different, I’ll change it. But what is really interesting is if you look down here, for instance, the names mentioned, that is really useful for the future to remember what sessions Bill Buxton was mentioned in. So he’s not here as a keyword. He’s here as a name. And, you know, we did not discuss this. We discussed timeline, you know. Yeah, we discussed on this because of the what I think is quite amazing, being able to do this, have all these different prompts in one. And what I would really like is if you or anyone else has an idea for a different prompt to kind of optimize this, or even to test the system, I would love to add it. Because the whole. Yeah. Peter.
Peter Wasilko: Yeah. I was wondering, is it integrating the chat bar transcript or not?
Frode Hegland: No it’s not. I have thought about that. And in the last meeting, because we’re supposed to be posting, I.
Peter Wasilko: Tend to beat a lot into the side chat. That doesn’t necessarily come out in audio, so I think it’s missing. A fair amount of input.
Frode Hegland: Right. Okay. So I think that’s absolutely a valid point. I know as I can. The second. This is for the prompt.
Speaker4: We’ve been said.
Frode Hegland: We’re in a meeting. At least say hi if you’re going to drop.
Speaker4: All right then. And she likes.
Frode Hegland: Yeah. Excellent. Can I continue my work now, please? Beautiful man. Okay. Right. Thank you. Peter, that’s absolutely valid. Thank you for the wave as well. Absolutely a valid criticism. What I’ll do from now on, I will paste in the chat transcript at the end of it. Perfect.
Peter Wasilko: That should fix it.
Frode Hegland: Very good. And I’ve also explicitly said here, take that into account. So yeah, that was definitely missing. Any other comments on this? You know, if you guys had a chance to look at the PDF. I just think it’s so cool. The. I mean, last meeting we did discuss uploading. There was a big thing. I think it’s incredible. Okay. Anyway, that was that. We do have an agenda today. As is here. Denny is not here. So we’re not going to do use cases. The vast majority of discussion today will be the work that Andrew is doing. And Adam is now looking to have more time to join us. And then we’ll be doing three things. He will be assisting Andrew further at Andrew’s requests. He will help make sure that we really optimize our uploading procedure, because Andrew has done work there. And to my great annoyance, and I say this with respect, both Fabian and Brandon are like, oh, it’s really easy. Well, let’s do it then. So Adam agrees that he also says, oh, it’s really easy, but he’s very happy to support you and make sure there’s nothing they’re lost. What we discussed on Monday is probably the way we’re going to go. You’re probably down 99% of the work. And finally, Adam is also going to do his independent stuff based on the same data. So what we talked about briefly this morning was that when we click on the link to our things and we now have the one sphere if when we get to, let’s say September in Poland, let’s say Andrew has two different things. There’ll be two spheres. If Adam has done three things, there’ll be five spheres, so it’ll be spheres in the circles you can click which experience you’re going into. So that’s kind of fun. So it means we can have a variety of experiences based on the same thing. And we have a unified entrance mechanism. That sounds cool, doesn’t it? And then Andrew, you’re happy with that, right?
Andrew Thompson: I just one clarification are those five spheres on the same web page that all load into their own experiences, or is this like there’s a landing page with the five spheres and they link out to the experiences.
Frode Hegland: To be decided? Okay, we get closer to the dates. I think for now it’s just going to be one sphere. And if Adam does something different, you will do sphere in his own thing and then we will decide later on, you guys will decide on the level of effort involved.
Andrew Thompson: Yeah. It wouldn’t make a lot of sense to stack multiple experiences in one. Page. Because you’d have a lot of redundant code. But it would make sense to have a landing page styled with the spheres, and each sphere links to the experience. And you’d have a sphere on that page, so it would it would be very consistent and it would work really well. I like your your steaming.
Frode Hegland: I agree with everything you said. So just following going through the announcement bit a little bit on Vision development. This is it’s going to sound like a huge detour. It isn’t. I have written a an email to Apple saying, please fix this, that and the other on the Vision Pro, and I sent it to Adam just to see if he had anything to add. Obviously. Fortunately, he hasn’t had a Vision Pro to to use yet, but just in general terms. And one of my things that I wrote is, oh, we can’t spawn more than one window. That’s completely wrong. You can spawn more than one window. So that changed the vocabulary. And before I get to this, there’s another vocabulary thing that Adam brought up, which I think is very important. And I think I’m going to get a big smile from Mark. Libraries. What is it? The thing in the JSON is not a library according to Adam, and I think he is right. It is a catalog. Because it doesn’t actually contain the data. So I think we should change our language that we have cataloged JSON, and then we have other mechanism to actually upload individual documents were available. Does that sound right? Okay, cool. And that relates to the question here. First of all, on a philosophical note, as I talked about the other day when Mark was off in the wilderness Mark pointed out when I did the map in author years ago that hang on, I’m just going to open it for, for the sake of discussion, then this map here that it may possibly be most useful at the beginning of a project, or maybe at the end.
Frode Hegland: It may not be useful during the project work. I think that was a very important insight because and I don’t know why it happened so strongly to me, but who here has seen the final episode of Shogun? Only. Adam. Okay, so we have one cultured person amongst us. Fine. You have to watch it. It’s the best television in decades. Anyway, at the end, there’s this scene where you have this big bad guy samurai warrior, and he has to write something, and he writes on a piece of paper that is maybe three A4 sheets wide, roughly. And it just really struck me that the a human scale for reading and writing probably is the size of an open, large book, roughly give or take, because beyond that, we have to move our head a lot, not just our eyes. Right? How does that fit in with using the whole page? And of course, there is a reason why a murder wall is a wall. That’s because of when we read and write. It needs to be in a focused card size space. But when we think that’s when we need more space. So it seems to be reiterating what many people have said. It’s just clarified for me that there’s a diff between reading and writing and thinking.
Frode Hegland: It’s blurred. It’s not black and white, but what that means is when Adam pointed out my complete mistake with the window thing, I just forgot it was possible. Here’s it’s kind of it’s just a feature in author, but I think it’s relevant for our discussion in a big way. Currently an author. We have this thing I called folding. So it folds into an outline, right? On iOS, though I’ve called it outline, not folding. And I’ve been going back and forth with the terminology. And here’s the thing I really want us to discuss. And that is. So until today, the idea was that whether you are an iOS version or Mac, you fold. And instead of just getting. This. You also get a list here at the bottom or other views for instance to see names like this keywords, defined terms, all these different things that uses the heading the outline. Right. But I still call that. Outline on iOS. And I wanted to put it back to fold. Sorry for making a confusing thing here, but what I realized is that if this button results in a new window on the side in vision, it’s not folding, is it? If you make a new thing on the side, it’s not a folding, it’s the wrong action. So in terms of our own vocabulary, we should probably call it an outline, shouldn’t we? What do you guys think?
Mark Anderson: What do you mean by a new sorry? When you say a new window, what is the content of the new window? Is it a folded view of the existing view?
Frode Hegland: I’ll show you. That’s the perfect question.
Peter Wasilko: So I had to call it a view or a perspective.
Frode Hegland: But yeah, that makes sense. But the thing is. I messed up a little bit here. So. On visionOS. Now what will happen is you have this window on the left writing away, and you tap outline instead of something happening on the same space as it does in iOS and Mac. This new window appears on the side. That window shows all the headings in the document. You can tap on them for this main view to jump to that section, right.
Mark Anderson: So So they’re only ever two two windows. That’s not presupposing they can’t be. What I’m trying to understand is so essentially the button for want of a better word in the bottom right is like a plus that opens a new window. But I think you’re doing something more specific than that. You’re not just opening a new window, you’re opening a specific new window of a fixed view type. So basically the button bottom right is a show hide toggle for a folded view of the current document. The thought behind the question is that I think the the issue about folding outline is interesting because another thing I’ve tripped across recently was a sort of completely pointless argument between two non-meeting sets of minds about people who expected an outline to work within the current text of the document, as opposed to people who were looking at the text of outline of an outline. And whilst that’s only a thing of scope, it does bring home this thing that the outline is very structural. The fold is actually quite a nuanced and enhancement. So if you understand what’s going on, they are one and the same. If you don’t, it might be confusing, because if you don’t understand what the folding is doing, you’re expected to see the outline, and you would expect to see all the elements of the outline at whatever scope you’ve given it, you know, down to a level, whatever, whatever you know is expected. I think the fold is more subtle than that because it doesn’t necessarily show headings for which there is no pertinent contact content within the folds. Question. Is that correct?
Frode Hegland: Kind of the new thing would be all these things at the bottom of the screen and the folded view, because, as you know, Mark in author, you can view only names, glossary, all of that stuff. Right? That should be defined. Thank you. So when it’s bold for headings, it’ll only show headings. If I now tap on highlights, it’ll show any highlighted text plus the headings above only those. So all these views get in here. Now the way it works in iOS right now. Is. See that? So that button, that’s hard to say. It was outlined by tap, you know. So now. And jump around the document. By that. So it becomes. And then I can hide the outline again.
Mark Anderson: Okay.
Frode Hegland: So there is no involving you guys in this particular issue. Is that because in vision these are two floating windows. And there can be many more. I hope to have the map and all of that stuff as individual floating windows. Doesn’t it make sense that this button to spawn this thing with headings plus should be called outline rather than fold? It makes more sense, right? Okay. I just find it useful sometimes to go through some of these. Definitions. Okay. So. Yeah. Peter.
Peter Wasilko: Yeah. Do you think maybe it would help if there was, like, a little mini icon next to a button that’s going to be spawning a new window? So the people could just see from the icon that if I click this button, it’s going to be creating a new spatial window, whereas this other button would be doing something in the window that I’m in. And because it doesn’t have that little icon next to it, I know that it’s just a local command versus a creating a new space command.
Frode Hegland: No. And there was a reason for that. I’ve thought about this type of thing a lot. I used to spend a ton of time wondering, what’s the difference between an internal link on a page versus external. So I absolutely agree with your questioning. The reason I don’t think we should do that in this case is because at some point soon you should be able to choose something like a long press spawns a new window, a quick press uses the same window. And this is why I’m also taking Andrew’s time with this. Because this the notion of over time, the user should be able to determine whether they’re reforming their current view versus spawning further views in addition.
Peter Wasilko: That makes a lot of sense.
Frode Hegland: Okay, so but thank you for the question. That was a really important question. And this is the kind of thinking that I’m going through. Because the email that I’m sending to Apple, I’ll probably share it with you guys before I send it actually, because it is It’s intended for a few things. There’s a few bugs that I want to fix, obviously, but also I want to strengthen our relationship with the vision team at Apple. Now that both Randall and Bruce Horn is in that team. So I’ll send it to you guys before I send it if you have any thoughts, I basically I want to say. Subtly. The vision isn’t selling very much. We’re working on making it a better working environment. Help us. You know. That’s the thing. Okay. And then on the agenda here, we have potential design day more Europe friendly time. At no point discussing that today, but when Adam is out of the woods, we’re dealing with family things. Adam, you’ll suggest today that in the time that works better for you, right? Yeah. I’m seeing you’re nodding. That’s good. So we’ll get back to that. And Andrew, that’ll probably be while you Americans are asleep, but it’ll be very high level. It will be recorded. But if something is of relevance to you, then we will send it. It’ll probably just be us having fun. And then we’re looking at Future of Text Social. We’re going to have to figure out some more timing on that. We’ll be dealing with that in the slack and in the and the Twitter thing. It’s getting complicated. And then the question is any other announcements? Come on guys. Mark, you want to talk about the joy of not being digital for a while?
Mark Anderson: No, apart from it was very relaxing, so I deliberately didn’t I did I did some sample news, but essentially I’ve had a week offline. And just just walking. Mainly I was extremely, extremely useful. Time for reflection.
