 here. This is the Fellowship of the Link on Wednesday, March 8, 2023. And we're talking about a small project I have going to do shared notes across six tools with six different people. And Chris, you're one of my hopeful volunteers for this project as well. And we're just talking with Aram, who is very interested in this kind of interoperability. And it's kind of a low hurdle goal, which is to take a simple format like Markdown. And so I created the page that I'm sharing in the chat. And the idea is, can we connect through this? And then what does it mean to share notes? And how might we improve that? So the low bar that I set was, let everybody link to this page from their note-taking system and then make a change to the page by adding a link to their system on the page, and then either make a pull request, send me the edit, or in some other way, cause that change to happen. And that may be too low a bar, or it may be confusing. I don't know. So I think that's a very good bootstrapping method. So building on Mardon and Git as the MVP, yes. And it can be extended in the direction of the inter-wiki linking that we have discussed at some point. And I know. So Timur, who joined, I think one of our goals only, has inter-wiki links in my Corrisa. Maybe we can build on that. There are conventions. My proposal, also building on that, or complementary to that, is you probably can see coming that we take the same repositories and also just put them in an Aivrata. Right. So my understanding is that your Aivrata will wrap itself around anybody else's kind of repositories. And so part of my question is, what does that mean, and how does that play out? Cause you just shared a link to, you just did that. Yeah, yeah. In general, as long as you submit, as long as you submit a repository to an Aivrata, and of course, an Aivrata is a reference one, but we can, I think will be maybe to set up a particular Aivrata for this experiment. So it's easier to play around and experiment with that. Once the repositories are in, what the Aivrata will do is like any time you search for something, it will volunteer every page that matches that something in some way that we can define, one after the other. And if you link using wiki links or hashtags, so anything you use, any wiki link that you have in your local repository will, by default, link to every other page with that name. So essentially the Aivrata optimistically promotes local links to global links, links. So in that sense, linking to share nodes across six tools in the Aivrata will not require linking to wiki relay dev slash share nodes across six tools. It will be sufficient to say wiki link, so square rack and square rack, share nodes across six tools. That link will be equivalent in the scope of the hour. Is that sort of like a meta backlink? How is that different from a backlink? And unfortunately, I wish I understood translation better, but I don't know where the limits of boundaries of translation are, but how does that... Right. So essentially, what happens, I have actually a diagram. Perfect. Showing, yeah. I can, if you want, show you quickly. So essentially, this is from my, yeah, this is the pre, the draft for the new program, kind of read this. This is the draft of what's going into print apparently in the, here we go, in the personal knowledge graphs book. I don't know if you can see this. Yes. Yeah. So essentially, what happens if you think about this, you know, so they are going to introduce a notion of node and node being like a nexus to some extent, each of these, about something. So it can also be thought as mapping to a search query. You search for something in an hour, you are visiting that node, which may be empty, but you know, essentially, the results page is like the node in our terms, because it's a graph, or we try to see it as a graph. And in each node, you have many resources. So it's like a multiplicity. So for example, like if we all, if, you know, in relate, you have shared nodes across these tools, and I have like, like the same page in my garden, or in another repository, they will, your page, and like my page will be sub nodes, so different resources in the node title, share nodes across these tools. And then if in your wiki, for example, like I see in relate, you have a link to tools for thinking, right? So in this case, that will look like this, right? In the sub node, share nodes across these tools in your, in relate wiki, you're linking to the node tools for thinking, right? In your garden or wiki. But the hour actually takes that and says, okay, if there are any links between repository, within a repository, I will actually build a link at the hour level, essentially by linking something in your context, you're contributing that relation to like the higher level graph. And this is essentially the core way in which we build the graph. Any link at a lower level gets promoted to a higher level by default. So just trying to mirror what you're saying and see if I'm understanding it. So this is kind of like a smart query, like a standing query that will find the same text string in whatever documents happen to mention that text string. Actually, it could do that. By default, it actually rather matches only the title of the documents. Oh, okay. So it's not looking into the text of the documents. Right. If you want to link, there is like a full text search. But that's how it's different from a permalink or a backlink, because the backlinks are actually text within the documents. And you're looking for the same, you're looking for people who've called something the same document. Exactly. Yes. I mean, a module of some things, right? Like you can have spaces or underscores or dashes. For the hour, it's all the same. Yep. It will match, like do some passing matching. And the links are built, precisely, the back and forward links are only the actual links you have in the resources. Got it. And are you looking for mistaken accidentally similarly named documents also? I mean, you're really looking for the same document that different people have ripped on, right? Exactly. Yes. So this is for something like the reference hour has only about 26,000 nodes. But like already, you can see that, you know, this is all just multiplicity that happened. Sometimes, because as a community, we said, like, oh, let's know in this location. Sometimes just because of serendipity. And this is, yeah, so for example, like, you know, I guess, you know, we have like 14, in a way, we have 14 definitions