Meeting 1 April 2024

1 April 2024

Frode Hegland: Try not to close the door. Try not to play right outside the door. Okay. I. Hello, Andrew.

Andrew Thompson: Hello. How are you doing?

Frode Hegland: Good, good. One of the questions is who is h r I t h I k? Is that one of the people who can ask you?

Dene Grigar: I was going to ask you, who is Hitha Tiwari?

Frode Hegland: I think it may be somebody we met on yesterday, Sunday. May. Hang on. I’ll be great. It’s quite a unique name.

Dene Grigar: There’s also Autopilot and Firefly. So, Heather, can you put in the chat? You joined from yesterday? Thank you for being here. It’s delighting. Delightful to have you here.

Frode Hegland: Oh, okay. Oh, so you’re a human as well. Okay, cool. Because all I’m saying is I. Okay, I see you’re separately. I just saw the the note taker. Absolutely. Fine. Absolutely fine. Okay, now we know who you are. Welcome, welcome.

Dene Grigar: Hey, Heather. Now you can see the computers in my home office. We were talking about the Mac lab yesterday.

Frode Hegland: It were indeed.

Dene Grigar: We got a bunch of them here.

Frode Hegland: I’m just checking a build thing here. While we’re waiting to see who will join us. I think we’ll be quite empty today due to the holiday. Is that a holiday in America land today as well?

Dene Grigar: We don’t get a holiday on Monday. We don’t. We’re not allowed to have these holidays anymore. Yeah.

Frode Hegland: So all of these are

Dene Grigar: That’s why Cindy is so precious to us. I mean, that’s the only day we get off. And I even had people writing me yesterday. My my bosses, my supervisors writing me for stuff yesterday. I was pretty angry. It’s like it’s Sunday. It’s Easter Sunday, for God’s sakes.

Frode Hegland: A whole new level of. What the heck? Right. Okay. Yeah. I don’t know who else is coming because of a European holiday, but I’m going to share a little bit of slidey thing here. First of all. Fewer things to go through today. Of course, today isn’t an agenda, but we have a kind of an agenda. This thing here. So that’s not what I wanted to do. I just wanted to be able to see you guys. Okay. We. Need to That’s the second slide. That was very confusing for me. So I think and all of these points are up for discussion. But I think that by Wednesday this week, including Wednesday next week, we need to have the use cases done for reading. You know, we spent time talking about it. I think we’re pretty close. Any objections or comments? Well that’s good. So that means that if you have a use case that you want to add to it, you know, please do. We’ll talk about it again on Wednesday. Danny has outlined exactly what it should be. Yes, Peter.

Peter Wasilko: Yeah. I was wondering if we might want to think about email integrations too. A lot of documents I get via email and then people comment on email.

Frode Hegland: We can’t extend the scope. We can talk about that in this Monday section. But when it comes to what we’re doing for Sloan, absolutely, we can’t afford to add to the material. We use it. Please continue. But please just keep that in mind.

Peter Wasilko: Yeah, I was just thinking, if you just. I guess it’s too far out of scope. Now. I have to give it more thought. And maybe in the next round of funding.

Dene Grigar: I say, can I say this? I think it’s a great idea and I think there’s a lot of. Features we want to add to this. And I think that we should have a list. I mean, we do have a list of things we want to add. So that should go on our, you know, what do we call that stretch goal. But right now I want to make sure that we got to make sure that what we’ve got done is in good code condition so that we can continue building on what Andrew’s done so far, which has been great. Bravo, Andrew.

Peter Wasilko: I guess what we really need at this stage is just to have some kind of an anchor associated with each document that could be referenced from other media, so that down the road, if we want to bring it in, we have a official designator that we can use and plug it into email or an RSS feed or whatever, that the system would be able to associate the PDF that we’re reading in the environment with whatever other metadata is coming in from whatever other source.

Frode Hegland: Yeah, I agree obviously with Denny because I always do. However, I had an interesting insight this morning. I saw some video, somebody was doing some work, and I noticed on their screen sharing they had 300 unread text messages on their computer. I don’t have any unread text messages on my computer. Text messages are very different. Click on the person, send a message. You’re done. Email is really clunky. I think you have to find a type in their address because it’s not based on the person. You have to do the subject and la la la and then threading is always bad. So my hope for the community is that we’ll start communicating through written stuff, maybe start with PDF, go over to other things. But I just think email is absolutely horrible. If we could improve it, we should. I agree with that, Peter. But yeah.

Dene Grigar: Can I say this though? Frodo, I really hate text messages and I it’s so daily I get it message from my people that follow me on Twitter, some Twitter messages, Facebook messages, text messages, email, slack. I mean, it’s I’m all over the internet just to piece together what people are asking of me every day, right? So I’d rather have emails so I can archive it. And I’ll know when I go back and make a like, for example, when I start the the report that I have to do in six months. Right. In three more months, I can go through email because it’s all archived under Future of Text Project. It’s hard to go back through text messages.

Frode Hegland: Yeah, no, I agree with you, Danny. Absolutely. And I do personally use different messaging for different reasons, and I’m sure we all do. And this is why I’ve started unsubscribing all kinds of nonsense emails that I didn’t care about before. But anyway for now, Peter, email is out of scope. So when I’m talking about use cases, I do mean reading an academic document either PDF or rendered in HTML and what to do when reading it. I have a few suggestions on the next page just to kick off a little bit of a background discussion. And then also, I think we should think a little bit more about really using the space. I’m glad you’re here, Brandel. Interesting frustration that I’m having at the moment is I can put all kinds of things with native vision apps all over the room, which is brilliant for spatial memory and so on. When I do that currently in webXR, because there’s no pass through, I can’t move. I’ll probably knock over something. And that’s a really, really interesting design. It’s just a reality right now. But we must somehow design for richer environments.

Frode Hegland: We also need to, for Wednesday, at least, have a few more people on the list for invited speakers and authors. We just talked about code review, so Andrew will go through maybe with you, Randall, if you have time. Fabian is also said he would be willing to not to, like, check the quality of the code. Of course, the quality of the code is as it should be going through. No, no, as it should be for having done all these wild detours and changes. Now, the question is, what should be the next steps forward? So if the code isn’t quite a mess at this stage, you haven’t done your job right. You know, we’ve been experimenting. It’s the whole point. So it’s not a code optimization I’m talking about. I’m just talking about some perspectives because it’s three months. Take stock and then we should if there are other design proposals talk about them now. Just to kind of kick off. What? Oh, yeah. Sorry, Peter. I ignored your finger hand thing. Please stop. I’m sorry. I just have.

Peter Wasilko: One final thought on the email score there, and that’s that. Let’s forget about reading email and dealing with email threads, but a use case that I have when I’m reading a document would be able would be to be able to shoot an outbound email, which should be a lot easier. So we don’t worry about anything coming in. But if we could just simply while I’m in the reading environment, be able to draft a little note and email it out to someone on a pre-designated list, and you could have a JSON file or something that had all of the names attached to the actual email addresses. Sending email should be a heck of a lot easier and much more narrowly scoped, and if we supported that, it might be a good teaser to get Sloane to say, gee, you can send an email from the environment. What if you were able to? Maybe they might even get the idea of giving us some more money and telling us to explore in that direction? Because there probably is drowning an email as I am. I spend about an hour and a half a day just going through processing all of the slack and newsletter subscriptions and subspace emails that come in. That’s a huge amount of my academic workflow is just spent triaging academic material coming in through email channels. But if I could send an email outbound from inside of the reading environment, that would be really invaluable to me. Otherwise I’d have to write it down. Hope to send an email later. You just sort of lose the immediacy of being able to shoot the email. Note oh, Brundle should see this. Let me shoot him a little email with a link to this PDF while I’m looking at the PDF that I don’t lose that train of thought and forget to ever do it once I leave the reading environment.

Frode Hegland: I’m just going to check something, okay? I agree with your issue. I’m not sure about the solution. You should definitely be able to share with the community thoughts and comments as you’re reading. So that’s definitely part of what we’re thinking, especially for next year. So Brandel, you you walked in, just as I was saying. We have a few priorities for this week, so we’re just going through that. The discussion today will go as a normal Monday. So when it comes to use case, I put up just a few things just to get a gist of what we’re talking about. Danny, because you’re the the pro academic. Just skimming this. Do you have any comments? But the main topics.

Dene Grigar: Well, I had already outlined them in a message to everybody before. Right?

Frode Hegland: Oh, I thought you hadn’t finished your.

Dene Grigar: Well, I haven’t finished the article, but I gave a list of the things that are, you know, made up. I made a list of everybody like this is. Did you put it on that back out of the archive.

Frode Hegland: Slack or email? I’ll.

Dene Grigar: I’ll pull it back out of archive and I’ll share it with everybody again.

Frode Hegland: But these are the kinds of things we’re talking about, right?

Dene Grigar: Kinds of things. But that’s not the list I gave. I did give I did give the preamble of what I’m writing, but I’m happy to find it and reshare it. It’s in my files. Okay.

Frode Hegland: Because we’re kind of Yeah, we do need to have our use case done. Otherwise we’ll be. Flailing a bit.

Dene Grigar: Well, it’s been Easter holidays and I’ve been working on a lot of other things, and I told everybody I’d get to it as soon as I could. So I’ll get to it as soon as I can. No, no, no.

Frode Hegland: Danny, that’s not what I mean. What I mean is you are writing a piece on it, which is perfect, but it really is an opportunity for all of us to go through our different versions of it. So I’m very happy to find what you have written and put that in. But you are the only one who has worked on this so far. So it kind of into the community. Everybody should have a little think of it, that’s all. Peter.

Peter Wasilko: Yeah. One other thought I had sort of a deep systems architecture level. If we could think about having some sort of a plug in mechanism so that we could go off and write little bits of code, that Andrew could just sort of drop in and integrate with his work without him having to tackle every single task himself. So if there’s just some sort of a very, very simple, minimal plug in API that we could build into the infrastructure, it would be extremely useful for going forward.