Frode Hegland: Just. That’s good. I’m just adding add notes to transcript. Just so I don’t forget, Peter. One more thing. Before we jump in here. Yeah. So for the the vision stuff, just to spur your thoughts, maybe I have something to add to it. On quest you can choose to have a pass through just for your keyboard. You can’t do that on the Mac. Oh, excuse me on vision yet. So you have to have half an environment, which is not ideal. They should fix that. There are some issues with the keyboard, which is bizarre. Most of these are just kind of bugs. The they haven’t polished the software. They. Yeah, the software very much. If you are an author and you select text using your eyes and dragging, you get an iOS style contextual menu. If you use a trackpad and select, you get a mac OS vertical menu. Right. And the iOS one isn’t very iOS isn’t very wide. We managed to make it finally on the iPad. We managed to make it fit the screen so you can easily tap when the iPhone is ridiculous, of course, but these are things they haven’t done for the for the vision yet. And then, of course, in addition to having more windows open, which Adama showed me how we can do, we need more control over that so they don’t just open up here or there. And then we have For their issues. Yeah, I think that’s it. Okay. Andrew, are you ready? Oh, hang on, one more thing. I’m so sorry. One more thing on the agenda. Okay. This is five seconds. Hang on, I’m just going to copy some links for you. We need to start inviting people soon. Please have a look at the invitation. Please have a look at the. Invitation less than. If you have someone, just email me and I’ll add it to the list. Okay, we really need to get further. We’ll have a little icon log on everything. That’s it. Sorry. Please, please go ahead. Andrew.
Andrew Thompson: Okay. I don’t know how many people have seen the the change log. There’s a whole bunch of like really little things that have been done. So I’m not going to go through all of those. But for main relevance, we now have a snap focus distance to see how that works. So you can select any of the text objects. In this case, it’s still just citations. But this next week I’m working on more. And you can, you can focus it so it snaps up close, adds a background and scales it down so it’s more comfortable. And if you load in a document from the library or catalog, I guess we’re calling it now. That will auto focus by default. You can only have one thing focused at a time, and it’s got a little animation that transitions between it. So that’s the main thing that’s been tested this week. Like I said, lots of like little changes as well. So you may notice other little Ease of use stuff. And. Yeah. And if anyone cares about that, it’s it’s listed in the. The updates.
Frode Hegland: Question. Yes. Upload choose. File has been moved top left, but it seems like I have to do it now.
Andrew Thompson: Rather than No, you just haven’t loaded the page properly. I’m thinking it shouldn’t be top left unless the CSS failed to load.
Frode Hegland: It’s not loading beyond that.
Andrew Thompson: Okay, let me take a look. It was running last night for me.
Adam Wern: Yeah, I had the same issue right now, a moment ago.
Andrew Thompson: Something’s probably broken then.
Mark Anderson: Yeah, I think I basically it’s not finding the files.
Andrew Thompson: Yeah, it’s it’s failing to find Yeah, I see it’s failed to find a CS. Let me go take a look. Sometimes NPM gives it the wrong. Url to the assets, and I know how to fix that really quick, but it may have overridden that one second.
Frode Hegland: This is so bizarre. I’ve taken the virtual Mac display and overlaid it exactly on the normal monitor. Global use of tech.
Andrew Thompson: Okay, try reloading the page. See if that does something.
Frode Hegland: Yes. Oh. Okay, I want I want everyone else to make a big sound. When you see what he’s done. Loads for you. Rob, you’re going to love this. Mark, can you see it? Adam, can you see it?
Mark Anderson: I’m into the. I’ve gone through the start with our little lines going on it, and I’m now into the display, but I.
Frode Hegland: How cool is it with the lines going through it? Yeah. Neat. I’ll show you why. It’s super neat. The special reason. So our invitation, which there’s a link to in our chat. Just put it again. Where is the agenda? Good question. Sorry. Usually I put it in the chat. It’s on base camp. But most talk. Oh, yeah. It wasn’t posted. I’m sorry. I’m sorry. It was still draft. It’s doing now. Yeah, this is really impressive under just these little animations. Makes me very happy.
Andrew Thompson: They look they look kind of cool as the menu. They look less cool inside the library. So I have some plans on how to make that better in the future, but for now, this was all I could fit in.
Frode Hegland: I think because of your brilliance here. When you look at a document, it shouldn’t be there. I agree it doesn’t work in the background. However, when you long press the sphere then it’s okay to have. And I see you’ve also got them on the on the little sphere. Andrew, this is so cool.
Mark Anderson: And one thing I just reflection. I wandering through the thing I found myself on the menu bar, click the export. And because I hadn’t previously allowed the site to download things I ended up going outside, basically back in the passthrough mode. It asked to do a download which I clicked on, but I then had to reopen the browser window to accept a a download minor implementation stuff. The relevant part being that it was a bit odd that I appeared to be dumped out of the simulation for no good reason.
Andrew Thompson: Yeah. That’s just how the the OS handles the browser if you don’t have an application enabled. I wish it wasn’t that rough. I’ve gone and enabled it for. The URL on mine and then it works. But perhaps we should have some little information tab. Or we could try. When you first load the page, maybe to download like a dummy file that prompts the user to enable these things right away.
Mark Anderson: Yeah. And one more quick question. Sorry, because I can’t see everybody else, but you can probably see from where my hands are. I’m trying to interact with the reading distance control.
Andrew Thompson: Have you tried touching it? Yeah, that one’s that one’s actually touch based. Everything else is au remote.
Mark Anderson: Yeah. That’s the other hand. Okay. Cool. Yeah.
Andrew Thompson: So that’s the thing. That menu was made a long time ago. And we’ve since changed a lot of design principles so that it is due for a redesign, for sure.
Mark Anderson: No, I know, I understand. Not that it appears to be changing anything I can, I can move the slider for reading distance.
Frode Hegland: Oh, it does change, but it doesn’t change the focused text. If you have a white background around the text, it doesn’t. That’s where I got confused.
Andrew Thompson: Yeah. I mean, I want to get that changed to I’m thinking of adding two different slider dots to the menu. So one for focus, one for the other layer. And of course they couldn’t pass each other. But since we already want to redesign the menu at some point, I’m not sure if that’s a priority or not.
Frode Hegland: So I have a question here. I’m very happy with this. So I’m now saying white background focused reference list. Right. But from what we have now, is it possible to interact with a piece to see an abstract or something in this view?
Andrew Thompson: No. No abstract yet? I, I put that dev time towards the orange lines, unfortunately. But this week I’m going to be working on abstract. And was.
Frode Hegland: That when I first came into this environment, I thought I saw.
Andrew Thompson: It. Yeah. The abstract still sits in the the library or catalog page. It doesn’t load in as a separate piece yet. But yeah, this this next week I’m going to be working on loading a lot more stuff from the document. So that’s my priority. And, Mark, since you mentioned the save load. It is worth pointing out just like last week, I don’t have the save load system yet working for the new implementation. So I fixed last week’s save load system, and then I ran out of time to implement this week’s. So the focus stuff doesn’t save properly. You’ll you’ll probably crash the software if you try to export and then load in one of those.
Mark Anderson: Okay, so no problem. And I’ve now got the hang the one bit I’m just trying to work out is when I go into the current demo you got uploaded. It starts with the gray background thing on the right. And then I do something and it goes to the white background, which is fine. Am I supposed to be able to toggle between those two?
Andrew Thompson: So yeah, that’s the long press on your your wrist.
Mark Anderson: I’ll give that a try then. Okay.
Andrew Thompson: Yes, that’s that’s to load in multiple documents from the catalog you’ve imported. It has won by default, but it’ll it’ll change.
Frode Hegland: So Andrew, many questions. One of them is, it seems to me now by my eyes that the text is more not good. It’s hard to explain, but it kind of looks more pixelated. Is it a different rendering mechanism, or is it just that you’ve changed it to the point where I actually notice?
Andrew Thompson: Are you talking like the the focus compared to everything else? Are you just saying in general it looks worse? Because I didn’t change anything there.
Frode Hegland: In general, it looks worse. The focus looks less worse.
Andrew Thompson: That’s strange. Yeah. I didn’t didn’t change the rendering system at all.
Frode Hegland: But I am on this headset. I’m wondering if it’s different in the quest with this type of text rendering.
Andrew Thompson: No, it’s the same text renderer. I’m not sure why it would. Be different this week as opposed you didn’t notice it like last week or anything like that. It’s just started now.
Frode Hegland: I think it’s quite simply that everything is better now, so I’m more critical. It may very well be that.
Andrew Thompson: Okay. Yeah. Maybe the the focus is just sharper. So everything else, in contrast looks worse. I’m a little bit unsure.
Frode Hegland: So in terms of notes and everyone else, if you disagree with any of this, please do say I think the background should be a touch darker. And the environment. Especially in the in the library, it is darker. I think the library darkness should be in the reading, and the library should be just a little bit darker.
Andrew Thompson: Yeah, we can do that. That’s that’s an easy adjustment. We can see how that looks.
Frode Hegland: Okay.
Mark Anderson: Can I learn from the greats from the library. Oh. Right, I I’ve done it. Thank you. Sorry. Answer my question. So if I pinch the library item, it loads into the main display space.
Frode Hegland: What? Oh, yeah. Yeah, yeah. This is crazy. Actually, they.
Mark Anderson: So, Andrea, I let me put this as a thought because I then really got it. It’s a really nice now seeing the document. So I’ve loaded a document, I see a citation. I’m seeing a very, very long, depending on the length of the citation given, which says to me, we’re probably showing more information than we need. Because in the first instance probably what you need to know is the authors, however they’re listed for that. So in this case it’s up to four authors or it’s blogs et al. The title. And then possibly for things that have a DIY or a DIY URL or a URL if they have them. In other words, if it’s got some link that we might want to, even if not now use, you know, something we might want to interact with. But otherwise I suspect most of the extra information might. Better be essentially elided for the first display. It just it will just give you a perhaps it will give us a slightly tighter information display. Does that make sense?
Andrew Thompson: Sort of. Are you talking about the citations or are you talking about the citations?
Mark Anderson: So the citations, one, exactly as it were, the white view when you’ve loaded it.
Andrew Thompson: Okay. Yeah. With the it’s got just everything that’s there just pasted in it didn’t care about what’s there. It didn’t trim anything. If you have sort of a format you want me to follow on what we can just erase or something like that, then absolutely. I’d be happy to make changes to that. It’s quite obnoxious to look at sometimes when they have like a massive URL that you can’t do anything with and it just stretches off.
Mark Anderson: Well, that’s another interesting thing is it’s, it’s it’s the balance of those sort of elisions where you want, you want something that isn’t too. So as you, as you might put it, too obnoxious visually and something you can absorb. So in the first instance, okay. So there’s a link there and it’s a link that’s too long to show. Fine. Okay. And then I might want to have a secondary interaction to see the whole of it. Because actually I do need to see the whole of it. Otherwise I might just need to know that, oh, there’s a URL I can follow here. And I take it that as long as you’ve got something like a JSON or a BibTeX underlying this, the citation information, then obviously you can get you can get access to those individual bits because if you’re correct.
Andrew Thompson: Yeah.
Mark Anderson: Can I have the authors date? It’s probably author year and title. The primary things that generally would be the most useful to surface. But I’m thinking of it, for instance, this is another interesting thing. So as we get as we get, say, get the documents from an ordinary PDF, we don’t have that. If we have a document with a PDF with visual meta, we do. Or if you have something like a JSON or BibTeX you likely do have that granularity. So this is this is an interesting thing going forward. In other words, our legacy publication, our legacy publishing styles actually aren’t that helpful to us.
Frode Hegland: Yeah. On this, we’re making an assumption that we have perfect metadata because this is a proof of concept, and you are the one who has provided this perfect metadata. So we are making the assumption that we had a big discussion over the last two weeks where Denny rightly was defending the need for PDF looking documents. So what we are doing now is trying to use HTML versions which are available from ACM currently. But rendering them to look PDF ish with the added level of interactive, because I do love PDF for some uses, but certainly not in XR. So so that’s what we’re going to try to do. That means we can have snippets and so on. Also, the kind of views and limitations you’re talking about, Mark, is something I would really like you to write down because you are an expert in that area or what would be useful to see. And furthermore, since we have both Mark and Adam, if you can all please click on the last link that I put in the chat. Chat. And for you, who was. Have you seen this before? Andrew.
Andrew Thompson: So maybe it’s trying to load in.
Frode Hegland: Okay. Please tell me when you got it.
Andrew Thompson: Okay. It’s here. Now, this does not look familiar, so I guess I have not seen it.
Frode Hegland: So with the risk of Mark and Adam coded this Mark provided the metadata, but so you can see something with a white background with lots of columns on top, right. Andrew? Yes. Oh, okay. Sorry, I couldn’t see your face. Right. Please hit the tab key on your keyboard. And please select anything. So what dynamically happens here is that So I just tabbed myself. It’s probably obvious, but that document you now see more of it. On the left, you see everything. It cites its reference section. On the right you see all the documents in our known system which references it. So I think we’re at the point now where in our library we can start doing this kind of view. I’ll ask everybody what they think. Okay. Well, I was going to ask Mark first, but with that nodding, I don’t think that’s necessary. Yeah.