of the favors, because we have 14 resources that are titled precisely favors. But you don't have a way of showing whether some of those are like the edit of a comma versus a different riff on the definition. You're just listing them all, right? Right. Yeah, exactly. Because to some extent, these are all unrelated to each other. They're all like sort of like, it's like a bag of things. Uh huh. So there's no notion right now of one is a version of the other. You could definitely have one like, you know, fork, like a particular resource, right? But we have another flow. Okay. And FedWiki sort of takes that approach is FedWiki is very much about versions of individual pages that then get shared by other people and cloned out. And I again, I still don't don't understand the sort of the promiscuous replication of pages of FedWiki. But it's trying to follow on a on a page track. Chris, go ahead. So hopefully you can hear me at least. Yes, totally. Like your next door. Good. So I know I am not yet in the Agora officially with a repo, but I am in there with respect to being tracked by the bot through Google and through Mastodon. So I can communicate into the Agora that way. And that's one way communication. And other than the bot saying, here's the page and the link back to me. I get no other communication about anything else that happens within that space, which is good in some sense. I, you know, at least I'm notified. Here's your thing. You can see the link and go. But if you tomorrow change something within that Agora, I get no notification. I don't know about it. Unless I choose to go back, which notification wise is fine. Now, if I did have, or at least for the Twitter and Mastodon section, but if I did have my own page, I would love nothing more than to have notifications of when you or Jerry update your related pages for me to go see, oh, there is an area within the Agora I'm interested in. I want updates or notifications from changes. So when friends of the link page changes, I know something has happened, I should go look at it. Is any of that infrastructure, does it exist? Or is it even an opt-in like, or a pseudo opt-in like, you know, web mentions, let's say, as an example? Thank you. Yeah. I mean, that's next. Yeah. First, you're sort of, exactly what exists now, pretty much. So, yeah, right now, the hour is very much like a fire and forget to some extent. And it tries, it doesn't notify lack. We do want to add that as an option. There's like some limited support in the sense that you have RSS feeds, right? So you can subscribe to, but it's a bit manual. And there's also like a matrix channels, it's sort of built on the RSS feeds and they can be used to like watch for updates, essentially. But I think the idea of flow will probably be something like watch node, right? And then you get notifications from anything you watch. We want to build that, but it doesn't exist. Okay. Well, at least RSS is a start. I can, I can live with that. Yeah. If I remember correctly, Julia, just RSS, actually, no, yeah, you need to add as for the feed, like I can feed, so please outreach. In the toward, yeah, there you go. Feed, please outreach gives you like a full feed for, for the node, this applies to every node, except for no all, which is like a hack and just gives you a whole hour. And it's very costly. That's interesting. So each node has its own matrix channel? No, not right now. No, no, each node has its own RSS feed. And you can, we have a, what? Interesting. We have, sorry, we have a, like a few channels that watch different RSS feeds, but right now it's only RSS. Now, you did hit on something which is like, you know, each node does have what we call stores, which are just like spaces associated with the node. So for example, like, you know, we have a Hedge document, etherpad and GCMeet, I don't know if GCMeet works. Yeah. Okay. You can have like a video call in there. Anyway, all very easy. It's just high frames. But we wanted to add precisely like a one matrix room per node, because it will be ideal, because the idea is like, you know, explore how cheap it could be to have like one space, one unified space. So that you know, like the two things you need to collaborate with someone is first, are you on an algorithm? And then you collaborate on things to make them by default in the namespace. And second, any node in an hour, you can say like photo, we will work on FOTL. And just with that, just with that bit of information, you have everything you need to like, you know, chat and so on. Now, out of the chat, we do want to add it. It doesn't exist. But there's a thing called cactus comments, like annotation with my matrix and another one which is similar to that one, which will be make it easy to create rooms per node. Yeah. So all these things we prioritize are going to interest. So if you're interested in, for example, like in these, we will prioritize. There you are again. Yes. Aram, how does this fit your conception of interoperating and the stuff that you're working on? Yeah, I mean, I think I always have thought that, you know, Anagora is very cool. I think it's maybe a little overwhelming for what I want to do. But we've had this conversation before, which is, you know, the whole point of this is we should each be able to use whichever tool we prefer. And I think that in terms of an RSS feed, that's definitely a good way to start in terms of picking things up. Is there an RSS feed for new nodes as well? Right. That would be the feed for latest. I mean, okay, so there is a feed for this page. Yes. Yeah, I think so. Yeah. Okay. Yeah. And this is what we watch usually on matrix is also show up here. Oh, nice. Oh, that worked. I certainly didn't expect images to work, but it did. Nice. Yeah. I'm going to try to draw that thing in Excala draw and embed the drawing instead, but for the moment, it's just a scan. Yes. And like, when it comes to like, as usual, like, you know, like, also, like, time is like, you know, drifting, like, I mean, drifting, but like, if we set up like the, like, for the experiment, I'm happy to remove things. They are going to sort of like a kitchen sink approach right now. The interface is a bit lower and so on. I'm happy to simplify and just deploy like a different operation. Cool. Now, if we, if we, if we, if we're Agra users, we should be able to go in and annotate those pages, but there will be no connectivity back to the original pages. In what cases and for what use, what use cases you have for us to actually