Andrew Thompson: And that’s what GitHub does. And it would be very cool.

Frode Hegland: Right. So on the topic, I mean, yeah, I mean, that’s that’s a fair point in a sense. That’s what I mean by code review. You know, to talk to, to go through the other webXR coders to see what could be done, to not have Andrew do every single thing now in that. But yeah, if, if you guys can have you. So I forgot to put my device on. Do not disturb, I apologize. Right. Okay. So Danny and I made a presentation yesterday. But just. I just wanted to. We just thought we would share that briefly with you guys since. Just so you can see how we presented what we’re doing so far. Oh, Peter, is your hand up for something else? Okay. Sorry, I.

Peter Wasilko: Forgot to lower it.

Frode Hegland: Oh, yeah. No problem. So just to go through a little bit. We talked about what’s at stake, you know, high level things, the opportunity things we talked about and of course, highlighting that academics is our user group and goal of this being an environment where people can think. And then we describe the three groups, the three sections of that, which is the software, metadata and dialog. We had a beautiful list of benefits for this specific headset and webXR. Thank you. Randall, we put in what you added to that. And showed a little bit of screen recordings, for instance, what it’s like to just work on a plain document in a pleasant space. You know what it can be like when you’re working in a public space and you’re sitting kind of close to people, which is not so comfortable. Also looked at the reader and author with video here on top, playing in Safari messages and so on, which is spatial. But as Deeney highlighted last week, it’s still forward oriented because it’s desk based. So then we just look briefly at author and reader. And by the way, to those of you who have the vision, if you’re not already testing, please tell me. I will try to re-add you because I can’t even test my own stuff. So what’s in these two little videos is way past what we’ve actually developed. We then mentioned what we’re doing in the lab. And we showed this cool stuff. And I did a walk through of What this is. Yes, you can have a feeling.

Frode Hegland: Show the how we interact. The hands, the spare. Now we can pull things out and so on. It’s just a really, really good place to be. Exactly where we should be for the end of of three months to to see where we’re going. Yeah, I got this one. Goes on a bit. You all know what it is. I mentioned visual media because that’s part of what it’s supposed to do. There’s a big visual metaphor. And then mention the whole thing of how visual media can explain what other data is doing. The LM thing, which is adjacent to what we’re working on now, but it’s something to keep in mind. And then there was the future design thing, which I’ll talk to you about in a second. I talked about the importance of being properly spatial. Here, it’s in a car. Because, you know, why not? The idea of driving through information showed our website and listed the dialog. We have any questions on that? I. Where did you go? Well, there you are. Yeah. That’s geographic is was for the LM one. It’s kind of cute bonsai, you know, to grow things. Okay. Right. I’ve been talking a lot. And if you don’t mind, I’ll talk. One more thing. Oh. Adam is saying he’s on the road. Yes. A lot of people are istring. Well, hang on, there’s something I may need to share with you guys. One point I would like to make about webXR on visionOS. Is that Crandall? Can you please use your computer mic? It’s hard to hear you.

Brandel Zachernuk: Okay, there’s a difference, but I will, I will.

Frode Hegland: Oh. Thank you. So this is better? Yeah, yeah.

Brandel Zachernuk: Go because it’s not consistently better, is the thing. Like, I tried this for a work meeting last week and it was still garbage. So That’s good. I’m glad. So webXR on visionOS is not available. It’s in a testing preview. And so that’s worth bearing in mind. That’s the sort of thing that, you know, the reason why people put things in a testing preview is so they can take them out at some point. But that that’s the that’s the basis on which it’s being supplied today, and it’s, and it’s worthwhile to characterize it as such. If, if pressed, if it’s necessary, you know, it is supported on on quest. It is available as a testing preview on on visionOS. Right.

Frode Hegland: Yeah. No. Makes sense. And that makes sense. Well, we just got an update.

Brandel Zachernuk: It’s not to say it will ever be taken away. It’s that that’s the sort of level of support and expectation and feature richness that it sort of supplied on the, on the terms of today which hopefully should mean good things in terms of there being a future, but that’s not that’s exactly the sort of thing that Apple doesn’t talk about. So.

Frode Hegland: Yeah I did. So, Tiwari, actually, how do you pronounce your name? You said the H is silent. Is it Ritwik Tiwari?

Hrithik Tiwari: Yeah. It’s Hrithik.

Frode Hegland: Okay. Yeah. He was there yesterday in our meeting with the webXR people, and I did mention that we have no idea what Apple will be doing in terms of webXR support, that we have a firm understanding that there is a real commitment to the open web. So, yeah, that connects with what you’re saying now. And Okay. You know, I’m just going to. Hang on. Boom. Sorry readers being updated. Yeah, I’m just going to show you something that is based on me desperately trying to write a use case. And, you know, doing use cases to this specific is obviously quite hard. So I’m going to share this and make sure you’re there. I can see you. You can obviously see this. If not, please scream. So one of the well, the kind of counterbalancing things we have here is when you’re reading a bit of text. Hang on. I have to tell the kids to be quiet. Please play in the front and get a meeting. And I need the window open for air. Sorry life, right? We want to have a richer three dimensional environment, but reading basically needs a flat background. Otherwise, unless it’s for a specific effect. Is easier on a flat background. So. This is in the environment where we kind of currently have. It’s also using interface styles from native. So the notion is just a top lower right and you get the outline on the right. This is for right handed people. It should be switchable if you’re a left handed person.

Frode Hegland: Point being that in this kind of space where you’re not using a cursor necessarily, you can just easily like a scroll bar on the right. That’s why it should be there. It wasn’t last week, that’s all. And on the left you can write notes. Can I both up? If you want to. And I don’t know if you can read the center one because the. Text bubbles there, but the point. Yeah. Just trying to move. There you go. So this would be exactly the references we have now in webXR. It’s just happened to be framed, that’s all. And you can then go into a wide view like this, either through your wrist or through tapping on some kind of a button. And as you can see here on the bottom, you have a list of how the references should be shown, either a list or a map. And on the left you have connect based on certain attributes. On the right is to show certain attributes, and no matter what we’re doing, we’ll need to. To have the ways to say what should be shown. And on the top there you can see saved views and this will be connected to your normal library. It shouldn’t just be in a frame and this is the same as you saw the other slide. It should be lively, three dimensional. So it was just to get a bit of discussion on. Kind of our approach to where we should go next.

Brandel Zachernuk: Cool. Oh. I’m not sure if folks have noticed. Is it still audible? I have my heater on now. Is it? It’s okay. Great. Great. I don’t know if folks in particular Andrew, have noticed in quest. Now, when you suspend the session, you can suspend a web session. You can actually see your 2D window up again. Still that mode is called visible blurred. And it’ll only work on a quest platform. So quest Pro Quest three. But what it means is that you’re in your webXR session. You get to see things that are portrayed in VR. But you also get to interact with a 2D window at the full fidelity that a 2D window has the ability to be displayed with him. That’s not a bug. That’s by design. And it’s something that it’s an official proposal by meta. I think it’s really interesting. And it means that you have a pretty fine grained surface and then native you know, Dom events and layout capabilities that you would be able to use to manipulate and navigate things. So in some regards, it’s the best of both worlds. The downside being that you don’t you no longer actually have XLR inputs. The state is called visible blurred. So it means that you don’t have the ability to or the benefit of using your direct input and controllers.

Brandel Zachernuk: But if you construct a system where you’re mediating sort of the contents of a document through the 2D page which is, you know, not ideal, but, but pretty interesting, then there are some pretty cool things that can come out as a, as a consequence. So definitely worth thinking about as as a way of being able to broker a mixture of rich text centric interaction alongside alongside the sort of the broader stroke spatial portrayal of that information. So the picture that you had of the, that sort of main document, sort of with other sort of attendant spatial things around it was really evocative to me of a thing that, that that could be pursued via this mode. It’s not. It’s not on. I’m not aware of any plans for it, but it’s the sort of thing that often happens by accident. During various testing and development. So if it turns out to be good, it’s not a hard thing for other people to be able to implement at a browser level. Obviously, it’s not something you can do, but. Yeah. But I just wanted to sort of call it out as an interesting thing if folks haven’t sort of taken a look at it at all then it’s, it’s worth understanding sort of some of the dynamics and practical consequences of.

Frode Hegland: Yeah. I mean, this is why we’re so grateful, Randall, that you made us think about this for a year and a half, two years before the headset came out. Because now the difference between webXR and Native and quest and vision are real. And we need to think within that. We also beyond it. And, you know, that’s partly why I’ve developed reader and author provision so that you know, to really get the feel for what that means. Yes. And then we can extend it. It it is very frustrating to not have passed through in webXR. Even if it was only for the viewer that it had no informational value. Sent to the Serbian web page, it could be quite, quite useful. Even so, that.

Brandel Zachernuk: Definitely. That would be a must have is to make sure that a web page is not a is not there’s a mechanism for a, for a web page not to be privy to that information. One of the challenges today is that There are a limited set of ways to do that via WebGL, which is the the main way that that webXR is rendered. And it’s a possibility that web GPU, which no one has implemented in WebEx or could be a more future forward way of being able to construct privacy preserving layers of information. So that’s something that, you know, I’ve been speaking to the folks at Google about people who are responsible for Webgpu WebGL. And I think there’s some links there, but yeah, it’s there are there are privacy and security challenges that are, are not trivial. It’s not it’s not just a matter of wanting to keep the fun stuff for the for the paid customers. It’s, it’s the, the very difficult engineering of finding a way that allows this to happen in a way that everybody agrees with.