Mark Anderson: No, absolutely. I think this is really interesting. And we’ll have to work out, you know, what document format sort of data. But I absolutely in short, yes, I think this is something that lends itself to the XR space very much and is a really, I think, a really useful way to explore things. Certainly it’s interesting that when I show people these views it’s something that most people actually get and there’s and there’s, there is a sort of sense of, well, why can’t I do this now? In other words, why isn’t this sort of a way to do it? I can see all the background reasons to it, but what that says to me is it’s pushing slightly on an open door in terms of the style of presentation is useful. I think the the white view is the one that seems to have the most instant sort of sense of people getting it. But the other interesting thing is people really like the idea of basically changing the view, spec the tab through to something else, and once you understand what that’s allowing them to do again, I’ve been remarkably surprised at how little sort of pushback there’s been. I certainly never shown it to anyone who said, I don’t understand what it is, and I never use it, which of itself is interesting.
Frode Hegland: So yes, on all points I’m leaning towards the beige design because I did that one. However, okay, I’m sitting now in perfect spacing. I’m going to go back into BR. Because now we’re at the point of really discussing this. So I’m now in our library. Let’s see. Documents on the left. I see abstracts on the right. Absolutely brilliant. Nothing happens if I have my flat hand. I do pointing thing and then. I get the laser thing. And the laser thing kind of goes away after a while. I have the little dot, which is absolutely great. This to me is perfect. If anyone disagrees, that’s fine. I’m not tapping on one. And that all it opens now, at least under. I can’t see your face. Of course, if I’m saying something wrong, please do tell. I believe what I’m saying is the reference section from the document I tapped on. Is that correct?
Andrew Thompson: Yes, that’s that’s what it’s doing. It’s just showing it can load content. And like I said this week, we’re trying to load in actual document content rather than just the citation.
Frode Hegland: Yeah. And then I tapped on one reference and I do find and document and it shows up in the background of the most horrible way, which is absolutely fine because we’re now working on the whole focus thing. So I’m not what I’m going to do now is try to unfocus. Focus, by the way. Please tell me, Gus, if you agree. When I point to the margin that expands to be greater interaction. That is what should spawn the menu to allow me to do unfocus and close instead of individual text items. Do you agree with me?
Andrew Thompson: On on selecting what I was writing a note.
Frode Hegland: What I think I mean is that I’m in the focus mode. I have reference section here. When I select an item, I get a pop up menu. What I would like to keep there is detach from group clone from group finding document marked as text. And close just to close the.
Andrew Thompson: Window, right, and move the focus somewhere else. But where did you want to put it instead?
Frode Hegland: I think that the moving bar on the left or something like that should produce that, because it’s an action that is for this whole rectangle rather than one item.
Andrew Thompson: Oh, okay. I see how that would make sense if you had like a mouse with left and right click. But considering we only have one button which is select. How would we do that? Because it would just pop up every time you try to drag.
Frode Hegland: That’s a good point. Hang on. I’m just trying to. I made a mess there in the corner with a few. Pulled out items. So the items that you pulled out this show and text. They obviously will need a focus option too, right?
Andrew Thompson: Yeah. And I think what’ll I’ll need to do is if you try to find in text a document that’s focused, it should auto focus that as well and not replace the currently focused text.
Frode Hegland: All right. So this is what I want you to do. And please tell me anyone if you disagree. As always in the pop up menu there is space between the title, the commands and the space between before close. Right. I think we need to. Design this menu a little bit better so that the focus is separate. So, guys, I’m asking all of you now, the button in the pop up menu that closes the pop up menu that’s currently called close. Should that maybe be a X top left hand corner or something?
Adam Wern: To take. Taking one step back, I would argue, still argue against the context. Menu. Okay. Okay. To me to to me, it makes much more first thing is that I trigger it all the time when working. Yes. So it pops it pops up all the time when it kind of clicking at items. So it really hides. It gets in the way all the time. I’m. I want to punch this little menu sometimes. But that is that is a problem with context menus. And most of them detach from group. It’s a bit unclear when you would like to do the detach from group. Most often you have a list and you pull things out. How? What would the use case be to actually remove an item from this kind of list? A fixed known reference list. Cloning it, I see you. That’s obvious. You want to keep. Keep it in a separate collection. So if if we look at all of them let’s say detach is not that relevant. Clone from group could be an arrow next to the to next to an item on the left hand side. Finding document could also be kind of an icon or the word find next to the item. So if you click it in place, it’s next to the item and not on top of it. Marxist text could actually be the four color, because with there is a submenu, I guess, for under under Marxist text. So it’s kind of clunky to go into several layers to get to mark an item. I would say that what if you had five dots or three dots of color on the right hand side of an item? You could just press it immediately. Or have a kind of a marker submenu on the right hand side.
Adam Wern: So and if you do that and focus would perhaps be better to have. On either on your wrist or on the focus area. If there is a focus area, kind of a close reading area that you point to it and then you get it there, and that would also be a quite natural. So when I say this, the thing is that the whole menu would disappear and you wouldn’t need a close button because there is no menu. The menu is spread out as actions in the next, next to the items. And many, many systems are designed like that where you have. Kind of the sub actions appearing in proximity to the to the list. A small either icons or a word like action buttons. But the good thing about that is that it would be much less modal. You don’t have a big pop up in your face. You have small actions that lead out from each item, and it could be added to quite easily without exploding or like building up a big menu that is in the way. And that would also make it. More user friendly, and that you don’t get the pop up in front of the items covering most of the So what? I would argue against the context menu perhaps for long press actions where if you really want to delete or export or do something a bit more. Involved in destructive actions or seldom used actions. It could be used to have a. But if you want to detach something, I would just point at it and drag it out to the side. I wouldn’t even have a clone button I would hold drag out.
Frode Hegland: Yeah, that was a lot of words and I agree, but let’s try to see what that actually would be. Should I do it in keynote and you kind of dictate to me, or should we do it differently? Because I agree in principle, but I think we should design it. Is that cool? Good time to do now. At.
Adam Wern: First, I think it’s good to have the discussion before we design so we don’t jump into something suboptimal. Or.
Frode Hegland: I agree, but let’s have a look at just a few conceptual rough things. Right? So let’s pretend this is the library. And you select something. What should happen when you select something? Should you see the abstract like we have now? Do you think? Obviously asking everybody. I just want to try a few rough reality, so to speak.
Andrew Thompson: Are you talking when you first load in from.
Frode Hegland: The okay.
Andrew Thompson: Library.
Adam Wern: So so I discussed the step one one step further down. Could we do that one first? Perhaps or else we will do. Now you’re stepping into a different interaction that is also important to design more.
Frode Hegland: I agree, but let me just Okay. So boom here is a piece of okay. So this piece of text could be a reference section. It could be the abstract. It could be any part of the document. Right. Yep. And now we want to do things to it. So currently we have a I’m just going to write it out. So we make sure we don’t lose out any functions. We have a context menu. This is so cool I can see it over here. So detach. Come on.
Adam Wern: So a comment on the detached. What was the thinking of detached? Was there a reason for it or was it just a test thing?
Frode Hegland: Please take this, Andrew, while I keep writing.
Andrew Thompson: I don’t remember who suggested it. Everything in the pop up menu was added because somebody asked for it.
Adam Wern: Yep, yep.
Andrew Thompson: So I don’t remember who was who was on.
Adam Wern: Yeah, not necessarily who, but what was the kind of thinking do we do we remember that? Well, y we have the detached. I’m sure there are situations where you would like to kind of drag out and remove things from a list if you kind of, if you have two lists, you may want to move one item to another kind of drag and drop.
Andrew Thompson: Yeah. If I had to guess, I think it was just kind of a natural progression for the let’s explode the document idea. So it’s like, well, let’s break everything into as many pieces as we can. Okay. It was added when we first added the pop up menu, which was months ago now, so I’m not quite sure. I don’t remember all the context for it.
Adam Wern: Yeah, it may make more sense in other contexts than reference lists. For example, I think cloning and copying is the most important action and that could be adjusted drag, pinch, drag out to the position you want to have it so you have one fluid motion instead of Press clone. Get the clone and then position. It is kind of that’s 3 or 4 actions instead having. Pinch and drag would be, you see a bit more adjusting.
Andrew Thompson: Removing the handle and just having dragging the text itself pulls it apart.
Adam Wern: Yeah. For dragging things out from a list. I’m just doing that. Yes. So pinch and drag is would be a new action, a new gesture, a universal gesture for lists or things that could be dragged. So if you have a list, if you pinch and and if you don’t release the pinch immediately, you could drag it out and get a copy to your to your collection.
Frode Hegland: What.
Adam Wern: Would be an equivalent in touch land?
Mark Anderson: What is forgive me, I, I’m if I’m forgetting what is group in this context? On the context menu.
Frode Hegland: So that was when you had a lot of references. And if you detached one and then you had another. I’m just putting this as a rough note. I think we should kind of forget about it and kind of start fresh. And, you know, I’m not doing any design here, but let’s say we have a piece of text and we are now not in the library, and you should be able to do things with it. So one thing you should be able to do is close it, right. Oops. That’s not what I wanted.
Mark Anderson: Close it in the sense of is this. Close and set focus back to where you were or is. Is this just closing a factory? What is the secondary window? I asked the question with sort of backstory to it, but I’m just trying to think, well, so part of the relevance is that if I’m looking at a effectively a small window of text, a particular point of focus, I would have got there from someone else. So in essence, am I is the close going back to where I was or is? Just saying I no longer wish to have this bit of display present?
Frode Hegland: The question is what you want it to be. You know, let’s.
Mark Anderson: No, indeed. But I mean, I suppose what I’m saying is, is close and sort of potentially implies two different things.
Frode Hegland: Yes. So it could mean close the pop up menu or close this piece of text I have now for the sake of what Adam was saying. I’ve just I’m not doing icons or anything, but what I’ve done now. Is just a rough idea of underneath whatever thing this is, we have a couple of options so they don’t go away. Maybe they’re not always strong. You have to point to them to really see them. But these are the key things we want. Maybe. Right?
Andrew Thompson: The one thing that I could point out if we’re trying to trim clothes could probably be removed. The pop up menu could just go away by. Pointing and selecting at nothing. Idea that might take care of that option.
Frode Hegland: Yeah, that would be very, very nice. So Adam, you can see what I’m doing as well, right?
Adam Wern: And not really because I’m in the headset and.
Frode Hegland: Okay.
Adam Wern: And and thinking.
Frode Hegland: I’m just saying that for simplicity without using icons. What if we had at the bottom of every window there was this grayed out thing, and then you point to it and then you get this options that we will decide, find and document is only relevant. You know, some of them would be context sensitive obviously would only be relevant if this was a reference. Mark and text would, you know, give you the option to color this thing, focus or defocus or unfocus or whatever it is will be toggled here. And then this will be to close this window. Not because there wouldn’t be a pope. I would just do this.
Andrew Thompson: Yeah, I see a potential issue. This only is designed for focus, right? So this would only work if you have it focused. How would this work for pieces of text in general?
Frode Hegland: So one of the things I said in the intro today is my personal thing of when we’re actually dealing with reading text, we probably want it. And focus, and we should probably allow the user to have more than one focus. So let’s say you have 3 or 4 of these up here. That should be allowed.
Andrew Thompson: I thought the point of focus was to only have one.
Frode Hegland: I thought so too, but now that I’m sitting here with my headset on, having your thing over there and having other thing, I’m thinking a bit differently then I’m just spitballing here to use that expression.
Andrew Thompson: Okay. The thing is, we.
Frode Hegland: Have.
Mark Anderson: It sounds like we sort of have two slightly overlapping meanings of focus, which I think I get but a separate. So there’s in terms of having multiple things in focus, what you’re really in a sense these are foregrounding things. So they’re they’re things within your visual exploration space that you want. And that’s in essence closer to you, perhaps in better focus or in more detail as opposed to the thing that actually has input, focus which may or may not necessarily need to be explicit depending on the quality of the tracking as time goes forward, that, that that may go away for now. It might be that you might have, say, four things essentially foregrounded, but we might need to know which of them has the focus, simply so that there’s clarity for the software as it exists at the moment to know what’s going on. Because obviously if it can’t guess what you’re trying to interact with, it’s useful if it there is at least an explicit way of saying, no, no, this particular text window is the one that I’m working with at the moment.