step into the Agra and edit or create there? Or is the Agra just kind of the meta web that connects our common interests? Right. So, so right now it's more like on the meta layer and it's not, I mean, we haven't been investing a lot in the flows that are actually being back. So the idea there will be, you know, embrace indie web standards for that. But essentially, the idea will be to focus on like enabling serendipity to some extent. So essentially maybe, you know, like take these six repositories, for example, and have like, you know, some basic tools here that allow us to like find the overlaps. And maybe they maybe have the Agra just say, you know, if, you know, you, we have these resources sort of in common and same space in this node that we share, like can we just like say, click through to like, you know, add them all to like a shared document and then work on that. So essentially like maybe depending on the weather ecosystem for like the actual more collaboration in one resource. How do you handle changes to the original pages? Like if I make an edit to one of the pages you scanned here, is there any way you're going to scan it again to check or? Yeah, if you make a change, it will affect the Agra within 50 seconds. Oh, wow. Yeah, I mean, as long as nothing breaks. Without you're doing anything or the next time you visit that page in the Agra? No, no, no. Yes, it's automatic. The Agra is always pulling in the background. So wait, you're pulling for changes on every page you've added to the Agra? Yes. Yeah, yeah, because it's essentially every repository. Git is quite efficient because it's just a Git pool. That's one of the advantages of the Git approach. Of course, it won't scale to like a million repositories right away, not with that latency, but I think it can go quite far, particularly for like a set of rules that are ready to collaborate. Just because like you can apply a heuristics for like, you know, how often repositories are updated and so on. Thank you. No, no, thank you. So other thoughts on this or other riffs on what it means to share notes? And a meta aside question, should we leave Jitsi so that Chris has an easier time joining us properly and go to either Zoom or some other kind of tool? Because Chris, every time you have to fuss with getting in here, I want to take you that annoyance. Well, I will mention too that for some reason Zoom has weird quirks too, and I'm not sure, although it takes less usually within like a minute or two, it sorts itself out. So just for some reason, like spends five minutes, 10 minutes, 15 minutes trying to find my microphone in my camera every time regardless of if I'm using the local one or other attached ones or... What OS are you on? Lately, it's been mostly my Windows machine. So strange. So is there is there any video conferencing system that is clean for you where you don't have this problem? Don't worry about it. Don't special case me. It's some weird quirk that eventually I'll flush out. Today, actually, I rebooted everything and cleared all the cookies in the whole nine yards. Oh, wow. In hopes that that would like get something in. What's even weirder is I can't get it to work at all in Firefox. It's very strange. But even Zoom and Meet have had issues and I'm not sure I don't think it's a hardware thing because I've tried three different microphones and two different cameras. So I'm not sure exactly what it is. That's very strange. And including the built-in webcam, bezel cam? Yep. Yeah. Crap. Yeah. All of the above. And I've even tried other computers doing it. Yeah. I don't know. Maybe you have a zone of video incompatibility about you somehow. Yeah. I mean, I'm almost to the point where I'm going to switch out my my router and further hardware upstream to see what the issue is. So I don't know. So weird. Anyway, sorry to distract, but back to the question at hand. Do we have any conventions beyond standard markdown we want to preserve here? Great question. Well, apparently, Flancin has thought of that. So go. Yes. I mean, this one, yeah, about Stormy, but I'll take a minute. So this is what we're going to link again. Yes. So let me find this. So essentially, okay, the way I, yeah, I, I, I, I focus this, this topic from, but I call this like a protocol level. So essentially, our protocol is a set of conventions beyond precisely, you know, the baseline that in this case we're discussing will be, that will be, you know, like my run on it. So to some extent, I, the base protocol is like whatever a community I reach as a set of conventions for collaboration or the lower, lower friction to collaboration. So essentially, you know, it's a bit like cheating because, you know, if you tell me, you know, oh, this is what I would like to do in a group to collaborate, I will like, I will like, yeah, that's another protocol. It's just my definition. And like in, in, in practice, you know, so I think the way I see it, it starts with like the intent, you know, if you have an intent to collaborate, that's probably like what you will learn. And then more like more concretely, well, we killings, they are an extension to Markdown. Of course, like they are like a really common extension. So maybe we consider them part of Markdown, maybe not. But actually, because we killing sort of like require defining how you are going to resolve them. Because they are not like Angkor in Markdown. I think they are this layer of like what I call our protocol. Same for hashtags. And then I have this specific like a proposal for essentially this will be an extensible protocol, basically an extensible protocol where we can say when I write these, I mean this essentially for a variety of things. So I propose like in this text, like a few conventions that on top of this, which are like go, pull, and push. Which I think we will discuss like shortly in other cases, but surely and go is a way of saying in the context I'm writing in, which in the other cases, the title of what you're writing. So for example, like a link in shared nodes across six tools. If you write any external link a year within that document and you hashtag go it, you're saying if you take any link here to be canonical for this, it's this one. It's like exactly. Yes. And thank you for asking that because the answer is a goalie. So go out a chapter in the Agua de la Francia will, this is broken, but also will direct to the document I'm presenting because this document has a goal hashtag. So essentially it's like a link redirect, sort of