WEBVTT 00:00.000 --> 00:08.760 Brilliant, okay. Well, hi everyone. It's nice to see you. This is my first 00:08.760 --> 00:13.600 fostum, so I'm learning a few things. It's very, very friendly so far, so please don't 00:13.600 --> 00:19.600 change that. This torque represents a very similar torque that was given at 00:19.600 --> 00:23.120 Jupiter-Con. I don't know if anyone here has been to Jupiter-Con before, but if 00:23.120 --> 00:26.200 you were there and you saw that torque and you think it was better the first time 00:26.200 --> 00:30.800 around, I'm really sorry, don't tell me. So I have a few questions for the 00:30.800 --> 00:35.680 audience before we start, just for my benefit. Who has heard of a project called Jupiter- 00:35.680 --> 00:42.680 Book? Interesting, okay, cool. Who's ever been interested in designing web apps? 00:42.680 --> 00:49.320 Okay, also cool. Who's a content author or a researcher or anyone who might think 00:49.320 --> 00:53.600 they could fall under that bracket? Okay, that's, well done, that's exactly the 00:53.760 --> 00:59.880 I was hoping for. So last question, who believes it was a really bad idea to 00:59.880 --> 01:03.400 wear a turtleneck before giving a fostum torque in a rim of this temperature? That 01:03.400 --> 01:08.440 would be me, but we're here now, so. Right, so hi, I'm Angus. I've been in open 01:08.440 --> 01:13.360 source in some lurking capacity for over 15 years. After finishing my physics 01:13.360 --> 01:17.160 PhD, I moved into software development full-time with the research software 01:17.160 --> 01:20.720 engineer, although at the time I didn't know that's what it was called, and I'm now 01:20.720 --> 01:25.240 in applications and infrastructure engineer as an organization called 2ITC. So I 01:25.240 --> 01:29.120 won't be speaking much about 2ITC today, but if you've heard of my binder or Jupiter 01:29.120 --> 01:33.760 Hub, 2ITC are a non-profit that are heavily involved in these projects, and we as 01:33.760 --> 01:37.960 part of our work with communities, we do a lot of upstream work, and our job is to 01:37.960 --> 01:41.680 provide the communities that we work with, a digital home. So that's all I'm going to 01:41.680 --> 01:45.440 say about Jupiter Book. Sorry, no, that's not all I'm going to say about Jupiter Book. 01:45.440 --> 01:49.800 That's 2ITC, and like finally, because I use Nix OS, I have to declare that 01:49.800 --> 01:53.120 publicly. They tell you that when you start using it. So there we go. This is why I 01:53.120 --> 01:59.600 offer my demos, bro. Okay, there we go. Right, cool. So at a high level, Jupiter 01:59.600 --> 02:04.240 Book is a tool that enables you to take collections of documents like notebooks or 02:04.240 --> 02:08.960 markdown files and build them into a nice user-facing book-like experience. 02:08.960 --> 02:12.560 We kind of say it in the long form, which is that it lets you create enriched and 02:12.560 --> 02:17.360 interactive multi-document computational narratives, which is an absolute mouthful. 02:17.360 --> 02:20.240 So I'm going to talk to you a little bit about the current iteration of Jupiter 02:20.240 --> 02:23.680 Book, and we'll give you a bit of a sense of the journey that it's taken, and we'll 02:23.680 --> 02:29.800 get into more of the details as we go. So here is an example of a Jupiter Book with a 02:29.800 --> 02:33.480 community that we work with called Project Pithia. I really encourage you to look at it. 02:33.480 --> 02:37.560 It's a really interesting project, and I'm not just saying that. But crucially, Jupiter 02:37.560 --> 02:41.960 Book lets you take a collection of documents of documentation, which is what we have here, 02:41.960 --> 02:46.200 and assemble it into something that has a more natural reading experience. 02:46.200 --> 02:55.000 Okay, cool. So Jupiter Book's quite old now. It's been around for a good number of years, 02:55.000 --> 02:59.240 and it's gone through quite a few different evolutions. It started off with a tool that 02:59.240 --> 03:04.040 involved using Jackal in a pretty hacky-templating system to produce the content, and it 03:04.040 --> 03:08.320 went through a really big redesign, which was a grant-funded project, where it was built 03:08.320 --> 03:11.920 as a basically a distribution of sphinx, where there's an ecosystem of sphinx extensions, 03:11.920 --> 03:16.600 and at the end of it you would get what we used to now call Jupiter Books. And most Jupiter 03:16.600 --> 03:21.080 Books in the wild up till now have been these. There are still quite a lot today that are 03:21.080 --> 03:24.960 not using the latest version of the stack, which I'll talk to you about in a minute. But what 03:24.960 --> 03:30.640 I'm here to talk to you about today, if I try and switch sides, is this bit? So Jupiter Book 03:30.640 --> 03:36.560 2 was released in October, and we talked about it at JupiterCon. It's a whole new text stack. 