#+TITLE: Adsan's junkyard #+NICK: adsan #+DESCRIPTION: I post whatever I want #+LINK: https://adsan.dev #+CONTACT: mailto:me@adsan.dev #+CONTACT: https://matrix.to/#/@adsan:adsan.dev #+FOLLOW: rossabaker https://rossabaker.com/social.org #+FOLLOW: omidmash https://omidmash.de/social.org #+FOLLOW: eoin https://eoin.site/social.org #+FOLLOW: emillo https://emillo.net/social.org #+FOLLOW: nickanderson https://cmdln.org/social.org #+FOLLOW: andros https://andros.dev/static/social.org #+FOLLOW: sachac https://sachachua.com/social.org #+FOLLOW: shom https://shom.dev/social.org #+FOLLOW: leandro https://cybervalley.org/org-social-leandro/org-social.org #+FOLLOW: alephoto85 https://www.alessandroliguori.it/social.org #+FOLLOW: concavi https://concavi.net/social.org #+FOLLOW: cherry https://cherryramatis.xyz/social.org #+FOLLOW: comacero https://comacero.eu/social.org #+FOLLOW: razzlom http://gemini.quietplace.xyz/~razzlom/social.org #+FOLLOW: arnebab https://www.draketo.de/social.org * Posts ** :PROPERTIES: :ID: 2025-08-16T16:51:54+0200 :LANG: en :TAGS: introduction :END: In the spirit of decentralization and returning to the internet's roots, a fully self hosted, informal social network seems like a great idea. So I think I'll put some thoughts here, whenever I feel like it. ** :PROPERTIES: :ID: 2025-08-18T13:36:19+0200 :LANG: en :CLIENT: org-social.el :REPLY_TO: https://andros.dev/static/social.org#2025-08-18T10:49:34+0200 :POLL_OPTION: Yes, I want to have groups (WhatsApp/Telegram/IRC style). :END: Groups seem like a great way to actually make people get together, rather than either posting into the void, or chatting back and forth. It's also better to split up the feed, for when there are more users on the network. ** :PROPERTIES: :ID: 2025-08-19T00:20:44+0200 :LANG: en :TAGS: org-social :CLIENT: org-social.el :END: The idea of discussing the development of the network on the network seems very healthy. It's somewhat like bootstrapping, at this point even with no outside tools (like github itself, with which I have problems connecting rn, it works only when I vpn out of my country??), three is enough of a link for the normal "git/community style" development to continue. Not saying that everything should be done here, accesible documentation is key. ** :PROPERTIES: :ID: 2025-08-19T00:40:38+0200 :LANG: en :TAGS: org-social.el :CLIENT: org-social.el :REPLY_TO: https://emillo.net/social.org#2025-08-18T14:18:02+0200 :END: I think that apart from "flat" and "threaded" views, there should be a place for any mentions/replies you got (from witch you could jump to the context - in threaded view perhaps?), regardless of where they were. Those messages are probably the most important to the user, after all. ** :PROPERTIES: :ID: 2025-08-24T16:46:55+0200 :LANG: en :CLIENT: org-social.el :END: It's *too* quiet... Anyone here? Hope we are not dead yet, just all taking a break for a few days? ** :PROPERTIES: :ID: 2025-08-25T02:16:59+0200 :LANG: en :CLIENT: org-social.el :END: [[org-social:http://omidmash.de/social.org][@omidmash]] Do you have some sort of ip rate-limit set on your website? I've been fetching feeds in semi-quick succession, and it seems that yours was the first to stop responding. Also, you should probably configure your nginx's mime to serve .org files as plaintext, instead of as a file. Just a convenience for browser usage 😉. ** :PROPERTIES: :ID: 2025-08-27T04:55:02+02:00 :LANG: en :CLIENT: org-social-rs :END: Hello world For the better part of a week, I've been busy creating an org-social.el alternative I like emacs, but don't really enjoy elisp, so contributing to the main client seemed unapealing. Also, I wanted to integrate org-social into my dashboard, powered by [[https://github.com/wtfutil/wtf][wtfutil]]. As such, I've created a rust library for parsing org-social, a subset of org mode, and some post and profile management. And paired with that, a cli/tui client, that allows for *some* interaction with the network. The project is still very WIP, slapped together in a very short time, and lacks a lot of features that a text editor just provides. But still, I think it's worth publishing, if anyone ever needs to use