Frode Hegland: But. Yeah, exactly. So there are two definitions of what focus would be. And that’s why I’ve added the term frame to. Now frame just means have this white background focus will go to exactly what we talked about. Meaning. Front and center.
Mark Anderson: I’m not sure focus is I’m not sure focus is useful for that. That’s not I must say it doesn’t. I find myself having to ask what it means. Focus to me either means it’s in focus, like I can read it or it is. It has input. Basically it is the input locus. Otherwise I’m not sure the term.
Frode Hegland: You’re right, but what we mean is front and center with a background in a comfortable reading position for your primary reading space.
Mark Anderson: Sure. So what I’m suggesting, therefore, is probably focus isn’t the word for that use we need. Oh I.
Frode Hegland: Agree. So what do we think might work better?
Mark Anderson: Except, well, if it wasn’t such a long word, I’d say foreground, because essentially what you’re doing is you’re saying you want these to the to the you want these more present closer to you than other things. So there are other things you want in your display space. You want you you want to know that they’re there. You don’t want them out of the space. I thought the.
Frode Hegland: Problem with foreground, Mark, I don’t think that’s bad at all. What the rest of you think? But I think that’s quite nice.
Andrew Thompson: It doesn’t work well for a verb, but doesn’t need to be.
Frode Hegland: So two foreground. Yeah.
Andrew Thompson: Okay.
Frode Hegland: I got battery warning for the vision. So there we go. That’s life.
Speaker7: I’m sorry. So the.
Adam Wern: The alternative to that could be that, that the different reading distances are. Clearly visualized in a in another way. So you just say where you want it. Wall close or close distance. Other. So they are visually. Visually. Apparent and that perhaps the icon icons or something, or the action is so clear that we don’t need to call it focus or foreground either because it’s just sending it to a visual place or a visual an icon or, or an action. That is the represent what we want.
Frode Hegland: So on that note, please.
Adam Wern: But I don’t think that’s we that can be experimented with if if it’s an icon or not, but it should be clear from the not just from the text, but also from some supporting visuals where the different layers are kind of the back layer, a wall or the three different distances if we have three of them.
Mark Anderson: I think it’s I think it’s difficult at this stage because we have the classic chicken and egg problem is, you know, how many how many layers or distances do you need? Well, the answer is you don’t know until you’ve tried. And a murphy’s law means the one you first chose isn’t ideal. But essentially, as I understand, we have two at the moment. We have the general background display, the things we put into the display area, and there are things at the moment only one item which we’ve we’ve chosen to sort of foreground those fruits rays, we might, as we go forward, possibly be looking to more and one reason you might want to effectively put them away out of the foreground is you just pull too many things to the front, which is something easily done when you have an exploratory interface because you keep pulling. It’s like pulling more, more, more books off the shelf until you’ve got a table full of books and empty shelves, and you actually need to start pulling some back because you just can’t find stuff.
Adam Wern: So there could be just an arrow front or back that sends us to the front row or back row what we call the wall and or reading distance. Other times. So front and back would be.
Speaker7: Yeah. Yeah, it was interesting.
Mark Anderson: There is, I think, and it’s something that got raised before. Is that in effect in the learner mode? You actually need to understand you probably text helps because it’s sort of more explicit. But probably once the, once the user is experienced, you are things will either be sort of almost, almost contextual to what you’re doing or might be reduced to a symbol just to give you a target to point at.
Frode Hegland: Yeah, I personally think the issue with that is completely correct. But once you know text, you don’t read the text, you know, letter by letter, obviously it becomes a symbol. So I am not an icon man. But if we find useful icons, you know, fine. The idea here, by the way, is to have an arrow forward or background, and half of it would be light gray if it was foregrounded or backgrounded. I just also wanted to show this screenshot of author iOS on the small iPad. Look how beautiful all these options appear. You don’t have to have a little bit and then hunt. Hunt, hunt. Because the thing is, we will have commands such as ask I. I’m not saying we have to have that, but there will be contextual, specific things so we can put that on the only select the bit of text. But often you want to do it for the whole thing. Like, analyze this whole document. So that’s why having a strip of commands at the bottom that are not too many could be potentially useful. I mean, they ask something like an ask I when it’s on Mac. It is. Like this. You know, basic hierarchy. Which I guess it would be here to.
Adam Wern: So do you think this is back to the context menu versus the more palette like thing that you described a moment ago, kind of a thing that stays in place and only lights up if it’s relevant to the current action. That’s one. And we also talked about the wrist palette at some point, kind of a one side of the prism thing was for kind of context menu like actions. And so where do you find or where do you think the best interaction or the most fluid interaction I think is a good thing to go for here. The quickest action as well, I think quick is a good measure of simplicity and elegance here. If it’s quick. I don’t think the context menu really does that at the moment. So we I think going back to the earlier suggestions of palette menu, hand menu. Close distance pilot actions that floats in space perhaps or context. Contextual side icon icons that that are next to on the sides of the currently selected. I think those three things are stronger contenders for a fluid action. Okay, perhaps perhaps some gesture as well as well, but they are more restricted. We can you only have a few of them.
Speaker7: Okay.
Frode Hegland: You can see my view, right? Yeah okay. So I’m going to address some of the realities of what you’re talking about now because I find it different in the actual headset, those things than what I expected. So here we have an author. You may notice that it kind of flickers in and out, which is weird.
Speaker7: I’m.
Mark Anderson: Not sure we’re seeing what you think we’re seeing. We’re seeing what.
Speaker7: Like.
Mark Anderson: A keynote?
Speaker7: Yeah.
Frode Hegland: Oh, really? Yeah. Oh, yeah.
Speaker7: Second, a second.
Frode Hegland: So you.
Speaker7: Suppose I haven’t.
Frode Hegland: Seen clouds and stuff.
Adam Wern: We haven’t seen anything. We haven’t seen a still image from keynote and you gesturing.
Speaker7: Come on.
Frode Hegland: Very entertaining. Okay. Yep. Okay.
Speaker7: So.
Frode Hegland: Let me try that again. I’m trying to share my screen to this device. So. If I do this. Hang on. I’m just going to disconnect this. Okay. That is disconnected now. Very good. Now. What do you say?
Adam Wern: You staring at?
Speaker7: Book I.
Adam Wern: I only see you.
Mark Anderson: Yeah.
Speaker7: Same here.
Adam Wern: I see Mark.
Mark Anderson: But were you saying speaker? Yeah.
Frode Hegland: For some reason. Share screen went off. Okay. Now you’re saying my desktop, right?
Speaker7: Yeah, I think so, yeah. Okay.
Frode Hegland: And now you’re saying craziness.
Speaker7: Yep. Yep yep. Okay.
Frode Hegland: So I’m going to go into virtual environment. Otherwise I’ll have this looping thing. Now you’re saying beautiful clouds in Hawaii. Yep. Okay. So here is the current version of author. So I find that trackpad and keyboard are a lot better.
Speaker7: So.
Frode Hegland: Just even to click in. So if I now move this thing down. So this is ridiculous.
Speaker7: To me.
Frode Hegland: That’s all it is. I have to hit that tiny target, which often the eye tracking gets wrong. And so which.
Adam Wern: Which topic are we on now? Now I’m a bit unclear. Are we on on context menus or palettes or are we. Have we switched topic?
Frode Hegland: No, we’re on what you talked about. I’m addressing the fact of how to select things in actuality in the in the native apps. Because you were talking about context menu versus other things. Right?
Speaker7: Yeah. Right.
Frode Hegland: So first of all, to actually select a span of text is hard. I don’t think it’s going to be much easier in webXR. So I’m trying to move that. So this is the first thing we do not want to do. We don’t want to have a tiny little pill menu like this with an arrow. It’s ridiculous.
Speaker7: But this should.
Frode Hegland: Be much, much longer. So if I now try to do ask, I look at that hardly anything. So. I choose this one and I get the result here. This is similar to what we’re looking at, is do an action on a thing that will result in something else, right?
Speaker7: Bad guys.
Frode Hegland: We will have these kinds of things like reformat this text or I or whatever it might be, right?
Speaker7: Yeah, I see.
Adam Wern: Some some actions are like that, some are producing new content, some are moving it and some are restyling it in some way or, or removing it, or some are triggering other actions that are further away. So there are kind of 5 or 6 or several different categories of. Actions here that do slightly different things.
Mark Anderson: Before we whiz past, I see Peter’s got his hand up. Peter, do you want to jump in with your question?
Peter Wasilko: Yeah. I just wanted to say that for. Sit down, XR, I really think it would be nice if there was an option to drive the entire user interface purely from the keyboard.
Speaker7: Now, you won’t be.
Frode Hegland: Able to do that because, well, you at least. You mean keyboard and trackpad or mouse, right?
Peter Wasilko: I mean your physical keyboard. If you have a physical keyboard or Bluetooth keyboard keyboard connected to it.
Speaker7: But you.
Peter Wasilko: Drive the interface by typing.
Frode Hegland: Yeah, but you mean keyboard and trackpad, right?
Peter Wasilko: Keyboard and trackpad, keyboard and 3D mouse, but mainly keyboard.
Speaker7: Okay
Frode Hegland: I agree I have the design for me to do. Author is complicated by the fact that it will be both with and without a Bluetooth keyboard. With a Bluetooth keyboard, it’s better. But just like what happened here accidentally, I selected something. I didn’t mean to do that. And then I overwrote something. And you will notice on the screen there is no undo button. Siri, remind me mention ado in the letter.
Peter Wasilko: How could they not have an undo button?
Frode Hegland: Yeah it’s crazy. Right? So these are the realities of this that we shouldn’t fall into the trap of. So I do completely agree with you, Peter. In my experience, having a trackpad, which I’m going to now use.
Speaker7: Because I.
Frode Hegland: Have a dedicated track by just for.
Speaker7: This.
Frode Hegland: It makes the interaction a lot easier. But of course we and for this loan work since you now it wants to resize but it doesn’t want to go away, right. But let me show you how the trackpad works. Yeah, okay. See here the cursor. Does it show up well?
Adam Wern: Okay, but this is slightly different from what we’re talking about. These are kind of native affordances that are not so they are somewhat related, but they also feel like a somewhat of a distraction to the discussion we had a moment ago.
Frode Hegland: Yeah. No, hang on. Peter took us on a bit of a direction, but I just want to show that when you have a trackpad, you don’t need a 3D mouse or anything like that, but it works really, really well. So what I the point I was making is that first of all, see this? This is a different menu because I use the trackpad. So that’s just stupid. Whatever we are doing.
Speaker7: Wait.
Frode Hegland: This is so dumb.
Speaker7: But this.
Frode Hegland: Is. And I’m trying to undo. And I won’t do it.
Speaker7: Okay.
Speaker7: Okay.
Mark Anderson: So I want I take the what I take away from this is that actually, you don’t want targets that are too small. So whilst it might seem by today’s sort of 2D sort of UI design standard that, that things should be sort of nice and and inconspicuous if you’re having, if you need targets that you can hit, especially by eye because that’s presumably where things seem to be heading where, well, we’re either using these or we’re hoping, in other words, for anything except typing to actually do things like text input, where the where things seem to be moving is we’re going to be using hands or eyes. And voice. Sorry. Yes. Well, well voice gets voice gets around the problem of, of the box being big enough or small enough.
Speaker7: Voice.
Frode Hegland: Voice is a bit of a problem though, because you still, in many cases have to select something in order for the voice to know what the voice should apply to.
Speaker7: Yeah.
Frode Hegland: For instance, I cannot say undo Siri, undo Cyril, just say, what are you talking about? Which message would you like me to unsend for some reason, right? I don’t know how that is, but you still have to.
Adam Wern: But we could do that in webXR. At least on Apple and with some more if we want to have a command that is a voice command, that is undo. We could have that in webXR. Webxr. Vision Pro has the web speech API that we can translate to commands. But they must be triggered in a very thoughtful way and also executed in a thoughtful way. So. So we don’t trigger them by mistake. And so it’s easy to undo the undo sort of say, or refine the undo if we want meant something else. But it’s possible and, and that’s will certainly be the future of, of XR voice. It’s, it may be one of the strongest components, but the the I think the interfaces have not caught up to that yet. It’s got to come online now.