03:36.560 --> 03:40.560 We don't use sphinx anymore. Lots of things are similar, something's a different, and 03:40.560 --> 03:46.480 I'll touch on a few of those today. Okay, get my step count in. 03:46.480 --> 03:47.480 What? 03:47.480 --> 03:50.480 If you off-camera, if you cross the line, the lines of the thought. 03:50.480 --> 03:51.480 Say again, sorry. 03:51.480 --> 03:58.240 The part of the room that's in camera, the camera over there, is in the, so if I want to 03:58.240 --> 04:03.600 make real gestures, it's over there, and then okay, cool. So one of the last things to 04:03.600 --> 04:07.000 say is we've gone through quite a big social change as well, and it's really important 04:07.000 --> 04:09.960 to think about the social parts of the project, and I'm really glad that the last speaker 04:10.040 --> 04:15.880 was basically talking quite a lot about that. We've gone from different organizational structures, 04:15.880 --> 04:19.400 different team members, and I will touch upon that later, but just wanted to call it out 04:19.400 --> 04:22.920 that that is not just the tech that's changed along the way. 04:22.920 --> 04:29.800 So next, describing the motivation reasons for how we got here, the one of the big gaps 04:29.800 --> 04:35.000 that we felt in Jupiter Book and in the wider ecosystem is the gap between the exploratory 04:35.080 --> 04:39.880 and the learning phase of doing research and of being a scientist or other sort of affiliated 04:39.880 --> 04:43.640 profession, and getting to the point where you're going to communicate that with your peers, 04:43.640 --> 04:47.800 whether that's sort of in an informal sense that you're publishing in an informal manner, 04:47.800 --> 04:51.400 or whether they're actually publishing to a journal, and anyone who's tried to publish to a journal 04:51.400 --> 04:56.440 using Sphinx before knows how painful that is, and so that's part of the technical and social 04:56.440 --> 05:02.200 reasons why we're here today. Finally, if anyone finds us, I'm going too fast because I've 05:02.200 --> 05:05.960 got really excited. Please raise your hand and I'll take it as a hint to slow down. 05:08.200 --> 05:14.360 So in order to talk about this topic, I'm going to break down the sort of space of personas 05:14.360 --> 05:18.360 into three different personas that we might imagine when we're thinking about publishing, 05:18.360 --> 05:23.240 about content, about science. So firstly, we have the learners and the learning communities, 05:23.240 --> 05:28.520 and these are people who need to gain new insights, to gain new skills, to learn the tools of 05:28.520 --> 05:32.680 their trade, and usually we think about undergraduates, but this is a very restrictive view of 05:32.680 --> 05:37.240 what is a very wide persona. Once these people have gained these skills, they've got their 05:37.240 --> 05:41.800 networks, they've started to integrate themselves into the researching sort of portfolio. 05:41.800 --> 05:47.640 We then think about research communities, and these are people who are becoming explorers and 05:47.640 --> 05:52.760 researchers and collaborators. They have people that they work with, they have science that they're doing, 05:52.760 --> 05:56.040 and they are building upon the skills they build from this previous section. 05:56.440 --> 06:01.720 Typically, these personas are coming up with their own experiments and their own ideas. 06:01.720 --> 06:05.960 There's a lot of idea generation happening in this middle phase. And then finally, we have 06:05.960 --> 06:10.440 the scientific communication phase. You've created something polished or not as most of my 06:10.440 --> 06:14.840 publication would have been, and you want to share it externally or formally with your peers. 