it. [[https://github.com/AdsanTheGreat/org-social-rs][Client]] [[https://github.com/AdsanTheGreat/org-social-lib-rs][Library]] ** :PROPERTIES: :ID: 2025-08-31T15:00:10+02:00 :CLIENT: org-social-rs :REPLY_TO: https://shom.dev/social.org#2025-08-24T23:25:09-0500 :END: I indeed did not see that for a while... Creating a mechanism to hop from user to user to discover anyone that gets mentioned, replied to, followed, etc. is on my todo list. The different source groups seem like a great idea, because seeing replies and mentions from anywhere in the network is probably the most important feature for notification systems. Also, for tracking conversations, posts that people you follow replied to should get fetched, even though the rest of the author's feed is not needed for you. ** :PROPERTIES: :ID: 2025-09-01T00:33:48+02:00 :TAGS: org-social.el org-social-rs org-social.nvim :CLIENT: org-social-rs :REPLY_TO: https://cherryramatis.xyz/social.org#2025-08-31T19:13:23-0300 :END: Wow, it's great to see another client. Seeing three different approaches to the same problem, in three different languages and environments, is lovely. Monopolies suck, you need differences to see real evolution. ** :PROPERTIES: :ID: 2025-09-03T18:04:36+02:00 :TAGS: org-social.el :CLIENT: org-social-rs :REPLY_TO: https://andros.dev/static/social.org#2025-09-03T12:12:57+0200 :POLL_OPTION: Yes :END: I think for consistency, the chronological timeline would be better in org-mode format. That said, if you would be implementing more abstract display modes, like threaded view, it fits the format less, and could just be more streamlined Your post made me finally implement polls for the rust client, I hoped polls wouldn't come up in the timeline again yet... ** :PROPERTIES: :ID: 2025-09-06T14:54:35+02:00 :TAGS: films tv-series :CLIENT: org-social-rs :REPLY_TO: https://andros.dev/static/social.org#2025-09-06T09:38:12+0200 :END: I do love Gunn's writing, though I find Peacemaker to be very weird and kinda all over the place. I think while the humor and plot are somewhat decent, the characters and acting feel off to me. ** :PROPERTIES: :ID: 2025-09-06T15:02:29+02:00 :CLIENT: org-social-rs :REPLY_TO: https://omidmash.de/social.org#2025-09-06T10:07:40+0200 :END: Hell yeah, a mint based school! The fact that you are both the school's sysadmin and IT teacher (as I think it usually is in schools) has some benefits though, doesn't it? At least you don't have someone over you forcing some microsoft bullshit or other proprietary hell. You have the power to shape your own workplace to your liking & ideals. It's a responsibility, but also an opportunity. ** :PROPERTIES: :ID: 2025-09-07T15:45:33+02:00 :TAGS: #emacs :CLIENT: org-social-rs :REPLY_TO: https://shom.dev/social.org#2025-09-06T22:36:20-0500 :END: I pretty much always do only a main(stable)/dev branches. Dev gets a lot of small&big commits, I don't care about spam there. Main gets merges with full squash from dev. I keep only commit messages in the squash's description. Main still gets commits when I need quick bugfixes/changes and there is no active dev branch. I don't need much rebasing, since dev gets remade each time. I think maintaining multiple branches is kinda pointless when developing alone, with one contributor and no needed communication, its simpler this way. ** :PROPERTIES: :ID: 2025-09-07T16:36:30+02:00 :TAGS: org-social :CLIENT: org-social-rs :REPLY_TO: https://andros.dev/static/social.org#2025-09-07T16:29:39+0200 :END: That makes sense, It's the same as poll voting with no text. Great addition! ** :PROPERTIES: :ID: 2025-09-09T03:20:00+02:00 :TAGS: book reading asimov scifi :CLIENT: org-social-rs :REPLY_TO: https://www.alessandroliguori.it/social.org#2025-09-08T21:10:18+0200 :END: Personally, I prefer my hard scifi closer to present earth than asimov's books. I've read most of his foundation series (I think there is also a running show about it?), and I found it definetely interesting, but I don't like when scifi starts to get a little star-treky, in a sense where the tech and environment gets more into magic and wonders teritory (and where there are simply overpowered characters) Most recently, I've been enjoying Ben Bova's "Grand Tour" novel series (it's so long...). While it's often more about politics than anything else, I find