Frode Hegland: Yeah, yeah. As I mentioned yesterday in the other presentation when asked. A few people who are real knowledge. People recently have asked me and I’m sure they’ve asked you guys too, why would anyone work in this stuff? And at this point I don’t have a good answer for them. And I think that’s very exciting. I don’t think that’s horrible or negative, because if you have a good sized screen or two and a good keyboard, most of this you already can do, you know if the system is there. But the thing is, it isn’t even developed for 2D. What we need. No, sorry. Important message.

Dene Grigar: No, we’re just text messaging back and forth behind the scenes. So he’s asking about asking other people to use cases. And my response is ask everybody. I mean, everybody should be doing it. So just go for it.

Frode Hegland: Oh, yeah. No, I thought you had to go out since you weren’t on video, but yeah, no.

Dene Grigar: I don’t normally come on Mondays, but I thought today would be a good idea to drop in, so I’m just listening while I’m working on Monday morning. So I usually am not here on Mondays, but I’m here lurking. Right?

Frode Hegland: Okay. Yeah. So, yeah, lots of points. It’s it’s an interesting point now being at the three month mark, which is not the quarter mark, because we’ll be showing our work in September. At the hypertext thing. That is our big goal. And Wait. There are so many ideas and things to do. Andrew, a question for you specifically.

Andrew Thompson: Yeah.

Frode Hegland: In order for us to kind of fork the code.

Frode Hegland: You know, of course, one of the things we’re talking about is building an environment of. Of kind of other aspects. Is that something we can do or should do or, you know, we don’t want to just have necessarily one thing. If we have one brilliant thing in September, that’s fine. But we don’t want necessarily want that for years. I mean, what would help you to comfortably move forward as my question to do all kinds of fun things.

Andrew Thompson: So with GitHub, you can easily fork the code. Should we I don’t know. I don’t like the best base code there. Now, if we want to have like multiple different applications running off of the same code people are welcome to try if they’d like. This has been built just kind of clobbered together with a whole bunch of different tests like I mentioned earlier. So if people see use in that, they’re welcome to use it. If we want to just have different experiences to demo, it might just make sense to have them make it in whatever is most comfortable for them. Like I know a lot of Adam’s demos were just totally their own thing, but because of that, they demonstrated things that my code couldn’t do. So if we’re looking for just a whole bunch of different prototypes, that would probably be the best way to do it. Of course, if we want to have a whole integrated system where they all can talk to each other and stuff like that gets a little bit more complicated, because then we have to all agree on some sort of fundamental principles on how the whole code works before we start building our own practices on top of it. And yeah, that probably be more of a collaborative effort before we split off. So I guess what’s the vision? That would be a better question for me.

Frode Hegland: What’s the vision? Oh. Do you remember when Randall, first time he did this and.

Andrew Thompson: I walked into that? Yeah.

Frode Hegland: That’s the vision. One thing we need to do, which is part of our initial. Promise is it has to be the user’s own data now. I talked to Adam briefly. He’s in this Easter thing in Sweden. He just has no real time for us. But he mentioned that. Well, rather reminded me that him and Brandel suggested a means to get stuff from the computer onto the headset that didn’t require a server. I think that should be our priority now, so that we don’t just have these references that are static, because I know you made it so they don’t have to be static. Andrew. But I think that mechanism needs to happen because sorry for repeating myself, but people need to see their own data. Otherwise it’s a just a demo.

Andrew Thompson: Right? The URL link that I put at the top I get the feeling no one’s tested that it does work for sort of any data, but it has to be formatted the same as the hypertext papers. So it’s it’s very much not any data. Of course, that’s a terrible interface. And that’s something that we will change with time. But the the base code is there for that part. So it shouldn’t be that difficult to connect it to whatever way you want to stream it from the computer. I believe the next step that we want to focus on is finding a way to integrate with like an export from reader. I think that was what your vision was for that part. I don’t know if that’s still next step. I’m working on the save and export bit, which has given me more hurdles than I expected, but I hope to have that presentable on Wednesday and then we can move forward on to that next part. Yeah I can other people can work on whatever they’d like in the meantime.

Frode Hegland: So I have a question for you, Danny. Before I answer Andrew’s question. And that is Hang on. One second, and this is kind of an important one. So I just see the question is on PDF because we talked about the user’s own data. But now that I’ve built reader, as for our community envision as well, I think we have fulfilled our obligation to show PDF and XR Erdeni. Sorry to drag you back in, but this is a really important question. Now that we have reader. Yeah, yeah, the question is, now that we have Reader Native in the webXR, do we still need to render PDFs, or should we render based on the HTML of the documents to have different views?

Dene Grigar: I’m going to say this, I think. Mark Anderson would prefer the HTML, as would I, but we all know that our colleagues in the field will never use an HTML page unless they absolutely are forced to, that their libraries consist of PDFs. So if we’re going to be building this out, as we said, we’re going to build it out. We have to use PDFs as our main document format. That said, we also have included in our language that will try a lot of other different types. So I think we shouldn’t turn our backs on other types, but we can’t turn our back on PDF. As much as I hate them and Andrew knows I hate them.

Frode Hegland: Oh, it’s it’s not a question of turning our back on PDF. Absolutely not. What I’m saying is because reader has for vision has been made for the project and will be open. And by the way, I have made the I don’t know if it’s done yet, but Adam will have access to all the source code for reader and author, for Mac, iOS and Vision Pro so he can do his own versions, which means we have different ways to integrate. Anybody else can have that as well. So now that we can have multiple PDFs floating in many dimensions, to be able to integrate from that library into webXR, I think that really fulfills the requirement where we can use Andrew’s time to have much more dynamic views and then still say we’ve done the PDF properly.

Dene Grigar: No, no, because our deliverables are based on the PDFs. I mean, it’s reading and writing with PDFs. We can’t say we did. We tried it and then we moved on. We can we need to stay with it and try things as side quests.

Frode Hegland: Okay. That’s really important because you did mean writing PDF, right? I mean, for all.

Dene Grigar: I know, but we’re talking about authoring and reading texts. Yeah. So we’re. I mean, what do you. I mean, we’re not going to code in HTML in the, in the headset either, right.

Frode Hegland: No, no, but what I’m, what I feel is that. To really use the different parts of the virtual environment to their best. What Andrew is developing is a swivel chair experience 360 not front facing, but it’s also not a walk around environment because you’ll bump into things because we don’t have the the pass through in webXR. So to me that means that we should probably really try to make that a multi-dimensional environment where we focus on references and things. References are very, very extractable, even from the PDF to put into webXR. So the benefit of being able to read a plain PDF in webXR, I think is much smaller than to read them in a native app. That’s why I built the native app so that we could fulfill that. But do you feel that the PDF has to be in webXR as well? Right?

Dene Grigar: Yeah. And I also think as bad as it is in the headset. Right. In the Activision Apple Vision Pro. But we also told the Sloan Foundation we’re working in an open source. So as much as I love what you’re doing with reader and author, that’s another side quest. It’s an important one. But it’s not what we it’s not what we promised we we did have it. And the original version. We’re working with author and reader, and we took it out and what we submitted didn’t include it at all.

Frode Hegland: It is included.

Dene Grigar: We didn’t get out.

Frode Hegland: Oh, no no, no, it’s there. In the latest version, it says it will be connected to other environments as pretty much exemplified by author and reader. So. So it is in there. But I bet that mostly refers to on other platforms like Mac or iOS. So. Okay. So so that means that we need.

Dene Grigar: Can I answer Peter’s question? I saw Peter’s question. I opened up GitHub to everybody who sent me emails. And so if you sent me an email, you should have access to GitHub. If you’re using a different email address, then you’re not able to get into GitHub. So right now in this chat give me the email address that you want to access GitHub with. And I will make sure Greg opens up GitHub with that email. And remember, I brought this up two months ago.

Andrew Thompson: A lot of what happened, I think, with the the GitHub stuff is people sent their email in. The invites were immediately sent back, but they expire after a week. So if you didn’t accept the invitation, it’s expired and you’re not in anymore. And we just have to do.

Dene Grigar: I also reissued them again. I did it twice. Brindle, shaking his head. Thank you Randall. I’m not losing my mind.

Brandel Zachernuk: Yeah, I recall something to that effect. I explicitly elected not to get access, so I wouldn’t be surprised if I tried and would fail today. But that’s kind of by design.

Dene Grigar: So drop your email address in the chat and I will make sure you have access. And Greg’s back from his holiday, so I’ll get it done today. Okay, Peter. Okay, everybody.

Brandel Zachernuk: Yeah. Thank you. I’m with Danny on the validity of PDF for this. In general, but also specifically to to satisfy the scope of this long grant, I think I think that there are things to do with the specific presentation of PDF. Pdf is so much more. I mean, you know, this is your talking points back at you. Proto is so much more robust as a representation of a fixed presentation of a document, and it’s going to have so much more resonance with the academic community as it stands today, that I think that it’s going to be a very useful thing to be able to do, as well as to be able to counterpoint against the flexibility of HTML to the extent that you wish to have that as a representation. But being able to keep that claim that what you’re ingesting at some level is, is to is PDF, I think is a is a really powerful claim to be able to have and using pdf.js to be able to kind of make that, make that hop like almost immediately is also totally fine. It’s just being able to, to really sort of defend that claim, I think is, is is really valuable. Yeah, I’m.

Frode Hegland: Fine with that. But the key thing that I really want to have locked down now, which is something we we looked at in the beginning, we need to have the users data be uploadable. Now that has to be the priority. So yeah. So, Brandel, the method of using the user’s computer to do that, would you mind walking that through that again so that Andrew can tell him whether he feels that’s a viable way of approaching this?