06:14.840 --> 06:20.200 You want to communicate it crucially using the same tools. Anyone who's done publications before 06:20.200 --> 06:23.160 will know that you might have a tool chain that you use all the way up to the middle, 06:23.160 --> 06:27.880 and then suddenly you have to learn later, like the final year of your PhD, depending on who you are. 06:27.880 --> 06:32.200 So, we really want to recognize that that is a big jump in the technical stack that people are 06:32.200 --> 06:36.920 using, and we want to try and, as a Jupyter book project, make that easier. So, we want to 06:36.920 --> 06:42.040 single tool chain. If I can come up with a catchphrase that can realize the goal of having an 06:42.040 --> 06:49.240 entire knowledge creation life cycle captured within its remit. So, oh, look at that, Mr is an ecosystem 06:49.240 --> 06:54.680 of tools and standards designed to meet. These needs, here we go. So, the idea is that the 06:54.680 --> 07:00.520 missed project or the missed ecosystem comprises tools and standards of composable building blocks 07:00.520 --> 07:04.040 that come together to serve the different parts of that life cycle that I just showed you. 07:04.680 --> 07:11.000 If anyone was in the, which talk was it, there was a talk yesterday that was on Dock Speck 07:11.000 --> 07:14.680 or something like that that talked about Dock and ASTs. If you happen to have been there, 07:14.680 --> 07:18.840 it's very relevant to this, and I'm keen to meet up with the author later, but for now, 07:19.160 --> 07:24.920 there's an interesting aside. So, Dock book crucially does three things when you run a build on a 07:24.920 --> 07:29.880 Jupyter book project. It parses in resolves all of your documents, so if you have input documents 07:29.880 --> 07:35.560 like mark down files or Jupyter notebooks, it takes them, it reads them and ingest them, and then 07:35.560 --> 07:40.280 produces an internal representation called the missed AST. Now, that's always the way that these 07:40.280 --> 07:45.320 sorts of tools work. What's different about Jupyter book, too, is that the missed AST is considered 07:45.320 --> 07:51.240 an asset that we care about. It's versioned, it's schemered, it's shared, and it's a structured representation 07:51.240 --> 07:57.400 of your documents. I think one of my jobs in this talk is make everyone hear care about AST, 07:57.400 --> 08:03.160 which I think is an uphill battle, but I'll do my best. And then finally, once we have this AST 08:03.160 --> 08:06.680 across all of the different documents in your project, so you might have different pages in your site, 08:06.680 --> 08:12.360 for example, we then want to render and export them out into your target format. That might be 08:12.360 --> 08:17.560 a website that's interactive to new users. It might be a PDF that's being published in a journal, 08:17.560 --> 08:21.960 and you can imagine, from those two statements alone, there's quite a lot of things to think about 08:21.960 --> 08:27.800 when you go from the AST into the rendering step. The benefit of having these separate phases 08:27.800 --> 08:33.240 amongst other things is that we have different extensibility points, so you can change how the 08:33.240 --> 08:37.640 documents are passed and you can add new features there. You can modify what happened when we have 08:37.640 --> 08:42.040 these documents and do interesting things. So for example, Jim from the cheering way, who's one of 08:42.040 --> 08:50.920 the organizers of this session, he don't think he's here. He added a plugin that he had a plugin that 08:50.920 --> 08:56.200 makes it possible to check, do your images have alt texts for an accessibility checker. And those are 08:56.200 --> 09:00.360 the kind of things you can do on a structural level by having an AST that you sort of share and build 09:00.360 --> 09:06.840 tooling for. It's the advice to say, you can also extend the final part. That's sort of a given, 09:06.840 --> 09:12.760 you can have different templates, you can have different PDF themes, and I'll talk more about that in a 09:12.760 --> 09:17.880 minute. So to build a new stack here with the guiding principles, we wanted to make something 09:17.880 --> 09:22.680 that was simple to extend and easy to use. We wanted the content to be canonical and machine 09:22.680 --> 09:26.840 readable. I love the word canonical, but ask me what it means, and I'll change my answer depending 09:26.840 --> 09:31.480 on the time of day. It does have a meaning, and then content should be modular and composable, 09:31.480 --> 09:35.400 which sort of connects to number two and will come on to that very shortly, and then computation 09:35.400 --> 09:40.520 should be a first-class citizen. I come from physics, and I'm sure most people have a connection 09:40.520 --> 09:43.960 with computation given that we're here, so computation is a first-class citizen. It's not 09:43.960 --> 09:48.600 something you do sort of as a bolt on or on a side. And then finally and most importantly, and I'm 09:48.600 --> 09:52.520 not saying that because of the last speaker, it should be governed by an open community. 