Mark Anderson: It may vary by language, but I know English can be delightfully imprecise when you get you know, this thing is not that thing. So there’s an awful lot of things it turns out are actually depend on a lot of context that the voice alone can’t, can’t actually deliver. But that’s a that’s a discovery we yet to go through.
Frode Hegland: So, guys, this is not a design exercise. What I’m doing here, what I’m doing is just trying to figure out what commands should be for different things, so we can discuss that.
Mark Anderson: Well, are we talking about? Sorry, I’m a bit confused here. Are we actually talking when we say it’s not a design exercise? Therefore it doesn’t really matter. We’re talking about the affordance of the command then, rather than what the command does or what the you know what the text is. I just say we’re because otherwise you slide around a bit. We seem to be chopping up.
Speaker7: What I’m talking about is.
Frode Hegland: What I’m talking about, not design. All I mean is whether it’s going to be a text button or whether it’s or whatever, that’s up for discussion. But we need to know what kind of things.
Speaker7: Well, I mean, one.
Frode Hegland: One thing that is clearly different is the reference section and other blocks of text in the document. Right? The reference section will have specific affordances. So You know, we were looking at the reference section earlier, but and we need to figure out what, what those affordances should be. But when you’re looking at a block of text in the reading mode, what are the things we want? We want to have some sort of way to say, get to my special foreground reading, and then maybe one saying, have a white frame behind it or not marked this text finding document and maybe some AI.
Frode Hegland: Any comments on this?
Mark Anderson: I think it’s a bit early to do that. I it would make more sense. I’d find this easier to understand. I for instance, I don’t think do we have a, do we have a sort of object model in the loosest sense of what the document is? So yes, we we know the an obvious thing is something like the citation list. So we clearly have text as well. And we’re talking about a bit of text. But is that like a paragraph block address system. So what are the bits and pieces we’re dealing with. That’s I’m not sure that’s something we’ve, we’ve, we’ve, we’ve worked out.
Frode Hegland: That’s for us to discuss and decide. But what I’m saying is that when we’re talking about a block of text that is at least now arbitrary, it could be one paragraph, many paragraph, the whole document what you.
Mark Anderson: Might want to do, the thing is, is a chicken egg there? Because what you might want to do with it might depend on the the granularity of the text that you’re dealing with. I suspect it will differ. If you say we’re looking at a whole section as opposed to say I mean, probably the smallest block that we display, I would imagine would probably be a paragraph. I guess you could drill down further. But if you’ve only got a sentence in scope, there’s not much you can really do with it apart from unless you’re drilling down because you want to use a drill down as a selection mechanism to give you the affordance of doing something else with that selection where you might be annotating it or you might be wanting to effectively copy it, to put or moving it or copying it to another place. So I suppose what I’m foregrounding is that that talking about what the affordances might be without actually thinking about what you’re doing can, could be a sort of exercise in slight futility, because when you come to actually use them, you find the things you’ve designed in there don’t get used in the place where you thought they would. I think that’s my my experience of learning tools along the way.
Frode Hegland: That is what I’m trying to do.
Speaker7: There are some. What I mean is so.
Mark Anderson: So if we don’t yet know what the granularity is, say, of the text is, then how can we decide what it is we might want to do with it? So it seems a more fruitful thing to do at this stage would be to have better understanding of how we deconstruct the document.
Speaker7: But we.
Adam Wern: Mark on that point. On that point yeah. We’re jumping back and forth between kind of text, as in yeah. Prose like or long, long text blocks and list like items. A library or a catalog is more like a list, like a thing. A reference list is a list like thing. An abstract is not a list like thing, perhaps. And so I think we’re jumping back and forth between those two and, and I think we should have two different separate discussions on how you highlight text and do actions or under which granularity and all the list like items. And I think many of the interaction design patterns that we have for list like items could be a bit similar in that how if it’s a context menu or different icons or palettes or other voice commands, but I still think there is a similar things you would do to lists like moving between lists, moving from lists, creating a new like detaching or dragging things out, highlighting things in list, pressing an item in a list. Pressing several items in a list. So there are things that are very list like. And I think just tackling those or is what we have started with or the reference list is list and the library list is a list. And we haven’t so far looked so much into the pros like things or long text blocks.
Adam Wern: I don’t think it’s wise to conflate them here in the discussion, in the design discussion. And I think this is design. It’s not styling things. And but it’s certainly interaction design. And one of the most important parts of this because this is what the people will see, the if we present them with a bit too slow interface or if it’s clunky in some way, or if we trigger things by mistake, which is a risk. Then when you create new interfaces, it’s really hard to do it well. And so that’s why I’m pointing like pushing for a design that is kind of removing a bit of the context menu issues. I think the current context menu is good in that it’s big and in front, so it doesn’t run into the problem Froud has with the small microscopic context menus and text selection issues there. But the problem is that it’s big, triggered by mistake. And then in front of in front of the actual things that you’re selecting. And if you happen to select something by mistake, it’s kind of hovering the things you’re working with, which is which forces things to be in your working memory instead of visualized for you.
Adam Wern: So I think we should look into ways of it could be that just a transparent menu would work better. It could be that it’s smaller. It could be that we separate different things of that context menu into different menus. It could be some things could be palette like things, others could be gestures and it could be a lot smaller. But I would argue that most of the things on the menu could be. More integrated into the closer to the item, more on the side and less in front of the thing as a big block modal block in front of it. That would to me that would feel more fluid, and I think the users would appreciate that. But it is also cost of implementation. But I think going forward from here, I think the building blocks are there. But if the context menu was kind of split up into different parts with a still being the buttons, they are, but not in the text box in front, I think we would have a much more fluid interface, and perhaps some things could be a gesture as well, like dragging a thing out from a list could be a gesture as its very fundamental. But but that is a slightly harder or more complex interaction.
Mark Anderson: So I understand this is probably one that well, Andrew might answer for me is, is that with something like with something like the contextual menu such you have at the moment, does it attach to the selected object or does it attach to the sort of the the view. So using as an example the reference list, if I select a, essentially, if I’m pointing out an item in the reference, is the context menu generated in relation to the objects I’ve selected, or is it just somehow within the the list, as it were, that I’m looking at? No, there’s there’s no hidden hidden angle to the question I ask out of ignorance.
Andrew Thompson: Yeah. So I, I’m not fully understanding what you mean by attach. So I’ll give you like both answers.
Mark Anderson: So what I mean is sorry where do you draw it. It’s drawn in relation. So in other words what is it anchored upon. Is it anchored upon the the list or is it anchored upon the item?
Andrew Thompson: If you’re talking anchor as in like physically attached it nothing. It’s just in the scene. It just generates at the lowest level. However, if you’re talking about what is it based upon with generation then it’s basically it’s fed information based on whatever you select. So right now say I have every, every bit of text anywhere has a type A tag I’ve added to it. And if the tag happens to be, say, a citation, then the pop up menu knows what to populate. And if it happens to be a reference, it populates differently. If it’s a library document, it populates differently. It calls the same pop up menu.
Speaker7: Just I think I was making something else.
Mark Anderson: Okay. This was reflecting on something that had cropped up about, you know, the, the the the menu was potentially occurring sort of in an inconvenient place. What I meant was it’s, it’s it’s it’s essentially like what I mean is, so if I have something selected and I wish to draw a context menu, the context menu has to appear in space effectively at an x, y position. Relevant to the thing that’s calling it is does that help at all? I was just trying to understand whether the context menu is, in a sense, something drawn within, say, the overall list or whether it was drawn at a position relevant, relevant to, I guess, one corner of the selected item.
Andrew Thompson: If I remember correctly, I think I draw it relative to the camera. So it’s always going to show up kind of where you’re facing no matter where you’re pointing. Yeah, I think.
Speaker7: That’s so that might.
Mark Anderson: That that might be inadvertently what’s, what’s leading to the potential of it sort of being in the way. So it might, it might want to basically if it was and if this is at all possible. Effectively be drawn off based on the object that’s being worked on. In other words, therefore.
Andrew Thompson: Like cursor position or something.
Mark Anderson: Yeah, yeah, yeah. And I don’t have a fixed view on that because I’m not I don’t have the skill to do it. But but you get what I’m driving at. It’s in effect that if we know that it needs to be offset from the thing we’re dealing in, then then its chances of being in front of it or obscuring it are far less so. I see Fred has his hand up.
Frode Hegland: I see. Rob had to go, but I see his question here is basically saying a user guide. And I think we should approach the design questions we’re discussing now, as though we’re designing a user guide or a walkthrough, because the actual commands will absolutely be different depending on what you need to do and what it is. But you will need commands. So in order for Andrew to start experimenting so we can better understand both through writing lists and to testing, I do think we need to do the design of how the option should appear and what the initial options should be. I don’t think that is jumping the gun because we have a pop up menu now that is far from ideal. So the user guide for Rob or anyone in the future needs to be. First you click on a link on our web page. That loads something that has a big sphere. It’s got some text on it. It says click unless you have your own library, in which case you load it.
Andrew Thompson: Well, I would say. Oh, sorry, Fred, I didn’t know you weren’t done.
Speaker7: No, no. It’s fine.
Andrew Thompson: I would say unless we want somebody to actively maintain this Tutorial of sorts. This manual. We should probably not write it until. Shortly before the presentation in September. Just because we’re going to keep changing stuff. Unless, like I said, if somebody wants to keep managing it and updating it.
Frode Hegland: I think I’m I think that’s my job to maintain it. I think it will will change. But I am worried about kind of a constant questioning of layers of this and that. I think we need to we can do that on Mondays. But I think on Wednesdays it needs to be. We think this works. Let’s try it. And then we’ll see how crap it is. The reason I went into author to show you things is there are so many unexpected things about selecting and text and all of this stuff. Even in an environment optimized by the biggest company in the world. So that’s why I think we should decide on stuff and go off and sweat and build it a little bit, and then change it next week if we have to. So I.
Speaker7: Have.
Mark Anderson: One thought in that in terms of the documentation is and also I, I shy away from the the visually descriptive it seems, it seems a seductive way to do it. But my experience of some decades of writing things for people who who are starting without my knowledge, is that that tends to not work. Because you don’t you don’t at that point have a common model. So it’s actually most useful to tell people what sort of things they may get. So I think one of the things that shines out from Rob Rob’s point is that, okay, so I go into this thing, what what do I get? What are the sort of things I can expect to interact with? And that that then leads you into the interactions you might do if you just say, if I just say to somebody, you’re going to see one of these, it doesn’t tell me what I can do. It just tells me I’m going to see something. So it’s not actually informative. It’s describing the experience. It’s not describing. It’s not actually helping the person understand what they’re doing.
Frode Hegland: I don’t think we disagree on that point at all.
Speaker7: I.
Frode Hegland: Think what we need to do is have a firm but to be changed model of what we tell a person, it is. So we’ve gotten to the point where they click on the thing. They’re now in the library and they have the thing on the wrist, so we have to decide. How do you interact with the library? You can, you know, scroll up and down and we have to decide when you stop on a thing, what happens currently it shows the abstract to the side. And if it if you tap on it, it opens that document into the non-library regular view. Right. That that’s what we have now, but what we have as the library. So this is under this is a request for you. Actually, the library as it is now is basically a reference section. It’s got numbers on the left. And Mark asked towards the beginning of the call. What does it not? Okay, okay, I just misremembering. Two different screens. Ignore that. That’s fine. But Mark pointed out we may be showing too much in the library. Ideally in the library probably should just show title author, date, not a URL or anything else. What do we show right now?
Andrew Thompson: We don’t show a URL either. Mark was talking about the citations.
Speaker7: Okay.
Frode Hegland: So okay, so in the library we need to decide when you point to an item, should the only option be to open that or should we have some sort of contextual options available. That’s the first thing we need to decide on.
Speaker7: Peter.
Peter Wasilko: Yes. What? I’m looking at the library. I’d also like to get a count of how many concepts are defined in that item, and maybe how many people or organizations or locations are referenced in it.
Speaker7: Okay, so you.
Frode Hegland: Would like to. That’s absolutely useful. So what you would like when we have it, when you select an item in the library, you get the abstract maybe plus additional stuff on kind of a.
Peter Wasilko: Yes or the count of metadata categories. So I have a sense of is this the real dense thing? It has a whole number of concepts defined in it. Yeah. One thing specifically.