Brandel Zachernuk: Yeah. So when you have a computer and you have a laptop and a headset on the same network, and you have the ability to set to have a small host on the, on the laptop or, or other computer. Then you can use a. Peer connection, using something like PJs to be able to have an open pipe. And then if you disengage the loading and parsing from the, the, the, the loading from the parsing, then what it allows you to do is inter intervene with a, with a message passing system that allows you to connect any arbitrary client to any arbitrary server in the sense of being the thing that happens to be doing the stuff. And and that just means that you can connect any device to any other device and, and sort of exploit the, the benefits of being able to do it on there. I, I haven’t tried drag and drop on visionOS, I haven’t tried I believe it should work. I’ve been told that it should work. You can definitely open files and and you get the files dialog. But drag and drop should work there. But, you know, in all likelihood, most people do not have, you know, a trove of PDFs lying around on their, on their headset. So having the ability to drag those in in the same way that I’m doing with that various sneaky paste URL that has, has the reader capability set up on it. So. Yeah, it’s it’s just a matter of being able to stand by having a A.

Brandel Zachernuk: A peer connection between one device and another. You can actually use the open source peer server, or you can stand one up yourself. It’s basically just for brokering the immediate and initial connection. So to my knowledge, that information is private from that point on. But I’m not a security researcher, so and then you just are sending messages back and forth, and so those can be as big as they can be permitted or as small as you could be bothered in terms of the granularity of the information that sort of constitutes a valid transfer of, of documents for the purposes of the the author parsing that I did. I guess, jeez, like over a year ago now. Right? I was passing messages back and forth at the granularity of the individual nodes and individual connections, and it processes in almost real time, it seems when you open it up, you know, if you have your headset open on the page and then you have the the document, the, the, your computer open on the page, you drag an author document, it seems more or less instant that I, that I see that that it loads in. And you can you can delay those things, you know, you can turn it into a message queue, all of those kinds of things. But but that’s the basic premise. So if you’ve used one before, it’s that if you haven’t I can go into more detail about how you negotiate a system via PJs or equivalent socket.io and that kind of thing.

Andrew Thompson: Wonderful. So thank you. Randall. The concept makes, like, a lot of sense. I remember that from last time you explained it. It was great. I’ve never done something like this in JavaScript before, so I understand the concept and implementing it are two separate things. But I will definitely go that route. Once that is the priority, right? Yeah. It’d be great to have.

Frode Hegland: So the question then would be, Randall, how much handholding could you give under for this? Because this is such a crucial bit.

Andrew Thompson: If he’s mentioned the peer to peer DJs, that should be enough to get me in the right direction. But if you have any like specific resources like linked to them, that’d be great as long as it doesn’t take you much effort.

Brandel Zachernuk: Yeah. Yeah. The CodePen that I have about auto peer should. Should should work, should sort of just work. And I don’t necessarily recommend that you use Autopia, but Autopia is a thing that I made, which is like 2 or 3 lines of code to have a peer coordination framework set up on a page for the purposes of zero effort information sharing. I think we looked at it in this call before.

Andrew Thompson: That sounds fabulous. Yeah, I think that was the one we looked at with was that the reader turning into the visual map that just goes global as soon as you upload it?

Brandel Zachernuk: So that that is just using a peer connection. Auto peer is a was a concept that I had here. And it’s the one entitled. This page automatically coordinates with zero effort right on CodePen. It may not work for me at this instant just because I am on my VPN, so that’s the one thing to bear in mind is that this this approach can at times be confounded by By you know, well intentioned institutions preferences for, you know, ironclad security. So if this is. And I should. It might be. Maybe it’s not working. Doesn’t seem to be working. Gotcha.

Andrew Thompson: We we would want it to be like a just on the local network rather than just open and global, but This is cool. I think this is one that you demonstrated before. I don’t remember, I thought it worked last time, but I don’t see the mouse working this time around.

Brandel Zachernuk: Yeah. I I’m not sure. Maybe it’s because I.

Andrew Thompson: In the meantime, I did have one other thing I wanted to comment on with the. Oh, there it is. Perfect. Now I see it. Comment on with the PDFs in the headset. I was under the impression that a couple weeks ago we had all agreed that using the HTML exports of the PDF was still a PDF, and that was acceptable for the project. If it isn’t, I do need to point out that most of the connections we have right now with the citations won’t work because it’s specifically using the HTML tags to find things which aren’t in a PDF. That was the whole point why we went the HTML route. So if we’re switching back to the PDFs in the headset We are going to be undoing a lot of work. That’s okay, because that’s the whole point of like testing and working through things. But just people need to be aware of that. And I’m not sure how we’ll consistently grab pieces of citations anymore. But we can discuss as a group.

Brandel Zachernuk: So, so my view on this and I will, I think, as we should all defer to Dean, his view of, of compliance to the, to the system is that as long as there is an entry point of ingestion with PDF of some level. There’s no particular need for it to retain the form of PDF there. And after, as long as there’s an open ended mechanism for for new PDFs of any sort of valid type to, to, to sort of enter the system, then we should be under no obligations to preserve the PDF in order to do that. Dean, does that sound off base?

Dene Grigar: And sounds fine.

Brandel Zachernuk: Right? Does that make sense?

Dene Grigar: At some point, yeah. I mean, I want to go on record saying it’s really hard because I don’t want to sound like I’m the gatekeeper. At the same time, you know, we’re looking at a grant with promises, and I’ve got to make sure that we follow them because we want more money, you know? Yeah.

Brandel Zachernuk: I’m very conscious.

Dene Grigar: That we can do whatever we fucking. Please know that we can do whatever we want, but we’re not going to. We’re not going to. We won’t get the second year of funding. Right. Number one. Number two, they won’t give us any more money after that. So I frankly, I like getting money from foundations. Yeah. No. This is.

Frode Hegland: Oh, sorry. Please. Sorry. Please continue.

Dene Grigar: No, just saying it’s important. And I just feel like I’m in this position all the time. Like. Like the mother. No, no, you can’t do that. No froda, no bad boy. And that just. It’s not the position I want to be in. And that’s why I’ve shared that document with all of you so you can see what’s there. So I don’t have to do that. I don’t want to be that person. Yeah, I’m. I’m lovely. I’m actually quite lovely.

Brandel Zachernuk: No, I, I definitely get that. I’m get that I that’s clear. I actually spoke to a mutual friend of ours, Andrew Hedges. I love Andrew Hedges. He was my boss. He was the guy who hired me to Apple.

Dene Grigar: Oh my God.

Brandel Zachernuk: Yeah. He’s been a he’s been a phenomenal mentor to me over the many, many years. Even though he was only my boss for seven months. He was he would love to see this. If you if you get the chance to run this past him at some point.

Dene Grigar: I actually have met him before. I knew. I’ve known him. Yeah, from some past life.

Brandel Zachernuk: Yeah, yeah. So he he was a I think he might have been University of Washington Vancouver at some point in the past, but yeah, he’s he’s, he’s a, he’s a really he’s a really, really great fellow. But anyway can I ask.

Dene Grigar: One more question? Can I ask you one more question? Have you come across a guy named Matt Jockers?

Brandel Zachernuk: Matt Jockers does not ring a bell. Good. Okay.

Dene Grigar: Here’s my former dean, and I didn’t like him. And I have this on tape.

Brandel Zachernuk: Yeah, he one needs to to probe the social network and. No, I’m so I’m. I absolutely sympathize with the what you feel that that kind of role puts you in a position of. But it’s absolutely critical. I’ve been involved in sort of research, creative technology stuff for ten years and at Apple and they’re just innumerable times when somebody will offer something promising, then change the goalposts. And the thing they’re offering is still promising, but because it doesn’t do the thing that they said it was going to do, they don’t get to do it again. And and it’s, and it’s and being able to put, you know, admittedly, you know, things onto an ever expanding list of interesting things to follow up on. One keeps you in better graces to be able to receive the next round of whatever it is, be it salary or budget or funding or or just whatever kind of opportunities exist within the common currency of an organization. And two, it actually changes what it is you have the ability to explore because there is something Challenging, like, undoubtedly, but nevertheless different and important about discovering what happens when you finish stuff. I don’t want to say.

Dene Grigar: I hate I absolutely hate Microsoft and I don’t like Adobe products. I’ve been mad at Adobe since, I don’t know Microsoft Macromedia Director, right? Since they killed director, then they killed flash. So they’re not my favorite people. And Microsoft has never done fake people. I do think that author and reader are far superior products for the kind of things we’re trying to do. I do, I do, I do, but the other thing that I have hoped for is open source. I do think we’ve got to put our energy into open source. As much as I teach in our program, the, you know, Unity and Unreal Engine and all these things that we have to teach Adobe products out the wazoo, we still have to be looking for solutions that are going to be better. And that’s just part of it. And I and I’m I’m heartened by the fact that Apple is not against web XR, that it actually is looking at webXR. Because Apple has had a reputation of being a closed ecosystem. And and as you know, it says, as I said to folks yesterday in our meeting, I’m not an Apple fan girl, but I’ve been I’m an Apple girl, right? I’ve been using Apple forever, and I do not use anything that’s not Apple at all. And when I was forced to buy a PC this year, it’s still sitting in the corner of this room and I don’t use it right. I pretend I use it, but I don’t use it. Don’t put that on tape. Anyway but I do think that it’s important that Apple wraps its arms around webXR. And we talked about this yesterday with our meeting with Hrithik. I’m hoping I say your name right. Hrithik. Hrithik?

Hrithik Tiwari: Yeah. That’s right. Just a double tick. That’s very thick. Okay.

Dene Grigar: Thank you. And Frodo. Did you want to introduce themselves?

Frode Hegland: Yeah. If you would like to. I just put a little comment here in the chat, but just on the on the current topic. I absolutely would like to have PDFs in webXR. That’s not really the question. The question is partly where to focus now. And I’m very, very glad that we’ve decided the local host will be the model. So you’ve got to keep your laptop going. When you put on your headset in the beginning, you know, that’s good. I also think that the JSON containing the user’s library on Mac should synchronize. That’s partly what this does. And of course, PDFs are a lot longer than a JSON string. So that means that we need some kind of a mechanism to tell the headset that this was that if there was a PDF open on the computer, transfer that first, it’s very likely that’s what they want to keep reading. So these these mechanisms are perfect. We also have to make sure we utilize the best platform and make it open to make the best experience for the person. So that’s it. I think we’re all in entirely in agreement on that. Right. So that’s cool.