09:54.680 --> 09:59.480 So these are our goals, feel free to ask me whether we've achieved any of them, and this is the 09:59.480 --> 10:03.960 first one. So Mr. Superpower should be this easy to use and simple for users to extend. 10:04.920 --> 10:09.240 The goal is to be simple and approachable for someone without a ton of technical expertise, 10:09.240 --> 10:13.720 but with the capabilities to extend into those for new use cases and bespoke needs. 10:15.880 --> 10:22.360 Easy to use. Document format. Everyone goes to mark down. It's ubiquitous. It's used everywhere. 10:22.360 --> 10:26.600 It's not used by the same dialect everywhere, which is really annoying, and I'm not going 10:26.600 --> 10:29.800 to tell you that we didn't create a new format of mark down for Geewed to Book because we did. 10:30.120 --> 10:36.440 And that's a good thing. But we start there, right? So here's regular common mark mark down, 10:36.440 --> 10:41.800 and it has mentioning features. You have title, italic text, bullet points, block quote, math, etc. 10:42.760 --> 10:47.960 And it's supported all over the place, but when you try to do more than this, it gets a bit tricky. 10:47.960 --> 10:54.040 So we can render this in HTML and look pretty nice. But imagine that you want to, for example, add 10:54.040 --> 10:58.440 a figure. So here's some common mark that represents a scientific paper, and you've got a figure 10:58.440 --> 11:03.640 one that corresponds to a graphic with a caption. To actually do this, you need to take what is 11:03.640 --> 11:08.840 a semantic concept of a figure and render out into mark down yourself. So you can't say that this is 11:08.840 --> 11:12.760 a figure with a caption what you actually often do is create an image node, and then below it, 11:12.760 --> 11:16.840 you put some italicized text with a figure one written there. And that works really well, but there 11:16.840 --> 11:20.680 are some challenges with it. Firstly, if you add a second figure above it, now you've got to remember 11:20.680 --> 11:25.640 to redo all the figure numbers, which I never did. And secondly, there's no real 11:25.640 --> 11:30.280 tool-level information here that what this is is a figure, and that's a more nuanced point that 11:30.280 --> 11:36.280 we'll get on to in a minute. So I'm just going to make my speaker note to not size 105. There we go. 11:37.320 --> 11:41.080 I think I've covered everything. Yep, so mainly this is fine, but it's not good enough. 11:42.040 --> 11:48.120 So mist mark down, we'll have a note. Mist mark down is a evolution of common mark. It's a 11:48.120 --> 11:51.400 super set of common mark, which means that we do all of that, but we also have some additional 11:51.400 --> 11:55.720 features. And one of the ones that I'm going to point to you is, oh, we have in line math, 11:55.720 --> 12:00.440 brilliant. That's not very exciting. What is exciting is we have block-level math as well. This is 12:00.440 --> 12:06.280 a directive node. If you've heard of MDX, it's kind of like an MDX component. If you've heard 12:06.280 --> 12:10.920 of programming, it's kind of like a function. And so here what we're doing is describing block-level 12:10.920 --> 12:16.760 content, and it has a name, it has an option that's been passed in called label, and it has a body, 12:16.760 --> 12:21.320 which in this case is a math expression. But it's not limited to to that. So that you can 12:21.320 --> 12:27.240 see this as one way that we can build extensible reusable building blocks. And we have quite a few 12:27.240 --> 12:31.240 things here, so we've got a math directive. Now here's the figure that I was referring to. So 12:31.240 --> 12:37.640 figure, we have a name, an image, some formatting information, and an exception. I can then reference 12:37.640 --> 12:43.320 that figure in my prose later on, and it will do interesting things I'm going to show you. So 12:43.320 --> 12:47.080 having that semantic information from that's a figure, and it's got a label, it's already enabling 12:47.080 --> 12:52.680 us to do interesting things. And we actually use the same syntax for citations. You can see over here, 12:52.680 --> 13:00.920 here's a citation in a bibliography, and that's using the same lexical approach to build that citation. 