like a short link, but it lets you, it's an example, right? In the sense that the Agua doesn't require you to do anything except write hashtag goal and it will parse intent. So it's sort of an example of how just by agreeing on a convention and writing, you can sort of use like the Agua as a command line and no matter where you write. Right. You can write in an OCDN and the Agua will still do this. Once your repository is there. And just shortly, before we move on. So is the go intended to indicate the main link of the page so there should only be one go per page? Or is it saying these links are all important to the definition of this page? Right. So from the two, I will choose the second because I'm not into like, you know, one way or anything or like one voice. And in practice also because if you want to do one goal link and I want to do another, we need to decide because the Agua only tries to think about nodes and the node will include bolder nodes. We need to decide which one gets redirected to and to some extent that's the layer of governance and ranking. So the Agua is trying to be like a generic framework for deciding that but right now it doesn't say this is the one. Well, it does but we can change how it does. And just thinking along the same lines. So what FedWiki tries to do with those little icons at the bottom that don't make sense to me is say, hey, here's a bunch of versions of that page. It would be interesting to have a graph view of the history of the page like here's who edited and what changes they made. So that would be pretty cool. You know, here is a fork. If it were simple enough to visualize, hey, this person forked the page, made these changes. And though the fork is kind of permanent and seems to have its own life, then you could kind of follow that around. And then fork and pull is an attempt to rationalize this so that there winds up being one main line or main reliable page, canonical page, for each page you're expecting, I guess. And the Agua and no. I mean fork and pull, I'm talking about GitHub specifically. Yes, yes. I think, well, I mean fork is, fork and pull are forking and then creating a pull request. So like I'm trying to merge, they are at the repository level. So they can apply to many pages, but yes. And when you're saying pull here in the Agua context, how different does that from? Yeah, it's different. So for example, this is transclusion just as per the Nelson, essentially. So here's the example, yes. So I created a shared node across these tools as a node. So now we have one node by Relate and one node by Atlantian. And this was incorporated within, yes. Hey, Chris. But I'm glad you're here with pixels, I know, eventually. So pull is just a request to embed. It could be called embed, it could be called like a transclude, include. But I call it pull because I went with pull and push just because I wanted to, I guess, simplify a bit. I don't know if I did or not. I don't know, an experiment. Also for disclosure, I mean, I was inspired by like the, you know, the expanse. No, sure. You were trying to build an Epstein drive? A what? An Epstein drive. Right. I mean, this is for the Internet. But like, yeah, you know, this interface, they have the floating whole around graphs and they're like grabbing things and like throwing them. So essentially, I was, what, this doesn't make sense in a minute, I think. So when you pull something, the other will say, after showing whatever is in the node, I will show that as well. So because, you know, because I said pull, fillet you the link. The other will include fillet you the link in the same context. So if you think about the graph, where you have like, you know, share nodes across these tools and fillet you the link, this one is pulling this one. So it will essentially be closer to, it's trying to reduce the distance. So these were like, I went with pull and push as sort of like physical metaphor. The only way that, I mean, the way that it compuses me a little bit is in contrast to fork and pole, which is an entirely different pole. I didn't think of that. So maybe it can be changed, we can have scenarios. Well, partly I'm thinking of like collaboration protocols, right? And when I'm describing that to the new leads, I'm like, well, there's one way you can do it is, you know, live interaction, and then you need things like CRDT or whatever else. Another way is to fork and pull, and that's how this works. And so you're describing a different operation here. Yes, yes. So I think it's an interesting, I mean, it's an interesting point. I guess I would think of fork and merge. By pole here, do you mean sort of follow? It's more like pull the resources there into this bucket, essentially. Yeah, or include, yeah. It's include, it's include. It can include statement. Yeah, this is literally an iframe. I'll mention that when approaching the agorah for the first time, coming up and understanding those definitions of push and pull, or even go pages, if you've never seen go pages before, those technical pieces are one of the biggest hurdles to understanding what's going on. What is this tool doing? Yeah, you know, even words like, transclude, unless you're maybe in a group like this, or at a certain technical level, a lot of people won't understand what that means, or what those technical words mean, unless you define them specifically on page one, which may not be a bad way to go. Quick question. You said a go page. Do these also create, the go directive creates a page as well? We just made a page with a bunch of go links on it. I think Chris meant the latter, probably. Yeah, no, go page is probably the wrong word. Go link is more standard, I think. Yeah, go links are used. But essentially it acts as a preview. Yeah, it's a lot of jargon. I got them, you know, it is a bit obscure. So sometimes this is like a variable, okay, this is the cop out, I guess. But like, I call it like a verbose bootstrap. You know, like, for example, I don't know if it comes across quickly. My strength is not like visual design or programming. You know, it's put together. But yeah, I completely agree. I'm pulling push and maybe a bit more idiosyncratic than they could have been. So I sometimes look back and think maybe. Well, and the other issue too is knowing how reliant parts of the Addera are on things like GitHub, which has a very clear cut push and pull kind of idea there, which doesn't completely overlap, makes the semantic part even that much harder. Yeah, that's completely true. Yeah, so