Brandel Zachernuk: Yeah. Yeah.

Frode Hegland: So Andreas focus now will be get stuff there. And when it comes to the JSON under if you can give me a sample request I sent to my guys. Branville. If you could please give me a pointer to where my guys should look to implement it on this side. Either or do you think should just be in a website? I don’t mind building it into reader. You click a button. As long as we make that code open, free and documented and available, then any developer can use it to write.

Brandel Zachernuk: Yeah. So what you’ll need on in a native app context is the ability to use some. You’ll need two main things. One is an endpoint to send the socket connection, whatever based channels, you’ll need to negotiate them as well. So if that means you want to use PJs with it’s. I believe what happens with PJs is that it does hit like a Google server somewhere for the intermediation and the, the the matchmaking of being able to say, this is a peer. This is a peer. You go talk to each other. If you don’t want to do that, then you will need to have some other mechanism, be it, you know, a server on the, on the computer or present in your application that is able to start that connection up. Right.

Frode Hegland: A problem we’ve had over the last in the beginning, we’re having again now, you’re doing a lot of ifs. Please consider me not anyone else but me. Ignorant enough for the IFS to fall on deaf ears. What you decide. And that Andrews happy with will just do it. I’m not saying you have to make if there are real decisions. Of course, as a community, we’re not just going to put it on your shoulders. But, I mean, a lot of this is I can’t usefully contribute to in a if situation. Does that make sense?

Brandel Zachernuk: Sure. I guess my what I was unclear on is the level of hardbound information security. You want to play by. So. So in the event that that is not your largest concern and that you don’t have the fear that you know, state actors are going to be breathing down your neck, then I would just go with the Google solution.

Frode Hegland: Do you agree with that?

Brandel Zachernuk: Danny? That’s not acceptable. Inside Apple is all for obvious reasons. So it’s something that I’m I’m used to having to hedge their saying, like, if this is the level of disclosure that we need for this project, then we need to stand up everything ourselves, and then we need to make sure that it all goes past Sierra is to go past infosec. So so yeah. So that in that context just use just use PJ’s largely as it is that will allow you to broker those connections. You’ll then need to know what a valid identifier is. I, I think I’m pretty bullish about the, the approach that I outlined and sketch in that code pen that I have the, the auto peer where you use an identifier that’s based on a person’s identity or anything like that. And, and then being able to kind of broken the connection on that basis. Because you need to have a unique.

Frode Hegland: But we talked about that bit last week was email address plus a single word or a single number. So it can’t just be guests. You know, I can’t just take Danny’s email address and therefore eavesdrop. It has to be two things. Not like a password, but just almost anything. Is that something? And two.

Brandel Zachernuk: Yeah. Yeah, exactly. And something generated so that other other participants wouldn’t have the ability to guess it simply on the basis of your credentials. But that doesn’t need to be robust. And it’s not a particularly interesting part of this sort of the conditions of the grant. People can do more innovative things with that, and you can just plug them in. So, so that and then the slightly more complicated thing is you need to agree on the message passing format. You need to have a basic notion of the moving parts that you want to be. You want to be shuttling back and forth across these participants. And that’s going to depend on what’s displayed. If it’s just the two dimensional textual document contents then that’s less controversial. But it means that you have less ability for the computer in the mix to be able to do determinations about the display. So it’s kind of at that point, though, it’s it’s easy enough for people to be able to bolt on whatever form of message as long as they you’re either using a single code base in order to, to sort of conceptualize what these messages are, or you have a well-formed representation, sort of an API doc of what it is that you’re passing back and forth between those participants. Hopefully that that makes sense.

Frode Hegland: Andrew. How did that make sense to you?

Andrew Thompson: And yeah, conceptually.

Andrew Thompson: Well, not exactly clear on how to implement it, but that is something that research into pure JS will do. So we should be okay. So and I’ll, I’ll ask you questions if I don’t get it. Yeah.

Brandel Zachernuk: Yeah. No feel free. Feel free. So like the what. All I mean is that once you’re in a JS world, then P are not pure. You’re in a situation where you’re passing messages and so then what you would have is a message like Createnode. And then you would need to say that a node has a name and Uuid and a and an x and a y and a z and a scale maybe, you know, and so you need to agree on the basic terminology of what, what needs to go in and who needs to invent what. But but, you know, there may be some things, for example, like the z coordinate that are irrelevant in the context that you’re constructing the node graph on the part of the computer because you just haven’t been in a position to to invent those. So then you either need to make those values up on the fly. Once that sort of lands in the immersive environment or or you need to make sure that the those things are being initialized and instantiated with reasonable null values like zero. Right. For for Z and stuff like that.

Andrew Thompson: Yeah. I mean that’s a great example with the z value. Because we’re going to have like a reader snap distance by default. So that can be changed inside the headset, but an export from author wouldn’t care. That’s right. But I assume, say, like, if you want to keep your your mind map layout road then an X and a Y would matter. So we’d keep those and then not worry about the third dimension. I think ultimately what it’s going to boil down to is you put in as much data as you want on your end. Because I feel like your developers are going to have more of a pain going back and making changes. It’s easier for us on our end to just be like, okay, well, we passed in like the author, but we don’t really need that, you know, something like that. Like, who cares? That’s such a minimal amount of data.

Brandel Zachernuk: That’s right. Yeah. And the speed of computers these days, it’s it’s a little sickening, honestly in terms of what it means that you can get away with wasting you know, printer drivers being a testament to that, obviously. So yeah, I would I wouldn’t be too worried. Another question might be, do you want to make the mechanism of transitions transmission stateful or not? My recommendation would be to not stateful means that you can go like, I’m in insert mode, I’m in edit mode, or I’m in put them into this child mode versus that mode. And in general, those things get you in trouble because dropped packages sort of mismatches of contexts can get in trouble, get you in trouble. So being more verbose and stateless is generally better for it. Sorry.

Frode Hegland: That’s a really, really good question. So and I’m so grateful you’re here today. The way that I view it is. But, you know, as I said, any client outside of the headset should easily be able to integrate with the system. So that means that they can take the advantage of whatever special things they have in the outside world. Right. So what needs to be updated are two basic things. One is basic information about every document. You know, the title, also the document name and so on, so that when the documents are uploaded we know what’s what. And two is the library information that will include which documents are favorited, which documents are hidden, and also the XYZ when it is used. Because to me, it is clear that the spatiality of the webXR environment will mean that we may end up doing some amazing graph map type things in there, and they need to be able to be sent back to the outside world when they’re received by the outside world. We will have the XYZ, but it needs to be. It can be stored as its own thing. Let’s say you try to open and say, no, this is three dimensional. You’re on a flat screen, forget it. Or there may be a rendering mechanism that says, I’ll ignore the Z dimension. That’s entirely for the developer of the outside software to do. We just have to make sure that we explicitly can load it and send it. And that’s totally fine to do for me. And I’m not speaking computer. I’m speaking human now at the end of a session. But does that make sense for you guys? Does that fit what you’re talking about?

Andrew Thompson: Yes. Yeah. Just to clarify what you said. I’m not done by a computer or done by you. Are you talking like you click an export button and then manually save it somewhere rather than it just flies back to author.

Frode Hegland: I would. I would like you to automatically fly back. I mean, the dream is the dream scenario that we talked about from the beginning is I’m working on my laptop or something in my library. Meaning in reader’s library or whatever. Software library. Oh, I need a better view. I put on my headset. The least amount of things I need to do to make the data go up the better. Similarly, when I’m done in the webXR experience the least amount of stuff I need to do to send it back if we have to for now, have a giant button that says send back. We’ll have to do that, but I’d love to hear what you guys think to make it more elegant.

Andrew Thompson: Yeah, I mean, for testing, that’s how it’ll start. But we can definitely aim for the the other route once we get the peer to peer setup. They will have to be something that determines, like when it goes back. Unless we want to constantly be streaming all of the data. Because from author to the headset, it’s rather small data. Assuming, you know, the biggest is going to be the PDF itself. Now, if you want to send it back, it’s a lot larger. And currently I’m dealing with export issues with troika text, which is stupidly large because of all the glyphs. And it doesn’t actually save. So it’s just like, useless data. So I got to figure out a way around that.

Brandel Zachernuk: You’re. Yeah. I was trying to save.

Andrew Thompson: With. Well, the current current project is to export the entire scene. Not like, literally the scene in code, but like the visible scene for the user. And then. So you can re-import it somewhere else, like a library. But troika text doesn’t save because runs off shaders. So I’m doing like workarounds on that, but because of the way I have it set up right now, it attempts to save all the troika stuff. So we have like a massive file and then it just overwrites it once it imports. So I got to find a smoother way of doing that.

Brandel Zachernuk: Right. So the thing that I did for my for timeline VR saves to Gltf and everything else it where I and you can probably use basically the same thing. Where what I’ve done is it it actually parses HTML and then creates do you know what an atlas, a texture atlas is.

Andrew Thompson: I’ve used them in game engines, not so much in JavaScript, so I haven’t created them from scratch.

Brandel Zachernuk: Yeah, so a texture atlas is just where you have a bunch of textures that you need to use laid up on a single on, on a single sheet. And what I do is dynamically generate a texture atlas based on all of the font variants in a page, and then then I serialize all all of the sort of Dom content such that you can kind of construct essentially what exactly what troika is, except it was using alpha blending rather than rather than a sound distance field texture atlas. The actual geometry of troika is perfectly appropriate for export. What you need to do, what needs to be done, is to convert it into something that can be exported as a gltf. That is a it’s a non-trivial project, but it’s, it’s it’s an eminently doable. All we need to do is convert the geometry to make sure that it’s exporting with the correct coordinate spaces and stuff like that. And then to convert the the atlas texture from something that is this signed distance field to an alpha threshold, so that such that the gltf exporter has the ability to to do that and that that may be as simple as, as augmenting the gltf exporter a little bit. I yeah, I hadn’t a bit of a look at that, but it’s but but it’s, it’s likely something that Adam can help a lot with.