13:03.000 --> 13:10.280 There we go. So because this AST is enriched, if I have a link to a citation, I can hover over it, 13:10.280 --> 13:14.760 and if I have a link to a figure, I can hover over it, and it brings the content, that's not 13:14.760 --> 13:22.600 very useful. It brings the content into the foreground, here we go. So I'm talking about something here, 13:22.600 --> 13:26.360 and usually the context of where I'm talking about something is really important. If you have 13:26.360 --> 13:29.800 a hyperlink, I don't want you to jump away and go and read that page, because then you'll 13:29.800 --> 13:33.400 lose the context of where you've come from, and usually I'm embedding a conversation about 13:33.400 --> 13:38.680 figure in some wider prose. So the useful thing with these hover references is that they are 13:38.680 --> 13:42.600 local, so you don't have to move away from where you currently are to learn more about the figure. 13:44.760 --> 13:50.600 So another super powerful myth is that it's a canonical and machine readable. So every missed page 13:50.600 --> 13:55.400 has adjacent equivalent, and you can go onto that with curl or whatever your favorite utility is, 13:55.400 --> 13:59.720 and you can look at the structure of the page. So we have lots of metadata here, like your title, 13:59.720 --> 14:04.440 and your authoring, and your licensing, and we also have important stuff like what is the content 14:04.440 --> 14:09.240 of this page in a structured way, and that's really excited from a reuse and sharing standpoint. 14:09.400 --> 14:17.160 I've already said I'm going to make you care about AST. The first step to doing that is that once 14:17.160 --> 14:21.800 you have this AST, you can then do interesting things with it, so you can export to different targets. 14:21.800 --> 14:26.280 You could build through a PDF or Word document or a late-tech document. You could also build 14:26.280 --> 14:31.240 to type it. You could also publish a website, like I've just shown you, or author to publish to a 14:31.240 --> 14:35.560 journal. There are lots of different ways that we can reuse this, and it's because we have an AST 14:35.640 --> 14:40.840 at the center. There are lots of templates that could mean the members of author to target different 14:40.840 --> 14:45.560 journals and different formatting styles, and that list is getting longer by the day. 14:47.720 --> 14:52.520 So here's an example from a company called Curve Note. They are members of the 14:52.520 --> 14:57.000 Jupiter organization, and they are heavily involved in the early development of missed. They 14:57.000 --> 15:03.720 worked with Cypai in the last couple of years to digitize all of, oh, not digitize, but to massage 15:03.800 --> 15:10.200 all of their proceedings into this format. If you go to Cypai proceedings, oops, that's too quick. 15:10.200 --> 15:17.720 If you go to Cypai proceedings, you can see. Yes, we've got all of the, I think 18 years, I'll say 15:17.720 --> 15:22.760 we, they have got all of the 18 years of proceedings in this format. I think it's over 1000 articles, 15:22.760 --> 15:28.520 and they will serve under this one, they are all served under one front-hand interface. 15:29.480 --> 15:33.560 So you can browse through, and because these are sort of digital artifacts, they can embed 15:33.560 --> 15:38.280 interesting things like figures and plots that you can actually drill down into and learn more about. 15:38.280 --> 15:41.880 This is kind of only scratching the surface, but you can already see how like having a medium 15:41.880 --> 15:46.280 that you're intentionally publishing for is really powerful from a exploratory perspective. 15:51.080 --> 15:56.680 Here's the Cypai proceedings template. I think it's very pretty objectively, and this has 15:56.680 --> 16:01.080 made pot of having a template that we've built to from the missed AST, and I'll talk more about 16:01.080 --> 16:05.080 the detail in a minute, but if you're targeting typist, which I think this template was, 16:05.080 --> 16:08.920 it takes one order of milliseconds to build that PDF rather than the minutes you're familiar with 16:08.920 --> 16:15.560 latex makers, etc. No shame on latex make, by the way, got me through my PhD. So this super 16:15.560 --> 16:21.800 path number two is that it's modular and composable, so content modularity is kind of nuanced, 16:21.800 --> 16:26.840 but the idea is that if you have cross references and hover references, we can hover over them and 16:26.840 --> 16:32.280 bring them into the local content and contexts, sorry, to explore them. So for example, here are 16:32.280 --> 16:38.360 cables that are highly detailed, and then you can also have citations that refer to your bibliography. 