maybe that's something that could change. I mean, I think the other synonyms for a few things, Go has synonyms, so it doesn't extend which words to use for transclude and for what push does. That can be part of the hour definition, essentially. I see it. It's part of the protocol. Yeah. But yes, thank you for the feedback because that feels a lot. And push is like publishing at a distance. So if you if it is sort of like it's not, you know, this is like it can be refined, but essentially this is this is more like this is probably more like when you do, we push it in a sense that you are publishing somewhere. So, so for example, if you wanted to trans-transclude something in share notes across six tools, you could say push share notes across six tools and whatever you pass will just show up there remotely. Will show up where? We'll show up in the in the Agora, for example, like we have an example here. Yes. So for example, you know, if I here I wrote, I do this a lot for writing in the journal. So no matter where you are in your system, you can say I'm writing something in my journal, for example, or like in any place. And you are like, oh, this is also relevant in these other places. So, you know, usually I will be like, oh, I write it here and then I need to go and like trans-clude or like use block references in Chrome or something. And it will push you just say like push there. And you just forget about it. Essentially, it's like you wrote in more than one place at once. Which would not which might also be achieved with a hashtag. It can be done. Yes, you can say hashtag push. But I mean hashtags are a way of connecting things across different spaces too. Well, right, you are by saying hashtag, you are being linked. Right. You are being linked in the hashtag. Right. Yeah. Okay. Right. I have all that. Yes. It's like registering what you're writing at the distance. Yes. It's like a strong link. Right. In a sense, push are similar. So does push? Push the current node and the directive node to the target or it pushes the current node to the target. So push A written in note B will include note B as a note on A. If you push, let me actually see if I have a better example. If you push, say you're writing in note A and you push to note B, what you put under, there's two ways. What you put under the push, like for example, in this case, oh, sorry, I'm going to show you this. Here you have push for Alexander in this node, which is actually 22 or 324. If you say push for Alexander, anything that is indented under, so like a children, like an outline, that is what gets pushed. If there's nothing. And is that now visible from the node Christopher Alexander? Right. So this is the node Christopher Alexander. Okay. So what did we just push there? How many seeking computations? So they're pushing, which means they show up below as attached as an appendage. Exactly. It's sort of like a publishing. It's like a feed. Right. So it's not changing the original page. It's just appending this node someplace else. Right. Right. The idea being that if you want to change it, this is to some extent because, you know, I was a bit frustrated personally with blocked references. Right. As you know, using Rome and OCDN now, I guess, OCDN does something different, like anchors, this parameter anchors, because the issue, of course, if you have your references and we have another, another garland elsewhere, we will need to synchronize the references using like a common like scheme, like a database actually, a scheme is not enough. But if we, so this is a way of doing transclusion without having to depend on a central registry, essentially. Right. So the limitation is that by default, you won't change the target. Right. It will appear as a separate resource. Now, one thing that is not implemented by the plan is if you are in Christopher Alexander, say, imagine like, you know, here, all the pushes are on my mind. You know, I imagine you push to Christopher Alexander and I, and it shows up as a separate sub node, like a blog, but it's relevant to a shared resource. I could be, go to a resource and actually pull or catch, I don't know, and say whatever is coming from there inserted here. So essentially like a way of, it's like getting a pull quote. Right. So, you know, those things we could play around for this. I mean, like all of these things under our protocol, they are sort of like experimental. Like I don't know, I find them useful. I use goblins a lot. Also here in the corporation, I use pull and push a lot. But essentially, I just dream this up. I mean, of course, transclusion is like a typical, but, you know, if we agree on any other ways of saying, like, you know, when I write this action, this is what I would love to happen. You know, our protocol just gives like a framework so we can make this happen, essentially. Explaining it as a block reference without the central registry is probably the most helpful definition I have heard because it takes something I know what it's doing. And in my own local system, it generates some random hash to give me that block reference. But then when I start playing with 30 other people, you can't guarantee it's not going to collide. And so that makes sense to me or it makes much more sense in the shared space. Thank you. Thank you. That means a lot. So use that definition as a means of explaining it, or at least be helpful for a subset of people. Yeah, thanks so much. Yeah, I think I added, I'm a very number of these teaching between competitive and complementary cognitive artifacts. It's the kind of thing where the definition itself is a cognitive artifact to me. It seems it's like a meta artifact because once I learn about the future, I start to see it everywhere. And I guess I like that these are readable. It's like push to for Alexander. Well, yeah, you can take a repository with such a push and say, like, I'm going to have a node for Alexander. And I know it's going to interact with this particular way. I don't know. Let's see. Shoot, my my link to the complementary and competitive broke. Oh, so I had a doc that with that name. But I am interface. Right. It is now broken. So yeah, I changed the reference because I heard this about this in a podcast with some Haris with the power. Yeah, I think I just found a new one. Joseph. Yeah, it took me a while to find a new reference, but I find it quite useful. Yeah. I'm Chris. I'm really interested in what note sharing means from your perspective with your systems also. And