Andrew Thompson: Gotcha. Yeah. I hadn’t considered going that route. But like, as a counterpoint perhaps we don’t go that route because all of the the geometry data, even if it can be exported, is very, very large. And right now, I basically like what.

Brandel Zachernuk: What what what kind of.

Andrew Thompson: Size are you talking. I’ve got, like, it’s like a one and a half megs for just the citation block, which will stack up very fast. So the way I have it currently set up is I’m just I’m saving the actual data of the text separately and then rebuilding that on import which would be a, a much smaller file size. I just need to also find a way to clear out the geometry data on export. Because just of the way I currently have it set up and my rather minimal knowledge of three.js to properly clear that, I would also end up wiping all of the the actual information, which would not be good.

Brandel Zachernuk: So one, one, you know, 1.5MB is a single JPEG these days. Two what is the geometry if you haven’t got the troika successfully exported at this point? Like, what is it that this what is what’s what’s constituting 0.5 x.

Andrew Thompson: So it is exporting correctly. It just doesn’t import correctly. Basically the shader information doesn’t he isn’t connected to the export from what I understand. So you get all of the, the base geometry information it loads in and it’s not there. It’s no longer troika text. It’s just like data. So you can’t edit it. You can’t do anything to it. You can’t see it. So I’m just I’m working on rebuilding it. And I know 1.5 isn’t huge for, like, in the sense of scope of an image, but for what it is, just a few lines of text. That’s way unnecessary. Yeah. And I know it’ll stack very fast.

Brandel Zachernuk: Yeah, I.

Hrithik Tiwari: I can give you some insights on making it better. Like how we are doing it, like with the hubs code base, which I work. We have a exporter in blender. And as you are also trying to export the data with the gltf export. But with this, if you have a custom exporter extension and while using your gltf export, you can include the data that you want to have with your custom exporter that you have built, like suppose your exporter name is is a exporter. So just while doing gltf exporter your gltf file, your GLB file will have a separate section, which is called like a export data. And in that you will have all the key value pairs that you want to use. So you can just essentially inside your blender itself, have a your data where you can put in the shader information manually, or either you can run a script where the blender shader data is taken and converted into the a exporter data that you want, and it should be like exported. In that. Right. So in that sense you can just take it in your three.js code wherever you want. And that particular component, you will search for that particular component. You will get the coordinates rotation sorry position and rotation and scale. And with that you will place the you will place your text wherever you want with those metadata and additionally your shader data. You will like have some script where you are decoding the data and, you know, mapping the values and showing it. I mean, if this makes sense I really find this a very good solution for. Keeping the like. If you want your shader data with your text without, you know If you are going to use a texture atlas, that would also solve the problem, but maybe that would also include more complexities. Into.

Brandel Zachernuk: Yeah. No, you raise a really good point. Like the the the point of this is not for the look to be preserved, but also in no small amount the, the intent and the capability of having text to be preserved. And so this is actually a really good point, is that if you have the ability to, to have an export or importer extension, then basically what you’re serializing to file is the messages that we are using to throw the stuff over the fence. So it’s a, there’s a, there’s a really wonderful synergy there in terms of having a well enough formed sort of schema for the messages, because then that means that the actual stuff that you dump into the gltf is also the messages. Does that make sense? Would. And he said, does that make sense as a way to to get around kind of both of these problems at once?

Andrew Thompson: Sort of I’ll admit it’s a little bit foggy on some of the details there. I didn’t respond because I thought you were still talking about rhetoric. Yeah. But yeah, it’s implementation wise. Yeah, I think I think I’ve got the idea. But we’re not. Never mind. I’m not going to run off of this because I’m. I’m just going to start, like, brainstorming out loud, and it’s not going to make any sense. I’ll just I’ll put a pin in it for now, but great stuff. Sure, sure.

Brandel Zachernuk: Yeah. So just briefly, like the the way that you build a document from one computer to another is by sending the commands over, which is functionally what a program does, but it’s all inside the same program. Once you make that break from having a program with function calls to sending messages that give rise to the function calls, then you’ve you’ve intermediated between those two sort of motions. Functionally, what a save file is, is just the serialized set of instructions that gives rise to a, to a document being composed from the very beginning of its life to its, its state to date. And so it’s just a compression of all of the stuff that you have there. But yeah, we’ll go over it again. But that’s, that’s in the broad strokes. What happens is that once you have this message passing sort of system, then you have the ability to disentangle it in time and in space. Right.

Andrew Thompson: And it’s like, I get that part. I’ve worked with save systems quite a bit because like I said, I come from the game dev side of things. And I have that part working right now in the current dev build. That’s not out yet. It’s just not every piece is currently saving properly. So I’m kind of going through and troubleshooting. That’s why I didn’t have any build last week. The bit I’m more foggy on is what the new suggestion was instead. And because I haven’t done it yet talking about sort of rebuilding shaders that’s a bit more complicated because I don’t work on that side so much. So maybe that’s what I’ll have clarification questions on. Once I get there, or if it’s even necessary. Like I said, maybe just passing the data and not worrying about the shaders and just building that from scratch like I’m currently trying to do. Maybe that’s just the fastest way to do it.

Brandel Zachernuk: Yeah, I think, I think, I think that might be right. And I think that’s largely in line with critics suggestion there. But may I?

Dene Grigar: May I interrupt for a second? So Frodo is going to be going on to Easter mode pretty soon. He’s heading to Japan, and he’s sent out an email message this morning and I frankly didn’t understand it. So before you take off on this, on this journey, can you explain to us what you’re talking about? When do we start this? What day? What time would that be for folks on the East coast? West coast?

Frode Hegland: Yeah, so hang on, I’m just going to paste something into our chat. Yeah. So. So what it is, is Monday the eighth and Wednesday the 10th, as well as Monday the 15th and Wednesday the 17th, meaning the next two weeks. Other than this Wednesday. Yeah. If we can change the time. So it’s 1:00 pm Pacific. 4 p.m. New York, 9 p.m. U.K. and 10 p.m. European. Because that way it’s 5 a.m. Japan. Otherwise I just wouldn’t survive.

Dene Grigar: I think it’s fine, but know that I teach on Wednesday starting at 130 and it’s I go all the way to 410. It’s a long class, right? So Wednesdays I can be there for 30 minutes and then I have to run downstairs and teach Mondays. There’s usually zoom meetings all damn day, like ten in a row. I’ll try to jump out and be here. But know that it’s hard by the time the day revs up.

Frode Hegland: Okay, so let’s make it this way. This Wednesday, we will nail down a few things based on today’s conversation. And then we’ll essentially have a two week holiday where those who can will be there and will chat. General will try not to be very specific. Sloan. And if something Sloan comes up, we will try to note it down clearly in the slack community. For the rest of us, whoever couldn’t be there to make sure we don’t make any inadvertent decisions or any nonsense like that. Does that make sense?

Dene Grigar: Yeah, we could move the Wednesday meeting to noon West Coast time, because then it would give me an hour and a half.

Frode Hegland: Yeah, that’s four in the morning.

Dene Grigar: Okay.

Frode Hegland: Yeah, yeah.

Dene Grigar: It’s going to be hard because you’re it’s going to be hard because you’re you’re going very far away.

Frode Hegland: Yeah. I don’t mind the four in the morning. In and of itself, it’s just the family likes to do things late in the evening. So I just have to follow the.

Dene Grigar: That’s okay. I’m just trying to give an alternative. That’s fine. And and we can spend two weeks working quietly. We need help. We can call a meeting or something.

Frode Hegland: Yeah. Essentially. Holiday. So. Right. So Okay, so my second question to you, Brandle. And to you, Andrew, this has been very, very useful to listen to. Is there a time that we can set aside? Where you guys. And maybe with Fabian or Adam, should it be useful where once Andrew has spent, I don’t know whatever period of time to go through it and really help him nail it. And if we’re going to have such a meeting like we kind of had Fridays, how long would you like to work on your own before we have it?

Andrew Thompson: I guess it would depend on what the topic of this meeting is. I can always use feedback. But it depends on what we’re talking about specifically. Well, if it’s if it’s a continuation of this about the saving and loading. I’m working on it, and I’m going to try to have, like, a base of it working Wednesday. So. We’ll see where I’m at then. Okay. But if we’re talking like the peer to peer, I haven’t even touched that. So that would very much be something that we can just pick a time.

Frode Hegland: Brandon.

Andrew Thompson: Maybe also post Wednesday, I suppose.

Frode Hegland: This will be a biggie on Wednesday in terms of kind of agreement day, I think.

Andrew Thompson: It’s also like the way I typically prefer to work when I’m on a concept I haven’t touched before is I’ll like I’ll gather the resources, I’ll try to implement it for a bit myself so I can get familiar with how code works. And then if I have trouble, I’ll bring up the question. So. Okay, I’m fine with meeting before I start, but I would personally rather start working on it first and then come and talk about it. Okay, that’s just me. I can do whatever people prefer.

Brandel Zachernuk: Yeah. And in terms of the timing, you would do you feel like it’s worth your time to try to chat on Wednesday? Do you want a little more time to think about these things? Chew it over.

Frode Hegland: The one step anyway.

Andrew Thompson: Right? Yeah. We’ll meet Wednesday anyways. Yeah. You’re having a separate meeting afterwards. I guess. I’m not sure where I’ll be. Yeah, I will have time. Oh, you want me to come to campus?

Dene Grigar: Campus? Yeah. Come to campus on those days.