16:39.640 --> 16:45.080 And these are all internal references. What's also more exciting is that same mechanism will work 16:45.080 --> 16:49.640 for external references. So you can imagine that if different people from different communities 16:49.640 --> 16:55.960 publish their sites as a bit missed sites, as Jupiter book sites, then it is possible with 16:55.960 --> 17:01.800 these into this new tooling to embed or hover over references to that content in your own. 17:01.800 --> 17:07.240 And the goal here is to think about, there we go, the goal is to think about federating content. 17:07.240 --> 17:10.680 So you can talk about distribution content across multiple different providers, 17:10.680 --> 17:14.680 and instead of having to copy and paste a figure into your paper to say, oh, this person did something 17:14.680 --> 17:18.760 cool, you just embed it in all of the citation and attribution metadata should be there. 17:18.840 --> 17:23.320 Now that is a goal, we haven't got all the attribution done yet, but for now you can certainly 17:23.320 --> 17:27.640 embed and hover preview things some other people's resources. And what this video was showing is that 17:27.640 --> 17:32.840 there's a, I think this is a kind of project-pithy cookbooks, no, this one isn't, but effectively you 17:32.840 --> 17:37.160 can have, I'll show you the proper one, you know, you go, project-pithy here. 17:39.080 --> 17:44.280 Here is, they had project-pithy had these tutorials called cookbooks, and the idea is that you can do 17:44.280 --> 17:49.080 sort of specific analyses here and share them. But all of these analyses are kind of doing the same 17:49.080 --> 17:53.240 thing in certain context, they all need to run on a Jupiter hard, and they always do x, y, and z. 17:53.240 --> 17:57.720 And so what you can do is the authors of project-pithy here can basically publish a guide, 17:57.720 --> 18:01.560 and then that guide's material can be embedded and referenced in these cookbooks when they're 18:01.560 --> 18:06.120 providing guidance for how do I set up a virtual environment or how do x, y, and z. And that's 18:06.120 --> 18:09.800 a moving target we'd love to do more on. I'm rushing through it because I'm aware I'm probably 18:09.800 --> 18:12.680 short for time, but please do asking about that at the end if you're interested. 18:14.680 --> 18:18.520 Computation being a first class citizen is something that we're thinking about all the way through 18:18.520 --> 18:24.840 the Jupiter stack. So this includes Jupiter lab as an editor. We want to try and blur the lines 18:24.840 --> 18:30.280 between code cells and mark-down cells, for example, so that you can have a mark-down cell that defines 18:30.280 --> 18:35.240 some code that's coming from a code cell. So I think down here we have a widget in-line in some 18:35.240 --> 18:40.840 text and we have a computed expression that comes down here later on. The idea is just to sort of 18:40.840 --> 18:45.160 play around and see what happens if you think less about a linear thing of blocks and more about 18:45.160 --> 18:49.480 a document that can live and breathe. Jupiter lab missed needs a bit more development work in the 18:49.480 --> 18:54.680 last couple of months, and that's one of my projects, but the goal here is to think about how we 18:54.680 --> 19:00.680 can reduce the barrier to interactive and rich experiences. Here's an example from microscopy, 19:00.680 --> 19:04.600 static figures aren't great. If you can drill down to them and you can actually zoom in and switch 19:04.600 --> 19:08.920 the column apps, etc. You can learn a lot more than a printing seven figures out statically in a 19:08.920 --> 19:14.120 PDF scan, and so it's really worth thinking about how you can lean into embedded and rich experiences 19:14.120 --> 19:19.400 and publishing for this kind of target. Some people have deployed compute resources like 19:19.400 --> 19:25.240 jute hubs and binder hubs and project pithia is an example, and project pithia cookbooks, for example, 19:25.240 --> 19:29.480 some of them can run on a binder hub, so here if we take the power button it's going to connect and 19:29.480 --> 19:33.400 spin up a binder instance, it's going to run the jute decernal for the current document, 19:33.400 --> 19:37.720 it's going to compute all of the cells and provide you with any rich interactive figures that 19:37.800 --> 19:42.280 might be there, and this is really useful from the idea of reproducible analyses that people 19:42.280 --> 19:47.080 can actually test out and share, and there are lots of nuances here again, I'm going to skip over them, 19:47.080 --> 19:50.920 but suffice is that it's really fun being as press a play button on the website, and that's the 19:50.920 --> 19:56.920 main reason