then and then sort of full circle after either of you, I'd like to then see what that does in the agora because I think that'll clear up in my head a bit what the agora is doing to weave together sort of these different nuggets of content in different ways. My side. Well, on my side, probably the easiest is to have a probably a subsection of a markdown text notebook in whatever form I'm going to use that in because I have several. And then so I have a couple different repositories. Most of my digital notes originate usually from hypothesis and then flow into either a markdown setup that's a private GitHub repo that syncs with obsidian. And then I have a separate flow that goes into my personal website that publishes it privately and only I can see it and search it. But I can always flip a switch and turn it on and make it public. So it's readable and shareable. And I have infrastructure there that does, you know, it's all HTML obviously, but there's also infrastructure for sending and receiving web mentions to communicate with other sites that support that as well as syndicating out to social media with a link back to, you know, a permanent resource. And I haven't set some of that stuff up on the obsidian like or GitHub repo setup. Although I've done that in the past with Tiddly with you, but I generally have not done that. So I'm trying and what I'm trying to not do is be invested in 15 tools and have everything separate so that I have to maintain multiple kind of databases for it, which is just insane. I think for this probably the best thing to do is to hook it into that GitHub repository which is ever so slowly becoming. I'm probably going to switch it actually and just make one that's a public that can draw from things so that I can have a separation between the public and private notes that I've had intermingled for so long. And you've been trying to do that in obsidian with nested vaults or something like that? Well, yeah. And right now it's split into two things. So there are nested vaults and eventually I can just flip a switch and say, here you go, it's part of the aggra now and have it make some sense right now. It's just such a hodgepodge of crazy that it's doesn't make sense. But I also have to just switch some of the workflows on the back end so that new things flow into the right spots. So let me see if I'm understanding some of this. So it would be trivial for you to look at the page I shared in the relate wiki and use hypothesis to make a comment or annotate the page. And I didn't realize how much you're generating original content in hypothesis over there and then sort of sinking back into your own infrastructure. That's really interesting. I hadn't sort of pictured or imagined it that way at all. But is there a link for you from a comment on hypothesis to actually just trying to edit and do like a fork and pull to a page that's available for fork and pull kind of cooperation? Not really. I could probably do that with their API. But I'm not sure how worthwhile that would be in the long run to do and support that type of infrastructure that's not my native code. Hypothesis into it, that's a whole other ball of wax. Or if someone wanted to dovetail with either their API or pull things in even through RSS, you could do that as well. And I've tried in the past to try and get hypothesis. And one of the other issues too is chunks of hypothesis is infrastructure even with relation to the web. A lot of their stuff is JavaScript didn't read territory and not raw, readable, parsable HTML, which is a similar problem to some of the things I was having with TiddlyWiki as a JavaScript resource. There are useful things that happen, but so much of it is JavaScript that isn't parsable HTML. It's just annoying. And as a result, it doesn't play nice with all the other HTML things. But it's reasonably easy for me to have a resource that sits on GitHub and is interactable, I think, is raw HTML with other things that can consume it that way or JSON or whatever the other thing is. Cool. And do you have this diagram dot any place? Have you written this out to explain to anyone else? Not fully and not in a long time. Although a lot of it's now stable enough and streamlined enough that there's only three moving pieces anymore or maybe four. And any time I get a chance to have fewer moving pieces, the better. So maybe I should draw it out at this point. It's interesting. All these things are hard to picture unless you know what the code is doing in different ways. It's a little bit like understanding how the web works in order to help somebody troubleshoot what might be broken because they can't see a web page. Okay, you're not getting this, which means the problem lies over here in the stream, not over here. And for what we're talking about here from Agora to your own system, to my own sort of imaginations of how these things work, they don't all fit neatly and elegantly. Well, and the other thing I tend to do for the last five or six years is I try to invest in things that have the easiest to use user interface and that are kind of broadly standardized. So if you're using something like some simple like markdown or HTML, I can buy into that a whole lot easier. Particularly if your markdown flavor is super plain vanilla and doesn't cause issues because there's 5,000 flavors of it. And there are fewer flavors of HTML which generally tend to be more backwards compatible. But things like hypothesis, to have a one button I can click and start writing and it saves it somewhere that I can do something with it. That kind of UI to me is stupidly easy and it's easy to see. It's got permalinks and I can point you at things that immediately take you to that page and show you my notes. But then also give me access to that data in a centralized repository that I can play with and change and do things with. So does Google Docs fail because of its document format? Because otherwise it's a way considering all the work we're doing here to sort of scratch our backs by reaching over to the Rube Goldberg device across the room sort of. Google Docs has it all beat by absolute simplicity of multi-party editing and so on and so forth. It's just that it's in a proprietary format on a proprietary platform. Which is also kind of a no, I'm not a big fan of that somebody else owning my data. So even in the last a month ago for reasons that are still beyond me and I can't figure out why or what Instagram turned my account off and locked me out and said you just can't play