Andrew Thompson: Okay. I probably won’t have time to meet then. I can I can be on campus. It depends on what I’ll be available to, like, message and stuff like that.

Brandel Zachernuk: Yeah, yeah. I mean, I can jump on slack anytime. Wonderful. I have a very busy day this third. This Wednesday, the third. So I may need to cut my my Wednesday attendance a little short or do it just over audio while I’m at Apple Park. I think some people are going to try to offer me some jobs, and I want to gracefully demur and try to have my cake and eat it.

Frode Hegland: No. That’s fine. I’m surprised you’re here today. Brandel. You did warn us. For the next few weeks, you will be very busy. So I’m very grateful.

Brandel Zachernuk: It was. It was the last. It was the last one. So I was up in Seattle last week for proposing model, and it went very well. So I’m really excited. Hopefully there I’ll be able to turn that into public documentation. I drew some teapots and I feel really good about them. I think they look really, just really good teapots.

Frode Hegland: Yeah. No, that’s that’s excellent. Yeah. I mean, the, the, the tenor or tone of this meeting is a bit different because we had a really good meeting last week. Well, two of them, one kind of taking stock and righting the ship. And now Dana and I are both putting in very different ways, putting our foot down and saying, okay, we’re going to make sure we’re walking in the right direction. I guess that’s why we put feet down, right to walk. Anyway okay, Andrew, you will work on this stuff when you are ready to ask my developers, just put together a paragraph and say, I need this, that and the other from you. And it’s my responsibility to make sure that they can do it.

Andrew Thompson: Yeah. Honestly, it’s it would work better the other way around. Because I just implement whatever you guys want, so.

Frode Hegland: You don’t need to give me a reason. Andrew, if you say that, that’s fine. What I do need from you guys, though And maybe. Randall, I hope you’re still here. What I would like to know is what to tell them. What in the world are they actually doing? Are they building a server inside reader, or are they. Is it something we drag, drop, drag onto a browser window? What do I tell them to build? How do you receive it? In other words, the actual JSON itself is fine. I’ll just tell it to do a JSON, but how to get it to you?

Brandel Zachernuk: I think it’s a little early to be able to give them an answer on that at this point. I think I would prefer to be a substantially. Understood. Solution within the web context on Andrew’s side before being able to, to make suggestions about the shape that a reader centric or a reader reader based interconnect needs to have in the broad strokes, it’s going to be the same. So it’s going to need to broker a connection via PJs. It’s going to need to have an identity, that kind of thing. But because the primary sort of scope of the, of the grant is to be sort of end to end within the Within the context of the. The web stuff, then I’d love to make sure that we have a sense of the broad strokes of it on that side before we make recommendations for people in a proprietary language to to to do that. Because my, my sense is just that, that code is inevitably going to be a little bit harder to change. And so getting a stronger sense of what it is in the, in the more flexible context is, is going to be instructive.

Frode Hegland: Okay. So once Andrew is ready to receive, as it were. We’ll look into it because it is in the spec, so to speak, that it should be possible for someone to send it up. So we’ll do it at the right time. Andrew, if you feel that I’m not delivering because of a misunderstanding now that Brandon has told me to pause please don’t feel it rude to repeat to me like we are ready. Now. You know, I may not have heard that. So whenever you’re ready.

Andrew Thompson: Absolutely.

Frode Hegland: Yeah.

Andrew Thompson: Yeah. We’ll probably talk more Wednesday. We’ll have more of a sense for that. But just to, like, reiterate, in case I wasn’t completely clear. The the content of what I’m given doesn’t matter as much because ultimately, I’m trying to just build what you guys want me to implement. And I can kind of make it work in different ways depending on what you envision. So I can’t really give you a list of, oh, I need to have these things. I need to hear first what you want, and then I can tell you what they need to give me. It’s kind of like a back and forth.

Brandel Zachernuk: Yeah. So that’s true. Andrew. But you’ll you’ll have once you’ve got the first bit of this done, you’ll know what you need to do, what you’ll need. You’ll know what you need to have in order to broker the connection between the laptop and the headset. And it would be ideal if that is the same style and channel of communication as happens between author reader and, and this framework as well. So it’s not so much that you’re the, the sole determinant because of your great ideas. Not that they’re not great, but that based on the, the expectation of the interop between those two clients that’s going to be the most future forward thing for the, the, the folks on the author reader side to be able to sign on to to, to make an addition or is it just up from I just see there and, and. I don’t know. Are you.

Frode Hegland: Is that a new hand, or is that an old Deanie hand? Deanie.

Dene Grigar: Listen. I’m sorry. What? Oh, I’m sorry, I didn’t mean to do that. It should have gone down automatically.

Brandel Zachernuk: Yeah, yeah, I just, I was, I just, I it was up and then it was up and then, and.

Brandel Zachernuk: Then I wasn’t sure if you were trying to get a word in.

Speaker7: I have a habit of.

Dene Grigar: Voting a thousand times at faculty Senate meetings. So. Yes, yes, yes.

Frode Hegland: So, so to address Brandel and Andrew on this and I’m glad we’re going in circles. These are very good circles. Yeah. The key is, in order to be truly open, the data has to be movable. And the notion of that is visual meta. Visual meta does not mean big tech and coding. It just means it’s a such a loose concept. It just means an introduction saying what the data is and then the data. Right. So to transfer an entire library, a contextual library, not the internals of every document in a JSON is entirely fine with that. And we will add a paragraph on top saying what it is so that an LM or whatever in the future can deal with it. Right? That is so core to everything we’re doing, because let’s say Adam does something completely different from your code, Andrew, he should still be able to access the same data. That’s really, really important. So that is why it’s very useful even now, Andrew, for you to when you’re defining your own environment. Because what do we want to do in your environment. Let’s look at it. And this is I think obvious but also very important to go through. Until we have passed through an XR, the person will be seated. They will be seated on something like a swivel chair. Whether it actually is or not, they’ll be seated at a central point and have a 360 degree view. That means that the primary thing they’ll be able to do is to move things about, and to tap on things to make things happen.

Frode Hegland: So the spatiality of the webXR is absolutely core. We will have a lot of lists because lists are useful, but things should be able to collapse and expand from that. So based on that, what you have now to be able to send that forget in a sense reader and author and all of that stuff for you to send that back to a thing and someone else to make a rich environment for that really is the key. So what you guys have been talking about for the last half an hour is fantastic to listen to. In a sense, since there is thoughtful silence. The Jason isn’t just a transport mechanism. It can also be thought of as a storage mechanism. It isn’t really, but it’s that important. I want to be able to get to a point where I compete with entirely commercial software that I make my two pennies on, and if you knew how little I made, you would buy me a beer right now. Or hopefully a coffee. But the point is, it should be possible to compete. But it has to be with the same data. And that’s why this incredible thing that you have been building so far under is so important to import stuff. So does that help you, Andrew, or that just that make you even more vague?

Andrew Thompson: No, no, that makes sense. Yeah. The data you’re talking about transferring between the different versions that we can all create here is now more defined than when we started this meeting. I thought you wanted just like something else that would we all be working off of? But, yeah, if we’re all working off the same sort of JSON exports that makes sense.

Frode Hegland: So great. So the second part of it is that architecturally, the way that I see it is that the JSON is the thing. What moves, how it moves is entirely secondary. It’s important to build now, but let’s say we have a great September and great end of year. And Sloan says this is fantastic. It has to be really secure, blah blah, blah. We can entirely change the mechanisms through which this happens, but the JSON stays the same. Plus additions. Because adjacent, it’s kind of cute. Obviously, we have the inventor of Jason in our community and the book. Jason is really cute. It’s just text saying what the world is. And how could that possibly be more core to what we’re doing, right? Literally writing it down. So a little bit to stretch the discussion a little bit on this. What we’re talking about for this library are in addition to the individual document. So in the JSON, each item in the library will be all the normal metadata plus any highlighted text in there because that can be used for views plus any entities. These are things we can discuss, basically useful lists of stuff inside to help us with views and other things. One thing that I dream about that’s not really for now, but I think we need to have a discussion, make sure it fits. Is having an NLM as being one item, obviously not internal as a thing, but referred to. So that means at some point we can build an environment of knowledge agents like we’ve talked about in, you know, in space all over the place, pushing and pulling and talking, and the library describes where they are, maybe even basic settings for how they were made or who made them, that kind of stuff. Right? So of course we can talk about 3D models and videos and stuff, but it should be anything that can make the environment more interactive and useful. Should be handleable by the JSON. It ain’t just references, obviously, or PDFs. We all agree on that.

Speaker8: And the great.

Andrew Thompson: Thing about Jason is if we decide to add more data to it later, that doesn’t break the older versions, it just adds more on top.

Frode Hegland: Yes, exactly. And furthermore things this whole what should be in there in this community. Over the years, we’ve talked about a lot of things related to timelines and places like DNA. And I had a fantastic week in Washington state. You were there, of course, for most of it, Andrew. But I could very easily imagine, like when we’re having sushi that day, if I add a thing to the library, it would be really nice if it knew when it was added as well as where, because this is private to me. So later on I can do that very basic thing of where’s that thing I wrote when we were having sushi in Washington state? These are really useful things that in this multidimensional environment, over time we can start to search for. Right.

Dene Grigar: Yeah. Step by step.

Speaker7: Yeah.

Frode Hegland: Step by step. I don’t want to stress you out, Danny, because obviously.

Speaker7: You’re.

Frode Hegland: Keeping the brains on this. All I’m saying is just to repeat what Andrew said. Actually, we have a basic Jason we add to it, describes what it is, and over time, we can have more dimensional views. That’s good.

Dene Grigar: I you know, I said earlier, I felt like a mother sometimes. Actually, I feel like a lion tamer. You down, lion down. Or maybe dominatrix. Who knows? Same thing. Get. Stop it, stop it.