we all do this, so there we go. Another mystery part is that it's super easy to use for 19:56.920 --> 20:02.840 developers to extend, so here's some very quick examples. I'll see you in turn, I think with 20:02.840 --> 20:07.240 curve note, created as ammo bar on the right hand side, every time you click on a figure, 20:07.240 --> 20:10.840 it has this trailing list of figures on the right hand side because you might argue that the 20:10.840 --> 20:17.080 figures are the foreground of the paper that you're reading. This was done by the fact that 20:17.080 --> 20:22.840 missed is built in two ways, I'm kind of talking about here, but it's got the JSON state that's 20:22.840 --> 20:27.320 the documents, and then it has this renderer that's currently a React front end, and that renderer 20:27.320 --> 20:31.320 can consume that state and render a different context, so you can either modify the renderer 20:31.320 --> 20:36.680 to do more things, or you can bring it into another context, so here, for example, is a website 20:36.680 --> 20:41.160 time that someone has embedded some section of missed in because they're just pulled in the 20:41.160 --> 20:44.760 React components and they're rendering the document there, and there's all kinds of things from 20:45.480 --> 20:51.480 CDNs and galleries that you can imagine doing with the ability to pull in parts of a published 20:51.480 --> 20:58.520 duplicate book site somewhere else. Sometimes you want to pull in bespoke visualization, so here, for 20:58.520 --> 21:04.040 example, is a sequence of visualizer. This is the kind of thing where you might want to embed 21:04.040 --> 21:08.120 a little widget that people can share and reference, and we're working on ways to make it possible 21:08.120 --> 21:12.520 to embed little of absolute things in your missed sites to be able to do this, and this is an 21:12.520 --> 21:16.840 example where we just hacked our own code to make it work, but missed is permissive enough to let 21:16.840 --> 21:21.720 you do that. I'm going to skip this one, which is not super important, let's keep going. 21:23.960 --> 21:28.120 You don't get to see that demo, sorry. Right, finally, and most importantly, I don't want 21:28.120 --> 21:32.040 to rush through it, led by an open community. So for duplicate work and for missed, 21:32.040 --> 21:36.440 communities, the most important thing, we want to be an open community that is governed by 21:36.440 --> 21:40.840 a collection of peers, a collection of organizations, and individuals that all work together. 21:41.880 --> 21:46.760 So we have a team compass, and in most duplicate projects, have these, and it describes the 21:46.760 --> 21:52.120 internal structure and organizational practices of how we work and what drives us forward. And we have 21:52.120 --> 21:56.120 it here so that people can understand the way that we work and the way that we operate, and 21:56.120 --> 22:00.280 see a path for them that they can start to consider how they might contribute. 22:00.280 --> 22:03.560 Whether that's reviewing pull requests, whether that's adding new issues, whether that's 22:03.560 --> 22:08.680 triaging things, whether it's checking for typos that I've made in the docs. If we have an open view 22:08.680 --> 22:11.960 to how we work, this should make things easier for people to engage for the first time. 22:12.840 --> 22:17.160 So this is the core team. I'm here. That was not apparently my partner says it's a terrible 22:17.160 --> 22:22.840 photo of me, but I've put it in anyway. So we have two ITC curve note, a black-cenship data science 22:22.920 --> 22:28.360 amongst other collaborators here, and the list is growing. Crucially, we have a much larger group 22:28.360 --> 22:32.520 of frequent contributors outside of the core team. So there are lots of different organizations and 22:32.520 --> 22:36.680 people driving the project forward, and that's one of the really exciting features that we have 22:36.680 --> 22:41.560 as part of Juvedberg. So you can learn more by looking at some of these links. If it sounds like 22:41.560 --> 22:45.720 something you'd like to be involved with, I'm already drowning in GitHub notification so a few more 22:45.720 --> 22:51.160 won't hurt. Please feel free to ping me, and I'd really love to hear from you. I don't know how much 22:51.240 --> 22:56.600 time I have left. Five minutes. Do I have time for a short demo? Always a time for questions. 22:57.240 --> 23:04.840 Okay, great. So I've got two tiny demos. So if I can figure it out, there we go. 23:10.680 --> 23:12.520 It's not, you're not supposed to be saying anything, don't worry. 