here anymore. Fortunately I've got all of the data I ever gave to them backed up and on my own space. But suddenly every connection I ever had with anybody on Instagram is now totally lost. Five people have called me and like did you and they know I have an anti-silo bent. They're like, did you just turn your account off? They've called their email like where's your stuff? I'm like well it's all on my website so you can find it. But they're irked. You know they think I just disappeared from the planet because that was the only link I had with them. So weird. Someone's busy trying to do it. No, they just literally yeah they just they pulled the rug out from underneath me and turned it off and I'm just stuck. I'm trying to gain enough distance from it and they've done this to me twice and I've jumped to the hoops and regained the account. But this time they don't seem to be responding at all. And I've been trying to go through enough, have enough distance from it that I just see pure red. And I would be even more irked and upset if it were the case that all my data just suddenly disappeared with it. But at least I kept it and backed it up so I'm not as pissed as I would be otherwise. But you know it's like and I just I don't have the time to to deal with it right now. I mean sorry now I'm seeing red. It's like they call them friends. It's like you build these relationships and then they are like no you don't have those relations anymore. Some people it will be very hard to reconnect with. Yeah and I you know if I was doing something crazy or bizarre or using had I been using that account for spam or some other weird insane thing or you know and it's not even like they have an API that I would have been hitting and butting up against doing something nefarious like you know what the hell did you and they won't say why they cut me off what happened what they think I was doing like not you know you know and it's all automated the interactions I can have with them so I can say here's a photo of me with your custom number as a selfie that I've uploaded to prove I'm an actual human being which I've done multiple times and still they turned it off and just said no and I'm like okay I'll move on. So I those types of things drive me bonkers and for that reason I don't really I'm not super keen on buying into a Google platform although my guess is they're much more permissive in terms of what I could and should do with it. Yeah I understand that I think like you know aspiring at first like you know I do have like very limited access to docs people yeah in single group so I will I haven't done this as of late but I will pursue to see if we can any chance we will interrupt you know like adult open standard for annotation for example and so on and there's some work for annotations in Chrome I know that was presented in Diana 2021 but anyway I can't really pursue that but also yes maybe we can find something as simple as to use as docs with comments yeah I think maybe Etherpad has comments or this route you know like one of those forks maybe has like an extension for comments flows maybe that well even things like you know they hypothesis and a few other related groups went through a w3c process and created a standard for annotations yeah but to my knowledge hypothesis is the only extant implementation of it which is essentially standard by fiat I think also war rain uses validations no and you know mimics mimics war rain yeah that may be but I've only ever seen the one implementation talked about war as an open source version yeah yeah completely and I think we'll have like who are running an open source version from hypothesis but there's not a lot of continuing work on improving or changing or doing x y or z with it so it you know it's it's it's a standard and it's there if you want to use it but you know there's no evolution or you know active work going on generally for it I think we can it would be nice to pursue a forward for sure there's a lot of interest I think I know like you know I spoke about what we've done like more than a year ago but like you know they're very focused on academia and like I'm making a hypothesis work but I think he's very interested in general in percent that eventually I don't know and I he has all the contacts in like the w3c so maybe we could experiment with like some flows related to notations and report back yeah or even or even making what we're doing here and I don't think there's been any talk or movement for you know that most of the groups I've known that we're doing this kind of work have shut down so things like you know the social web working group I think was open for about five years and created a lot of the indie web building blocks and things like some of the standards for activity pub for mastodon but a lot of that work has now since they've closed the group and it ceased and they had they were successful in kind of trying multiple routes and avenues but I don't think you know it would be interesting to see a wc3 group doing this type of work either basic text and or html kind of coming out and people working on that sort of you know and maybe that's maybe that's us we're already doing it it would be easier so I had to step away that was my bank calling because I just set up a bank account for rel 8 and they were having trouble convincing their back office what rel 8 does so I had to kind of explain a little bit more oh I can imagine it's very funny we're gonna run out of time soon Aram I'm curious what this what sharing notes looks like from your end uh yeah hold on a sec I'm just adding the information here about the web annotations yeah so I think that there's a lot of there's a lot of interesting stuff here that I'm just trying to think my way through what I'm going to do here we go let's put useful specs on what I'm going to do what I need what I've been working my way towards and sort of recording information notes from here on um is right how to publish my notebook essentially um and then be able to make it interoperable with the rest of the notebooks right because I'm not coming into this with a particular system um I have a different set of like priorities um than of some of you which is like we talked about it's fine right my interest here is I don't want to have a server um I want to build a static I think that's important for like making things last longer um and I have a pre-existing notebook that doesn't have necessarily quite the depth of practice um that some of you has which allows