Frode Hegland: Yeah. So what what we all agree on, when you point to an item in the library you should get further metadata. Right? Yep. So then one of the other things we need to decide is the demo that we saw that Adam made quite a while ago of the ACM library, where you can select an item and then you see instantly what it cited and what cited it. Do we want this to be in our library view?
Mark Anderson: If it’s feasible, I’d like it. I think I might my my my experience, just from the unprompted comments of people who’ve seen that sort of display seem to be that they they find it fairly immediately tractable. They sort of they get without much explanation what’s going on, which suggests to me it’s quite nice and it’s something that I would suggest works. And others here can correct me, but but actually can transition quite, quite nicely from a 2D to a 3D to A to an X, let’s call it Excel space. In that.
Speaker7: Okay.
Frode Hegland: So so we agree we we want we want something like that right. Marcus said so. Yeah.
Mark Anderson: Well I don’t take my word for it. I’m just one voice.
Speaker7: No, no, no.
Frode Hegland: But Mark likes it and Peter has done a heart. So then the question is currently when you select an item in the library, the metadata abstract appears to the side on the right. Do we want to continue that and then have the further metadata, such as what Peter was talking about and whatever’s available there, or should we have it left and right? Because one thing we have discussed.
Speaker7: Of course is.
Frode Hegland: These views here.
Speaker7: Quite a while ago.
Frode Hegland: Where we have. So, so many hundreds of these. This kind of stuff where we have different views of what’s happening. My feeling is we keep what we have now because it’s nice and then all quote unquote. I’m sure it’ll be a ton of work, but when we have An item highlighted in the library. And on the right we have this little section of abstracts. To the right. From that we have all these. Lines showing what documents cite this document. And on the left we have.
Speaker7: This. Yeah, but I think we.
Mark Anderson: Need to be. I think we need to be quite pragmatic.
Speaker7: Just a quick.
Frode Hegland: Thing. Andrew, could you please send me a screenshot of the library right now while they talk so we can together look at something pretty, please? But. Sorry. Mark. Yeah. Please. Go on.
Speaker7: No, no, I.
Mark Anderson: And I and this isn’t a this isn’t a sound of panic because I really like this. What you’re showing. But but I’m conscious that that that very rarely do you have that the sufficient metadata in a akin to the stuff that that Peter was mentioning, which is extremely useful if you have it. But the reality is most of those things Mark.
Frode Hegland: A really important thing for the Sloan project is we have it.
Speaker7: But hang on, hang on, hang on.
Frode Hegland: I really need to really need to underline this. What we’re doing is a proof of concept, how interactions can be if sufficient metadata is made available. So we’re using.
Speaker7: The. But I’m not.
Mark Anderson: Arguing against what I’m. What I’m trying to say is that just just bear in mind that what we want to do is in designing this is, is, is to allow for the fact that it may or may not be there. So in other words, if you just have lots of blank boxes, I think it I think it’s sensible to presume that. So, you know, we design we design for the, the full set, but it’s just to be mindful that probably the majority of the time will certainly until we have much richer, cleaner metadata sets for all of this that most of the information will never be there.
Speaker7: Yeah, but we’re not.
Frode Hegland: Designing for that. But but fair enough. But okay. So let’s say so. Let’s say this is what we have today. With the exception of these commands down here, let’s ignore what they are. They just place holder for stuff. It may not even be there. Right. So we want to show who cites this document and what it cites. One way of course is to have this stuff.
Speaker7: The pair in line.
Frode Hegland: That’s kind of easy. But if we have. Because. Sorry, I’m just trying to make this a bit smaller.
Mark Anderson: I would say we can much more easily know what IT sites than what sites it, because the document itself tells you what IT sites. Without without a white a much larger metadata set and without re arguing the point about having a perfect data set. A document does not know what sites it. It’s only it can only know what sites it from a larger a larger scope. Data.
Frode Hegland: But a library. Excuse me, but a catalog does.
Speaker7: Right.
Frode Hegland: So one thing we can do is do it this way then. Right. So we have below is what it sides and above is what sides it. That’s one way of doing it. It’s not the best.
Speaker7: But
Mark Anderson: Time generally flows left to right for most people, I think. So what goes before, what it cites being on the left and what cites it being on the right would seem to be more culturally. Okay. I see Adam’s sitting right back. So, am I saying something?
Speaker7: No no no no no no no.
Adam Wern: I agree, but we have a we have one one, two few dimensions here. Because the list takes up and down space and so we it’s a bit complicated or if we’re going to use the Z direction, the, the depth direction in some way here to have the list or the time dimension visualized if we need to.
Speaker7: Yeah.
Frode Hegland: I mean, what is the library currently organized?
Speaker7: Under I.
Andrew Thompson: It’s not. It just generates in the order that the JSON was created in. We can sort it, but for now, it’s not.
Mark Anderson: Yeah, I think it’s very difficult. I appreciate Andrew’s room, because, you know, you can pick anything and so immediately say, well, that’s not what I want. I mean, the reality is that there are a number of things like date and author are probably date, author and possibly the, the publisher with a very small p and I that’s what dive into what that means at the moment. But I think those are the most likely sort of sorts that you might want. And the reality is it’s it if you had to fix it, it’ll never be right for everybody. So further down the line, it was something. But I think what one could say is it will work on the presumption of there being a sort.
Andrew Thompson: One thing that is one thing is kind of a fun idea with the sorting is if we just offload it to the JSON library. No one will ever blame our software because they just imported a library that wasn’t generated the way they wanted. And it’s like, oh, okay, that’s not our fault. But we could, of course, put sorts in if we want.
Mark Anderson: I mean, the hard part is that dependent on the source. I mean, I know we’re using ACM stuff. The hardest part is, is actually getting the bit of the author you wanted to sort on because name, sort name sorting, it gets horrendously complicated. And there are all sorts of cultural variables you have to put into it. That of course, we normally don’t see because it just works because some other poor person spent a nature on it. Year is. Year is easy. And that might be, that might be a low a low hanging fruit as a default sort.
Frode Hegland: So this is obviously rough and bad, but what do you think of the logic of something like this? A better designed and using the space. You have a main library list off to the side. We have a summary and whatever metadata we have to the right of that. If we have any documents that we know, cite this document. And to the left we have what the document cites using the kind of connecting line, highlighting it in the library as being what makes it coherent.
Andrew Thompson: So why are we printing a list of all of the documents cited? In the preview that we saw earlier that Adam had put together? It was just highlighting them with color. Why don’t we just do something like that? It seemed elegant.
Frode Hegland: The thinking for me is that difference between reading space and thinking space in the library. It is a bit of a thinking space where you have lots of small things. You don’t read long paragraphs of text. So I believe, I feel that while you’re in the library, that’s a good place to see how things connect, and this is one of the ways you can see how they connect. Now, ideally, if you click on any of the documents to the far left or right, or double click or click to open or whatever it might be, that will then be the center document. So you can see what that site’s incited by. So it.
Andrew Thompson: Is. I see. Because you want more buttons then? Gotcha.
Speaker7: No, I want less buttons.
Frode Hegland: No.
Andrew Thompson: No, I don’t mean that in a negative way. I just mean, like you want them to be clickable. And that’s why just highlighting the color doesn’t work.
Speaker7: Yeah, yeah, yeah.
Frode Hegland: What I want is you know, some moving up and down this list and all of this stuff is instant. It’s amazing. And then I get to this one. That’s interesting. Oh, so and so sited it. Tap, tap. This thing is now here.
Speaker7: You know, we go.
Frode Hegland: Around, you do things, and hopefully what this thing here is indicating at the bottom is commands we may very well have. And again, not necessarily visually like this. It may be icons or maybe disappeared. One here, which is sort.
Speaker7: All right.
Frode Hegland: Rotate whatever. So you know that becomes a view as well.
Andrew Thompson: One other thing I think we’ll need is some kind of method for scrolling if we start having large libraries. Because. You know, right now it fits like, say, 50 documents or something, but you start getting into like 300. It’s going to stretch so far up and down that you’re not going to be able to realistically see the top and bottom documents. So we’ll need some kind of UI that scrolls it, whether that’s another handle or if that’s an actual scroll bar that you drag. Whatever we decide. Or we can just choose to never make a library that large because of load time.
Speaker7: But well.
Frode Hegland: Here’s the brilliant stuff that Adam designed. To make sure you know what you’re looking at. Andrew. The purple one is the document we’re reading about here. So everything younger, of course, is what excites everything. Sorry, older. Everything younger is what sites it, so you point to it. You get a bit more information about what that is.
Speaker7: Very good use of space.
Frode Hegland: So I’m wondering if we can take some of that logic with the scrolling and with the kind of. Lensing effect or whatever.
Speaker7: Because this.
Frode Hegland: Looks nice. I mean, this is awful. This, you know, visually, this is absolutely horrendous.
Mark Anderson: The different exploratory modes. In a way. The interesting thing is going back for a second to the other display, the the sort of like the temporal line. The top part of that, to my mind, lends itself to an XR space, because you can just make that as a sort of big as you want. There is an element of being able to think to actually zero and the thing you want. But but for instance, in, in an XR space, the, the abstract, the bit at the bottom of that display could in essence be a floating panel closely to you, but you could have a very large listing space, as it were, and the sort of those subtleties of highlight or color or things would allow you to, to effectively explore that. But it’s a different sort of exploration, that one to the, to the one that’s on screen at the moment. The first one I think is, is interesting in terms of trying to see where, when you may be looking through a corpus or something. Just to sort of see, you know, is it, is it just a singleton there that just exists in the corpus or is it something that’s richly interconnected? And it lends itself to all sorts of potentially further filters on it because you could, you know, say start to filter people without don’t take this as a feature request, but I’m just thinking what it will then allows you to do in due course, which is to say, filter on the number of of connections.
Mark Anderson: The view probe was just showing I think is is perhaps more deliberate in terms of looking at the, the actual citation tree and only that in terms of one thing removed. One of the things that does immediately say to me is you definitely want a history because you’re going to want to undo. Otherwise you hit the Wikipedia problem. You’re thinking, how on earth did I get here? And what was I doing when I started? And that might be that you in two forms. It might be that you literally want to go back one, or you might want to allow for having a sort of a start point in almost like a waypoint. That actually where I want to do is I want to scroll back 50 explorations. Because what I was doing when I started this was I was looking at paper X, and now I need to get back on track and go back to that. So those those two, those two sort of history points those two, those two affordances, both both run off having a history. If that makes sense.
Speaker7: Yeah.
Mark Anderson: And in this sort of despair, I think it’s interesting. I know things like Tin Box used to have a history, and it got rid of it because people said they never used it. But I think in this sort of exploratory mode, a history is innate. You’re very much in a browsing experience. So the history is relevant going backwards and forwards. And you have you have less you have less sort of noise of, oh, I click the wrong thing. Well, that doesn’t really matter. Because when you’re actually dealing with the the documents that lie behind all these list items, it gets slightly different because often you you’re looking at something, but you’re not really visiting it, if you see what I mean. And that’s something that the user interface always struggles with, because understanding the user’s intentionality in looking at something can be, can be, can be hard to guess from their inputs alone. But I think when you’re navigating around what is essentially a tree, I don’t think that. I don’t think that’s problematic.
Speaker7: Right.
Frode Hegland: So so one thing we can do. This makes the user work for this. Meaning so I’m just going straight back into this thing. So imagine we’re in this view. And the user has to do the thing of clicking a button. One will be, let’s say, open. So it’s explicit. Another one will be. Or whatever you click on citation links or whatever. And it does this. It hides the library. And we still have options here. Like open.
Speaker7: And then we have something.
Frode Hegland: Like back to library. Because now that you have this explicit, we may even title it.
Speaker7: Something like this, right? With it.
Mark Anderson: They seem just. Are they needed? I mean, in a sense, if if you if you’ve had the, the view explained which you would need the wants, then that’s just extra noise. So extra text rather you don’t really need to read.
Speaker7: Well I mean.
Frode Hegland: You said that left to right thing. You know, that’s in a Western alphabet. Some alphabets do go the other way around.
Mark Anderson: No, no. But I mean, to use the thing the first time, you would need to know that you know something. The first time you see this, you will just say, you will see something. We’ll show you. We’ll show you a document, and then it will show you the things it cites and the things it’s cited by. I don’t think it then becomes relevant to or necessary to label them. You can have them read other way around if you want, and that would be a preference. But that’s way down the line for what we’re doing here. I just don’t I’m not sure that you need to have more text because it’s all something else you have to read, and the more stuff you have to read that you don’t need, the more the load.