Frode Hegland: Well, the You know, architecturally everything should be possible. Implementing is something else. But yeah, no, I think we’ve made real progress today. I’m very grateful. And also we weren’t.

Dene Grigar: We did not follow our, our our outline again.

Frode Hegland: Well, the outline today we don’t really have an outline but.

Dene Grigar: We did we actually did.

Frode Hegland: Yeah. We kind of we did.

Dene Grigar: We never do though. Froda.

Frode Hegland: But on on Wednesday we really need to at least nail down the core use cases. And one of the key issues is, you know, like we’ve been talking about forever in this community. Once we know that, where are we going to put it? Pdf, web page, slack, email, text messages in our heads through a dub. You know, once we have these, we’ll put it down several places so we can keep referring back to it. This is what we’re doing. This is what we’re doing. This is what we’re doing. And that’s it. We’ve added three months orienting now it’s, you know, following the wind and sailing. By the way, are any of you watching Shogun on Disney+?

Dene Grigar: No, I just finished three body problem.

Frode Hegland: Shogun is very good. And it gets better and better. We’re going to start three. Body. We started one. It’s good. Right?

Dene Grigar: Oh my God. Yeah. And it’s funny because I’m just finishing up a book called James, which is a A black Writer’s View of Huckleberry Finn, which is fantastic. As soon as I finished that book, I’m going to pick up three Body Problem and read the trilogy. I’m so into science fiction, it’s just like, it is so great. I’m now on the newsletter list, so I get the newsletters about updates and things. So it’s really, really good. Brandon, do you watch it? You’re shaking your head.

Speaker7: No, I.

Brandel Zachernuk: I read Three-body problem. I intend to watch it. I’m also going to pick up a Apple TV+ subscription and a bit so that I can watch some of the specifically also the Vision Pro content that’s being put out there 3D movies and stuff like that. But I watched I watched Moving Earth or Wandering Earth, I think it’s called it says it’s another it’s a Chinese language version version of a of a session. Lu story and similar to the Chinese version of Three-body problem because they did like a 40 episode version of it there. It’s amazing, but it’s also stunningly like adept, but pro-communist propaganda which is really, I don’t know, funny to see being so expertly executed. I learned yeah.

Dene Grigar: As I say, I learned I was reading about the book because I’m fascinated by the story. But what the author did in the book was they he put the Cultural Revolution information in the middle of it. Yes. So that it would pass the censors in the actual television series. It’s laid out the way he really wanted it laid out, which was it started the whole thing. Right. It kicks the whole thing off.

Brandel Zachernuk: That’s really funny. Yeah. I’m reading Too Much to Know by Ann Blair at the moment. And it’s a wonderful book for understanding information management. Anti-war. It was at at Marc’s suggestion. It’s a it’s a really fantastic book. Blair is a has she ever participated in the community?

Frode Hegland: Frode Hegland I did try to get to her. Tom Standage suggested but I didn’t hear back. And if we have other access to her, it would be wonderful. I do have her book right there. Yeah.

Brandel Zachernuk: Once I’m. Once I’m finished reading, I’m going to try to reach out to her, I think, because. And I’ve got some. I think she’s at Harvard. I think I think it would be really interesting because she did a, an interesting conference back in 2010 called Why Books? Which was about the past and future of the book as well. So, so a pretty interesting character. Too Much to Know is a great read. It’s not exactly a page turner, but I’m sure you’re used to that.

Speaker7: I live.

Dene Grigar: I live in that world. It’s funny, because I’m reading this book this this book by Percival Eliot, I think is his last name. He did the American fiction. Remember that? Remember the movie? He did the book for that. And so I’m reading his books now, and it’s funny because I can read a novel in like an hour. It’s like, so easy just to consume it, because then because I read so many things that is just so kind of take notes and I’m under. But even this book I’ve underlined whole passages is fantastic.

Speaker7: So, yeah, it’s it’s just amazing.

Brandel Zachernuk: Going back to fiction and realizing how effortless and fast it is and how hours can go by. Because you cannot do that with these books, no matter how well written they are. They you just you have to be aware of when they’re no longer going in. Anyway.

Speaker7: I have a I.

Dene Grigar: Have a habit of throwing books across the room when they make me mad. And so I have one called the Gutenberg Elegies, which was written back in the 90s, and by Sven Birkerts, in which he criticizes electronic textuality. You know, anything that’s digital and because you can’t smell it, you can’t touch it, you can’t taste it. I mean, it’s not like paper, right? Yeah. And I get like, I read five pages and I would throw it, I’d pick it up, read another five pages, and I’d throw it. So the book is like a freaking mess. But it’s such a visceral response to asinine, you know, commentary.

Speaker7: Yes. Yeah. No, I’ve.

Brandel Zachernuk: Realized that I’ve taken far too grim an attitude toward my books, and I’ve sort of bent and deformed this in ways that that make it much more legible to me as a, as a body of text. But one of the reasons why I brought up Blair is because she talked about somebody who purported to have reviewed 9000 books over a period of 30 years, and that’s just not not reasonable. This is in the early days of of writing of printing. But yeah, like it’s a it’s amazing what you can get away with by. And so would have actually read like a, like a small minority of those texts to any to any like at all, let alone to any level of depth. And it’s, it’s, it’s amazing what you can get away with by burying stuff in the middle.

Speaker7: So I’m going.

Dene Grigar: To say this, that was a literature major writing college and then went on for humanities after that. And so I’ve always been a reader in my as poor as we were growing up, we had a library. My mother had a library for us. So it’s not. I haven’t counted how many books, but 9000 wouldn’t be out of line for me. I mean, I know the college reading lists. I’ve read every any time somebody gives me one, it’s like, oh yeah, I’ve read all of those, I read all of those I’ve read. I mean, I just read all the time.

Brandel Zachernuk: But you wouldn’t have published. You wouldn’t have put your name on a published book review for all 9000 is what I’m saying. No.

Speaker7: I’ve only that’s over 100. Right, right. Exactly.

Brandel Zachernuk: So that was the that was the thing that was striking about it. It’s like that’s cheeky.

Speaker7: No, I should go back.

Dene Grigar: I mean that’s something you could do in your retirement is go back and review the book. In fact, what I do is I go back and reread books that I thought were hot and cool when I was young, like Hermann Hesse. I’ve read, like, everything Hermann Hesse ever read. I went back and read, you know Narcissus and Goldmund. I thought, well, this is crap. You know, 50 was crap, you know? So it’s kind of fun to go back and reread things.

Frode Hegland: It’s kind of interesting. The benefits of paper book reading. One of them is, of course, the rich annotations you can do by scribbling in them. And the other one is, I think when you read in the pages get looser, you own it more. That’s how I feel. But also you can hold your thumb in the spine and it’s easier to hold than a flat thing. But I just because I now have author and reader for the tiny iPad. Right? I have this amazing little stand. So when I put. I mean, how crazy is this to have a full sized keyboard and there’s nothing to carry around? It’s really nice and minimalist and elegant, but I do think we need to talk a lot more about what annotations mean. You know it should be possible to scribble. It should be possible to do it in a much richer way than currently. So maybe that was a huge part of what we might do.

Dene Grigar: There’s a book that I that I picked up years ago called marginalia, and it’s this person has done an incredible study on a margin, on marginalia, going back to the ancient period. The early texts that were held by the Moors in in northern Africa, Spain, in that area. They had actually written into Aristotle’s and Plato’s text. And when folks in the West picked up those texts, they thought that was Aristotle’s writing. So early marginalia was not even separated from the actual text. And it took a while for people to understand the difference between writing in a text and the text itself. Because for the West, everything was text. Everything was, you know, the same author. So it’s a wonderful book. And so I’m going to recommend that to you if you, you know, want to pick it up.

Brandel Zachernuk: That sounds excellent. And yeah, very much in line with the sort of managing of scholarly information. And so it’s, it’s all about note taking reference materials all that kind of stuff. And to realize that text hasn’t always just been text is something that it’s obvious in a hyper text world, but that there are centuries long traditions of having to manage this kind of information and reasonable solutions for doing so. Something that was embarrassed to say quite new to me.

Frode Hegland: I think this goes tails really well, back to our JSON discussion, because the notion of the JSON representing library. What is a library? Is it the outside or the inside of a document? It probably should be both. I would really like this JSON to include not only things we have highlighted in our own library, but also thick texts were written annotations and comments because that means we have access to them from the outside. And one thing in addition to this that I would really like to see is to have documents easily tell what they refer to. Citation is what a section refers to, but we don’t have a way for documents now to say that this document is in response to. Because that way we can start threading things even in our own environment, which will be interesting. So yeah, there’s a lot of stuff we need to keep in mind of. Useful data for us rather than just the basics.

Dene Grigar: So I need to go to the next meeting. So. Hey, Frodo, when do you take off for Japan?

Frode Hegland: Thursday. So Wednesdays are clean and happy day for me.

Dene Grigar: Okay, I’ll see you Wednesday. Brandel, thanks for coming. It’s good to see you again. And I want to thank you for being here today. Thank you for coming. And I want to thank you for yesterday. That was a wonderful experience for tonight. We’re going to report back to everybody about it. But we didn’t get a chance. But it was a wonderful experience. Thank you for the opportunity to let us talk about this project. It was the first time that we had the chance to kind of present it. And so we have made some pretty great slides, and it was our first dog and pony show. Right. So thank you for that opportunity. We’ve got a lot more ahead of us.

Speaker7: It’s great to.

Dene Grigar: See all of.

Frode Hegland: You. I’d like to second that and just add Mondays are usually all over the place. Right now we’re going through an existential crisis of the most fun kind. What have we done and where are we going? So we’ll be continuing to do Wednesdays project work Mondays all over the place. So until Wednesday for those of us who will join us Wednesday. Otherwise, Monday, Wednesday, next week at the weird time zones, everybody to email me whatever.

Dene Grigar: Thanks, Peter. Bye, everybody.