me to be a little more flexible so I think essentially this means I now have sort of a requirement list and a set of directives that I need a parser to be able to understand um that allows me to create pages in a way similar to how other people are creating them as well as be able to read pages from your systems in a way that they will then operate similarly um so I just published it into it into a blank repo here um with the idea of uh sort of what my conventions are how I understood the conventions we've discussed here and uh how they can operate together uh the main thing I think that is going to be there is the sort of YAML stuff is stuff I already use because I heavily use YAML the folder structure is stuff I already use the meta statements I started with stuff I already used but also added sort of some adaptive things from what we've talked about and the conventions are pretty much all coming from Floncian's conventions um as stuff that I want to be able to support and then some notes about other things that I may want to link in um and so I think this gives me a set of tooling that I can understand and build out with um I have a bunch of pre-built tools for this um built when I was building contact center um yeah I just pulled that directly from your page the calendar one I was just looking at engora stuff and and pulling or sorry agora stuff and pulling it in um so that is um yeah I guess what so go calendar node what what is that going to do is that a URL structure right no right right so the agora just tries to when you pass like a serial like we think with slashes it will try to compose so if you say go calendar node it will see if it finds a calendar link and it will relate to that one in node okay um so like it'll go find the URL that's the calendar URL yeah and forward you to it yeah for any for any hashtag that you use okay uh let me make a note of that because that was not clear to me so yeah no I didn't mention it okay uh yeah this is like a composite goal so go has superpowers depending on what arguments you pass it I guess right right yes and it would generalize right so the idea being that you pass certain nodes it will try to like find one in the other or one of the other one they try to like promiscuously find something to do is um is the goal link so central to agora that the g o in agora is there a reference to go ah good question no not really yeah yeah damn i thought i was onto something there but you know what we might uh but uh back uh retcon right yeah yeah so like this is good because what it means is that one I now have rules for essentially building tooling that pulls the notes that are marked as public from my private notebook into a public repo that's one and two is I now have tooling with clear rules that I can establish I use markdown dash it as my markdown parser that's a node markdown parser um and I can build these rules in right like now I know if there's a particular hash like go I can start saying if it starts with go then between hash go and the end of the line there's certain rules I can build out um and if you're curious I have information here let's see um I'll put this into the link I did a retro on my use of markdown it for my dev blog explaining how it works and what I'm using it for and why I ended up deciding after sort of struggling with it that I do want to keep using it um so now I have rules that I can turn into plugins and I could build those plugins into markdown it and then I can publish those plugins and use them across multiple projects and other people could theoretically use them um and that allows me to now establish here are my rules here's the program or the functionality that derives from those rules um and here's how I could build them so I have I have a product description and I have the next steps and I now have some idea what I want to build for my notebook publishing tool um so this is useful to me in that sense and I think the the sort of last question that is I don't think we necessarily have answered is sort of where um how we create the matched links I think the gore has a clear way to uh to do it which is um right it's just you're gonna list sources and then pull those source matched nodes in from those sources okay so I think um that can go into my interlink section and we can say um establish list of sources and uh where those sources have url slash matched path um path pull that source in as a link or potentially a node for the matched path on my own site um I guess that leads me to one last question then does this therefore assume we have a flat structure right because if we're matching based on paths then we either need to have a flat structure so we all know that url slash path is where we can find what we're looking for or we need to have some way to say here's the set of folders that you can hop to and check for matched paths there right yeah I think I personally don't know what the right answer I think it may depend on the work group it may be part of the protocol whether because there's it's a trade off no what the algorithm does do is like instead of uh you so you have like a full as a nodding and you can also have like a real eight full for example so a user like a user like so like a syntax where that means full only according to real eight which is not the same as a path because it's not really a path it's more like a selector right so it doesn't mean but it's actually maps to resources under the path you know matching where that source actually is deployed and is it your convention that it's always at really uh yes that's a convention we use essentially it's like just like out as a user it just means the repository so if you check out the sources yaml yeah anything that is under garden is a user name so forgive the programming ignorance this is a little bit like content addressability in the sense of if we could ignore the URLs of the permalinks then we would be finding stuff because we know how to mention it and then I was I was getting the impression that go links were meant to be a kind of inoculation from URLs in the sense we're talking about but I don't understand enough what it's doing under the hood yeah this is the intention and then I'm trying to do the simplest thing that could possibly work in this little demo of sharing notes so I'm like if we just use a permalink as the place to meet then awesome and the question is to what extent can we just read read and edit read and pull request or in some other way collaborate through this node