Speaker7: Okay.
Frode Hegland: So that that’s fair enough. In many ways. But also if we look at and this goes way back to our initial discussion of menus and options, let’s say you’re in this view and you’ve decided that this reference section is something you want to work on. You want to be able to tear it away. So to have a title area which may be then you select somehow gives you affordances for this. Maybe instead of having the bar on the side to move it around, maybe the heading becomes the thing we move by. I’m just spitballing here. I’m just saying don’t necessarily think. Because also they will be like this by default. But maybe you want to do this for your workspace, so then it’s nice to know what it is. Were you waving at us, Adam?
Adam Wern: I’m waving to Andrew, who had to head out now.
Speaker7: Oh, okay.
Adam Wern: He left.
Speaker7: Oh, yeah. Need some.
Adam Wern: Water?
Speaker7: The rest of the time? No, I take your point.
Mark Anderson: I mean, that that’s moving ahead of where I thought we were, because now you’re. The difficulty is I could see 101 ways in which one could customize all these things. But if we were starting with the notion of being able to just see something in our library in the relation of its inbound and outbound citations, then I don’t think you would need by default to show the the labels for them. I don’t think you need the labels as a, as a selection point because as we’ve shown, you know, some of the other views, you can point at it and you can just show a effectively a bar that you can use as a target, though I totally accept that if you are to start to move the items around, you may wish to have a label to remind you what they were. So the information can be there, whether it needs to be displayed is, you know quite contextual, but I wouldn’t say in the, in the, in the initial case, I think less is more in terms of.
Frode Hegland: I think in the initial case more is more because it. So when I get lost we can minimize over time. But I think it is good to have labels to begin with.
Mark Anderson: I’m not convinced, but I’m happy.
Speaker7: To put Mark, if.
Frode Hegland: You can move this in different places and pin them, maybe you want to have 3 or 4 different reference sections. You should be able to do so.
Speaker7: And you know, but but there.
Mark Anderson: Are lots of yes, I understand, but there are lots of ifs here. So we’re now talking. We’re now jumping from a view to a highly customized version of the view. And I just wonder if we’re getting ahead of ourselves, because the ability to do the latter, I think, is some way down the line.
Frode Hegland: But for us, it’s obvious that the left is one thing. The right is something else. I don’t think we can assume. No, I.
Mark Anderson: Didn’t say it was obvious, but you will need to describe for the first time that people use things. What they are. Why would we.
Speaker7: Would we expect.
Frode Hegland: Them to remember it for all futures?
Mark Anderson: Well, I. I’d be very well. I really doubt that someone who’d used this once would actually forget which is which is what? And anyway, you can tell by the dates. Because if something is if something is citing something, it has to have a later date. And unless you’re not showing the dates, I mean, that’s the most obvious way you can. You can work things out. So, you know, it’s self-evident from the data there, I would say it’s.
Frode Hegland: Self-evident for someone who wants to read the dates and compare the dates between three different places. Okay. Anyway, Peter.
Peter Wasilko: Yeah. Maybe we could just have a drop down menu over each side column, and then the user could pick what they want to have on each side, and whichever menu then automatically be the label.
Speaker7: Yeah, well, this sort of do double duty.
Frode Hegland: This is exactly the kind of discussion I think we need to have. Yes. On what kind of What means through which we provide the options and and how we do the labeling. Absolutely.
Speaker7: What was like.
Frode Hegland: This on the right here wouldn’t have numbers. So just for clarity, I’m getting rid of the numbers.
Peter Wasilko: Well, it’s 1:00. I’m going to have to drop now.
Speaker7: Okay.
Peter Wasilko: Then. Monday.
Speaker7: See him under. What are your.
Mark Anderson: My thoughts earlier were driven partly by when I’ve shown when I sort of shown people the, the sort of two source visualizations which sort of talk about here that Adam did. The one the question that’s never come up is people not understanding, which is what. Which is why I just. I just think, you know it’s about paring back because there’s a lot of there’s a lot of data in here. So the less that we show that we don’t need, need, I stress need to show, I think the better. And it’s kinder to the user.
Speaker7: I don’t think.
Frode Hegland: It’s kinder to the user. I think that’s a kind of a blanket statement to make, Mark that, that that yours is kinder than what I’m suggesting because, well, okay.
Mark Anderson: But I’m only just going off the experience of showing thing, you know, these things are cold to people. And so it well, you know, experience is what it is. So it’s worth doing. I mean, this is why you show things to people. If that means nothing, fine. But I, I’m not I’m not going to.
Frode Hegland: I’m not saying it means nothing, but what I’m saying I’m my my software is demonstrably minimalist, so I clearly agree with the notion of not having clutter on the screen. There’s no question about that. What all I’m saying is, at this point, when we’re still debating, deciding and experimenting how to put options, how to display information, I think we should be careful about pruning back. I think it is useful to be able to say these are references, and this is what it’s cited by. I don’t think it’s a big thing. And if it turns out that later on we don’t need it, I’ll be very happy to get rid of it.
Speaker7: I mean, I think.
Mark Anderson: I’ll restate my thought. I think it’s way too early to be worrying about this level of detail.
Speaker7: So why so.
Frode Hegland: Why are you so worried about it? Then? Why can’t we just leave it?
Mark Anderson: Well, because it keeps being dragged out, so. No, we must have it. So I’m. If if that’s what we have to discuss. I’m just responding. Yeah.
Adam Wern: It’s you who’s pushing this. You push this into the library discussion. So I don’t think it’s fair to to criticize Mark for that.
Speaker7: No, no, I’m not.
Frode Hegland: Criticizing Mark for having the library discussion. I’m just saying.
Speaker7: No, no, no.
Adam Wern: But that is not what I meant.
Frode Hegland: And I’m not sure we need to remove it, so.
Speaker7: But but I.
Adam Wern: Think there is a valid overall discussion in on as clearly we fitted the whole ACM hypertext set on the regular screen laptop screen. So one question is what does Xoar do here that can’t be done or what does XHR do better. So there are some kinds of volume things for example, that will display even more documents. For example, it could be that it displays a history that it all your interactions build up for working session that you start with a library. You pull out a few documents, follow the references. And that leaves a visible kind of a browsing session. So it’s offloading instead of having lots of back buttons and undo history, we could actually see where we came from, where we went, and kind of build up a traversal of the information space, kind of spatial hypertext in that you lay it out, you follow this reference, you saw this paper, you took this paper and put it in the in the pile, this paper in the pile. You went there, then you went back again because you didn’t want to follow that trail. So that’s one way of visualizing time, your interaction and for this session or a couple of sessions, but for a, let’s say, a browsing session that could be spread out in time as well.
Mark Anderson: Another thing I know comes up that I mean would would, will hopefully filter through the use cases is Aside from whether I have this document or not, and we tend to assume that all these things are available to us, is that whether I’ve used it? So, for instance, if I’m looking at something, have I have have I have, I have, I used it, I’ve heard the word cited it. But in in other words, have I used it in, in a past paper I’ve written. So is it something that has has another aspect of importance to me? So what that what that element is showing is, is this new essentially is this thing that I’m now looking at new to me, or is it part of an existing corpus that is. So there’s some there’s some sort of implicit knowledge or recognition there. That would be way too difficult to feed into the original things, plus the fact the metadata wasn’t there to show it if I do know, because actually, Adam Carney asked me, would I? One of the things you put in there I remember was we colored up some special hypertext papers, but that’s sort of an example of really what I’m driving at is. Yeah. And the key, if the, if key wording were better, which I don’t think it’s going to get better any time soon in terms of papers. Then that would be tremendously useful as a way, again, of visualizing stuff.
Adam Wern: Yeah. But the lowest hanging fruit there, I think, is collections, loosely defined collections, either by you that you want to pick out the things you you basically a list or a collection, maybe with some list annotations as well, where you say why you have this list or why this this item is in the list. But as its most basic set is, it is a named named list, or maybe not even named, but a pile of things. But to be able to send it and send it to others or remember later where it was from, I think it needs some light annotations, titles or annotations. And so it’s that doesn’t really need a good keyword system. It just needs some sort of intuition or what you put in each pile. And I think that that is something everyone wants. We have folders, we have desktops, we have PowerPoints. We’re pushing, we have reference Spanish where we do it. Making lists or collections is one of the most basic things in all, all kind of library or knowledge management. And I think there is much to be done there that could we’re actually XOR. I remember when using our prototype the the flat screen version here. It’s a I wanted to collect papers along the way, put them into a pile, and with Xoar, we have the opportunity to actually see all the things we have in the pile as well. Much more. So even if the whole data set could be visualized in a very compact way, like lights lighting up like we did before, the actual papers could could pile up and in a visual way and actually be represented visually on screen and to be taken to another device later on, or what.
Mark Anderson: You’re describing is, is an essentially a form of exploration. We’ve many of us probably would like to have done, but we couldn’t do before because we didn’t really have the interface. So now and making some assumptions about data, but but now essentially we, we, we have that data. We can put it essentially onto the big screen. So going back to Brandon’s comment about just having, making insanely wasteful use of space, we can do that. We can have a very big screen. And obviously there’s some accommodation has to be made in that the bigger it gets, you know, the more we have to think about sort of being able to focus bits we need. But it’s absolutely possible in a way that as Alice rightly said, you know, there is a problem with, with a finite center of a browser window where you can only put so much in if you make it too small, you just can’t see anything. But I think and one interesting part, again, thinking of the well, actually both of those both of those visualizations, is that the the thing that’s in focus actually is something we could, in an Excel space, pull out so that can be closer to you, sort of readable and big, and then you can have a much bigger hinterland of, of this related of this related data related by be it by citation or or membership of, of some other set. So I think that’s, I think that’s a, that’s one massive win. In relation to references that you get with Excel.
Adam Wern: Hold on. I have guest apparently I need to see what guest I’m here but I’m muted for it okay.
Frode Hegland: Yeah. No, it’s It’s past the hour. So yeah, at some point we need to decide where to have buttons. But I guess that was not for today.
Mark Anderson: Well, I think a useful thing that we seem to have have, you know, you’re absolutely right. And something I think that would make that easier to discuss. Fred just mentioned we need to think about buttons in the general sense of the word, not literally skeuomorphic buttons, but, you know, things we can interact with. And I think that’s right. But one of the things that today’s reminded me of is and maybe I’ve missed it, but I think we’ve we’ve whizzed past a point where we are just had a, had a sense of the. No, the addressable parts of a document. I know that’s sort of angels and pinheads because you can keep drilling down, but but there definitely are a number of or there’s an almost like an explicit map of the document. So there are things like, you know, reference references are clearly and there an obvious one to think of and things like the abstract, but because in a way it’s in a, in a hypertext of the document. And as many of the things that we are visualizing are sort of lists and outlines and maps having a greater view. And why do I say that? Because it then links to, well, what are these buttons going to do? Where are they? Where are we going back to and from with these buttons. Otherwise we will just be looking at things like editing and annotation, which are fine, but not that, not the whole of the users need.
Frode Hegland: I agree with the importance of addressability, but practically with author the way I’ve been doing it on the different platforms you have global and then you have selected when it comes to the documents, not when it comes to the references necessarily. When it comes to the selected, that’s not always useful or easy, but
Frode Hegland: And also, do you mean something like you have a cursor and something and you tell the system, I want to have a command based on this sentence or this paragraph. What do you mean by it?
Speaker7: No, I was.
Mark Anderson: Thinking about so you were showing an example of of, say, some text. And I was just thinking, yeah, I’ve, but I was thinking I find it hard to think about what that might be because I don’t know why. I don’t know what this text is in relation to its source is how how much of it is why I opened it. Because why I opened it will dictate what it is I think I want to do with it. So just just seeing just having a box of some text in isn’t, to me meaningful as a starting point to work out what what the controls would be.
Frode Hegland: That text is your document.
Speaker7: Yeah.
Adam Wern: But it could be interesting to go back to. We talked about that fraud in a, in a life design session where we brainstormed around that to pulling out parts named parts of a of an academic document, pulling out the abstract and conclusion part, for example intro, abstract, conclusion and gluing them together and stripping away everything else. Perhaps put it, or having those as toggles where you could have a view like a different parts of a standard parts of an academic document that are either extracted from good data, like a very pure text or slightly handheld either. Perhaps an LLM can assist assist in finding it, and you adjust it per document once for a, for a just a single time per each document that you. Preview what it what an LRM thinks is the abstract, and then you say, okay, thumbs up. This is abstract, this one. And then you have an object that is.