23:12.520 --> 23:36.920 Okay, let's try this. Whoops. La-ba-ba-ba. Cool. Right, say. Cool. So here's a Juvedberg website. This is what 23:36.920 --> 23:41.080 you might see when you start up right in some content. It's about fake birds, and then on 23:41.080 --> 23:48.440 how to the fake bird conspiracy. Cool. I'll show you more later. This is a Juvedberg page. It's got 23:48.440 --> 23:53.000 lots of interesting markup. You've got this of definition lists. These which previews that I told you 23:53.000 --> 23:57.800 about, sorry, Mr Bird, out of the way, and then also I have some code in here. That's a really 23:57.800 --> 24:02.920 not good look of my face that I took earlier. So you can embed all kinds of stuff in here. You 24:02.920 --> 24:10.440 can do all kinds of interesting things, and the source code for this just because I want to show you 24:10.440 --> 24:19.400 was if I open up my file, there we go. So we have some information about what's running the code, 24:19.400 --> 24:25.160 some abbreviations, my name, because I'd like to see that everywhere. The title, some code that gets 24:25.160 --> 24:29.960 then embedded in a subsequent expression, an interesting figure that wasn't very interesting, 24:29.960 --> 24:34.680 and it's a finally a code cell on the bottom. You can see how actually the amount of markup there, 24:34.680 --> 24:38.840 although probably I'm familiar to some people, is quite thin. What's captured in there is more 24:38.840 --> 24:42.680 interesting is the AST, and that's the bit that I was spending all this talk talking about, 24:42.680 --> 24:46.360 but I've run out of time to go any further. I'll skip the other conspiracy theories demo, 24:46.360 --> 24:57.640 and I'll hand over to questions if anyone has any. 24:57.640 --> 25:11.720 Yes. Yes. So there is support for Jupiter light already. Computations that you really 25:11.720 --> 25:15.880 know, there's quite a lot of interesting parts to it, so there are projects like end script and 25:15.880 --> 25:19.880 forge, which we don't support yet that I would like to, because that makes baking actually complex 25:19.880 --> 25:24.840 environments a lot easier. There's also three kinds of ways you can execute with Jupiter book. 25:24.840 --> 25:29.480 You can execute locally when you're running the build. You can execute in browser using the 25:29.480 --> 25:33.080 binder mechanism that I showed you, and then there's a third one that I've temporarily forgotten, 25:33.080 --> 25:38.200 but suffice to say there are different ways you can think about computation, and Jupiter light is 25:38.200 --> 25:43.640 pertinent to the in browser form. We also support passing in a Jupiter server URL, so if you can 25:43.640 --> 25:49.000 start a Jupiter server on our remote binder or on Jupiter gateway, you can run it in CI, but use 25:49.000 --> 25:52.760 a binder hub to actually build the book and do the complex computation, which is something that 25:52.760 --> 25:56.600 too ITC worked on recently. So there's lots of interesting places to take this, and it's quite 25:56.600 --> 26:00.120 nascent, so I think there was a lot of interesting runaway we could explore, so thank you for your 26:00.120 --> 26:03.240 question. You know what, last point. Oh, sorry. 26:07.240 --> 26:09.720 The link I will go back to, and then can I ask you off? 26:22.920 --> 26:24.440 Is there anything special? 26:33.080 --> 26:34.600 Well, the digital. 26:34.600 --> 26:44.600 a markdown in the east or to provide a pipeline of documents on this case and for the best 26:44.600 --> 26:49.600 parties of the talk before we communicate. 26:49.600 --> 26:53.600 So my understanding is about an editor is something we'd like to have. 26:53.600 --> 26:57.600 We haven't done a lot of work there yet, but members of the Dubb Book community have worked 26:57.600 --> 27:01.600 on making like a rich whip what you see is what you get editor for this. 27:01.600 --> 27:07.600 Injusting a dock ex, like a word file, and then using it as an input file, we don't support that yet. 27:07.600 --> 27:11.600 But from having seen other talks this week, I don't think that's incredibly difficult. 27:11.600 --> 27:13.600 And it's something we probably could do. 27:13.600 --> 27:17.600 The difficult part is figuring out what subset of word documents you want to support, 27:17.600 --> 27:20.600 and the more restricted you can be kind of the easier it gets, really. 27:20.600 --> 27:21.600 Does that answer your question? 27:21.600 --> 27:26.600 Let's talk after this. 27:26.600 --> 27:27.600 Thank you. 27:27.600 --> 27:40.600 Thank you.