Speaker7: I agree.
Frode Hegland: On that. But when it comes to. Being able to issue commands. I don’t think that’s relevant because you will be issuing relevance based on something. You can tell the system what it is, right. And when you’re reading a document, whether you’re reading a background chapter or a specific chapter, you know it’s basically text in the document. You will select a group of text to tell the system what you want to do. Do I agree that we should have views, for instance, where each chapter is a vertical scroll, for example? Absolutely.
Speaker7: Right?
Adam Wern: Well. Well, I think we’re talking about different things, perhaps. I’m talking about the library view or library list. And in relation to document specific kind of previews or snippets from that document and metadata snippets from the document and perhaps references as well, but they are more clear cut in what they are. But if we want to have a, a way of browsing a library, of which some papers are a bit unknown to you, you don’t know them by heart. Most papers are perhaps like that, you know, some by heart, but most of them are in your school, squirreled away collections, things you’ve downloaded but never read there. You want a good way of getting a sense of what this document is about? The summary is one thing. The abstract is one thing. Perhaps keywords, author and institutions or of course very important. But there are kind of views into the document. There are small, small snippets from the document that need to be extracted, either by good metadata or by by hand, or perhaps an LLM where you approve of where you approve the previews and store them, persist them. So that is what I’m talking about.
Speaker7: I agree with.
Frode Hegland: All of this, but I don’t understand why we can’t start.
Speaker7: You know, we have.
Frode Hegland: A pop up venue in the current version, which is not very good either in terms of.
Adam Wern: Which view in the library view.
Speaker7: No, all of them.
Frode Hegland: It’s the same kind of pop up. It’s in the main library. Not not not in the not okay. It’s not in the library view. It showed up. It was a mistake that it showed up. It was a rendering error, but okay.
Adam Wern: Yeah. So not in the library view because there we have a list and abstract or something list and intro right now.
Speaker7: We. But I’m not sure. What’s going on in.
Frode Hegland: This discussion because. We will need the means through which we can tell the system what we want it to do.
Mark Anderson: And we’re going, okay, we’re going around in circles because I understand that the thing that you, that it shines out, that you want to do is to start, is to start listing the interactions that you have for certain blocks. What what’s being brought up is the fact that what those blocks are is not necessarily clear. So in order to understand what.
Speaker7: The blocks.
Frode Hegland: Are, we’re talking about library now.
Mark Anderson: No, you. Well, the thing that you keep going back to is wanting to be able to know what you should do with a window of text that seems to where we keep being rotating back to. And that was the picture he’s showing us earlier. I understand that, and the point we’re trying to make is that what goes yes. You want affordances. And the style of that is something that that we’ve had some useful discussion on. I mean, you know, the, the in terms of the intensity you give it and the way that you think about foregrounding and backgrounding and things. But exactly what’s on there are is they’re just they’re just labels to buttons that point to something we don’t yet know because we haven’t properly discussed why you’re looking at that, that text, something that this really shines out in a sense from the, the, the use case. Not everyone uses documents the same way, even within the even within the context we’re discussing. So understanding why you’re looking at that piece of text is actually very relevant to what goes on there. Otherwise you end up with 64 options which you don’t use. So and there won’t be a one size fits all. What we really need to understand is so is to not tie the button labels too clearly into the buttons, so to speak. If we can usefully describe the buttons, I wouldn’t worry about what they point to at this point, because we just don’t know. We haven’t we haven’t, we haven’t said.
Speaker7: What.
Frode Hegland: Do you mean commands or text?
Mark Anderson: The the commands because I mean, there are there are there are a small number of things that are given.
Speaker7: Mark, Mark, Mark, Mark, Mark.
Frode Hegland: We do have specific use cases. We’ve gone through them with Denny. They’re listed on the website. We you know, the first one is reading documents to prepare to write an academic article. So we do know some of this, right? And there are some commands that we do know that we want we want, for instance, when you’re in the library, you want to be able to open a document.
Speaker7: But that is.
Frode Hegland: One thing we do know.
Speaker7: Yes, but.
Mark Anderson: That’s not what we. That’s not the interface you were showing us. Something was to do with having a piece of text and what you might want to do with it. It was about asking AI and stuff, and I’m not sure we’re there yet. There are the senses. If you have a bit of text, you might want to the senses. Are you going to take it and use it somewhere else? Are you going to annotate it in the loosest sense of the word AI? You might be tagging it in some way, or you might want to highlight it visually or something, but let’s class those as annotations. There are a whole grab bag of things, which include things like AI summaries that you can do, but those are add ons that that some people might want, others might not want. They’re not necessary parts of it. I mean.
Frode Hegland: There you you cut me off. Yes. But there are some things that we do know. Okay. If you’re looking if you have a piece of text on your Non-library reading area, as we’ve discussed today, you will want to have things such as should it have a white background or not, which we may call framed for the sake of argument. Also, should it be foregrounded or should it be in the background space? These are interactions that we know we want other.
Speaker7: Yeah.
Adam Wern: And we and we covered them today. But what we’re in the things you’re kind of leading the design on here. We haven’t really specified what use case. We can’t just say all use cases. It must be a specific or 1 or 2 use cases perhaps could be combined into one. So when we’re when we’re designing this we’re designing a bit blind. Or maybe you know which use case you selected from the long list of this lone use cases or use cases, but that hasn’t been discussed at all. Or we researching a paper, then we want to to skim the collection in a certain way. Or we reading our own favorite papers. We need very different buttons appearing on different abstracts or references. So that is the essence of the kind of confusion we’re designing without the use case.
Speaker7: Yes. Yes.
Frode Hegland: Okay.
Speaker7: Hang on. I’m going to.
Frode Hegland: Stop recording for a second. Yeah. Sorry for the interruption there. So the thing we’re talking about is there’s one thing about interacting with the text according to a use case, which is really important, and we need to have much more discussion on. There’s also the mechanical interactions or where to put something foreground background and so on. And we need to look at interactions for that. I’m being simplistic and just putting text labels. Adam wants much more intuitive control, such as gestures and icons and that stuff. I completely agree that these approaches need to be experimented with, and I hope that next week we can continue. Hopefully some of us can do some of this offline. And I’m very, very grateful. This is a very, very difficult discussion, but it is an important discussion. And and from my side summary I would just like to thank you for this. Any other comments from you guys before we go? Evening time. All right.
Speaker7: Evening time.
Mark Anderson: In slack.
Speaker7: All right.
Chat log:
16:25:33 From Frode Hegland : https://thefutureoftext.org/invitation/
16:25:38 From Frode Hegland : https://docs.google.com/document/d/1Uv60IHr7sjNSldzeSYCAZwCn7nHsphrk5iJyNhKY5Z8/edit?usp=sharing
16:25:44 From Frode Hegland : https://docs.google.com/spreadsheets/d/11_Wo7x6gNtuNdFnJ8xi7_41dBvhGQm-Ropo91_baYbM/edit?usp=sharing
16:25:46 From Adam’s iPhone : Where is the agenda?
16:29:50 From Frode Hegland : https://thefutureoftext.org/invitation/
16:30:13 From Frode Hegland : https://public.3.basecamp.com/p/q6mJPiQzKmTXGvB3pFAXUsFM
16:38:05 From Frode Hegland : https://www.shoantel.com/proj/acm-ht/visualisations/index.html
16:38:18 From Frode Hegland : https://www.shoantel.com/proj/acm-ht/visualisations/demos/index.html
17:06:41 From Rob Swigart : Speaking both as a naive user and a former manual writer, I need a manual to use this project. It would be a work in progress of course. 1. How do we get things into the Vision? Inside it, I’ve been navigating from the FoT site, but it is not very intuitive and I’m not sure I’m in the right place. Inside I would like a list of gestures and what they do. A glossary of terminology ( focus, Detach from Group (what’s a group ? what does detach do? Where does it put detached things, or what can I do with them… This is getting quite rich and potentially interesting even to me. I’m pretty impressed by how much has happened, but it raises a lot of questions I think for the naive user.
17:07:45 From Andrew Thompson : That’s great insight Rob, I think a manual is a must as the development continues.
17:07:52 From Peter Wasilko : I would really like to see us adopt a Literate Programming methodology.
17:12:13 From Rob Swigart : The transition to XR is complex and quite different from what we are used to. The mechanics environment asks for a lot of relearning. Somehow we need to ease it. Now I’m afraid I have to run. See you all next week.
17:43:13 From Andrew Thompson : Library (catalog?) screenshot
17:55:23 From Andrew Thompson : I’ve got to head out, great discussion today!
16:03:13 From Frode Hegland : ou are an expert note taker and analyser of meetings. We have two weekly meetings where we discuss the future of text, in particular as related to XR (Extended Reality, also referred to as VR, AR or Spatial Computing). Part of our work is funded by the Sloan Foundation. Please this transcript and provide me separate lists for all of the following. Do not show bullet points or numbers, just list one item per line. I also need to add a blank line after each of your headings, as well as an empty line after each time you discuss a speaker.
List of participants in a line, separated by comma. Do not put a heading on this section.
General one paragraph summary. Do not put a heading on this section.
16:03:26 From Frode Hegland : What was important for each speaker? Make this heading ‘What Speakers Expressed as Important’.
Summary per speaker in a paragraph per speaker. Make this heading ‘Speaker Summary’.
What agreements and disagreements were there? Make this heading ‘Agreements & Disagreements’.
Add a heading of ‘Topics Discussed’ here please.
What was discussed regarding WebXR?
What was discussed regarding Sloan?
What was discussed regarding Visual-Meta?
What was discussed regarding visionOS or Vision Pro development?
Were other topics discussed?
16:03:32 From Frode Hegland : What named entities were mentioned? List command separated. Make this heading ‘Names Mentioned’.
Did the group seem to get any new understanding in this meeting? Make this heading ‘New Understanding?’.
What would you suggest as an agenda for the next meeting? Make this heading ‘Suggested Next Agenda’.
Thank you for helping me put together this very useful summary. This will be used by us and others in the future to better understand who we went about our work any why we built what we did and came the understanding which we did.
16:25:33 From Frode Hegland : https://thefutureoftext.org/invitation/
16:25:38 From Frode Hegland : https://docs.google.com/document/d/1Uv60IHr7sjNSldzeSYCAZwCn7nHsphrk5iJyNhKY5Z8/edit?usp=sharing
16:25:44 From Frode Hegland : https://docs.google.com/spreadsheets/d/11_Wo7x6gNtuNdFnJ8xi7_41dBvhGQm-Ropo91_baYbM/edit?usp=sharing
16:25:46 From Adam’s iPhone : Where is the agenda?
16:29:50 From Frode Hegland : https://thefutureoftext.org/invitation/
16:30:13 From Frode Hegland : https://public.3.basecamp.com/p/q6mJPiQzKmTXGvB3pFAXUsFM
16:38:05 From Frode Hegland : https://www.shoantel.com/proj/acm-ht/visualisations/index.html
16:38:18 From Frode Hegland : https://www.shoantel.com/proj/acm-ht/visualisations/demos/index.html
17:06:41 From Rob Swigart : Speaking both as a naive user and a former manual writer, I need a manual to use this project. It would be a work in progress of course. 1. How do we get things into the Vision? Inside it, I’ve been navigating from the FoT site, but it is not very intuitive and I’m not sure I’m in the right place. Inside I would like a list of gestures and what they do. A glossary of terminology ( focus, Detach from Group (what’s a group ? what does detach do? Where does it put detached things, or what can I do with them… This is getting quite rich and potentially interesting even to me. I’m pretty impressed by how much has happened, but it raises a lot of questions I think for the naive user.
17:07:45 From Andrew Thompson : That’s great insight Rob, I think a manual is a must as the development continues.
17:07:52 From Peter Wasilko : I would really like to see us adopt a Literate Programming methodology.
17:12:13 From Rob Swigart : The transition to XR is complex and quite different from what we are used to. The mechanics environment asks for a lot of relearning. Somehow we need to ease it. Now I’m afraid I have to run. See you all next week.
17:43:13 From Andrew Thompson : Library (catalog?) screenshot
17:55:23 From Andrew Thompson : I’ve got to head out, great discussion today!