it much more intreresting than normal political novels, and also more relatable than some more abstract works. ** :PROPERTIES: :ID: 2025-09-15T02:37:04+02:00 :TAGS: org-social org-social-previews :CLIENT: org-social-rs :REPLY_TO: https://andros.dev/static/social.org#2025-09-14T10:40:53+0200 :MOOD: 😎 :END: Previews are *the* way to integrate org-social with other forms of communication. Do you know what is also great for mentioning org-social in other social media? Embeds Dropped a simple embed implementation your way, it should work most places. Now we can go around, advertising our posts left and right! ** :PROPERTIES: :ID: 2025-09-15T02:46:18+02:00 :TAGS: graphic-novels :CLIENT: org-social-rs :REPLY_TO: https://andros.dev/static/social.org#2025-09-14T09:22:09+0200 :END: You know what? The League of Extraordinary Gentlemen actually seems interesting, might give it a try. ** :PROPERTIES: :ID: 2025-09-18T17:38:15+02:00 :LANG: en :TAGS: hn :CLIENT: org-social-rs :END: I have no idea what I just read... [[https://ipfs.io/ipfs/bafkreieiwashxhlv5epydts2apocoepdvjudzhpnrswqxcd3zm3i5gipyu][An Afternoon at the Recursive Café: Two Threads Interleaving]] ** :PROPERTIES: :ID: 2025-09-21T17:35:58+02:00 :CLIENT: org-social-rs :REPLY_TO: https://cherryramatis.xyz/social.org#2025-09-20T12:55:56-0300 :END: That is what was my thought when creating the rust library. Though, as it turns out, that most of the work needed is not just parsing org-social specifically, but also parsing org-mode itself too, at least for my purposes. Though I suppose creating a simple parser to just extract posts and metadata is fairly easy, most of it could probably be done with some clever regexes. ** :PROPERTIES: :ID: 2025-09-21T18:24:28+02:00 :LANG: en :TAGS: programming organization :CLIENT: org-social-rs :END: I find focusing on a single (programming) project at a time damaging to the quality and efficiency of the work. In my experience, whenever I'm creating something, it's way better to have exactly two projects to dynamically hop between. Those projects have to have very specific traits though: - They both should have at least one simmilarity in their topic, environment, toolset, etc. - but not too many. For example two completely different programs that share only the language is fine. Or working on a simmilar task in two different languages works too. - They should both have defined goals - be it roadmaps, needed features, issues, etc. - The work should be able to be done iteratively, where it's not "all or nothing" - but you can actively try to test it as a human while creating it. My theory is, that two projects are the best when optimizing for total efficiency - it's slower than doing only one thing, but overall faster than doing one thing to the end and then the second one afterwards. The idea behind such project pairing is to use simmilarities to your advantage. While working on an algorithm for one project, one can find a way to modify it and apply it to another. There is also the benefit of shared research - if you have to learn a specific part of a language or environment, it could be applied to a different project with the same language or environment. Fully focusing on one thing doesn't have such advantages. If you get stuck on a problem, you have to keep hammering at it until it's solved, you can't move on to a different thing while still considering the old problem in the back of your mind. There is also the risk of burnout, it gets boring doing the same thing for a long time. Even if it could be faster on that specific project, the total work done is less if splitting focus. It also doesn't work for more that two projects - the more topics there are to be considered, the harder it is to remember the specifics of each project while switching contexts. It's becoming more chaotic, with not really many benefits. The second project could in theory be replaced with some task fully unrelated to programming or computers - perhaps a sport, some hobbies, etc. Though that has the drawback of actually reducing the total time spent working on projects - so it brings a reduction in total work speed. My thoughts are probably applicable to other things than programming itself, though that's mainly what I had in mind. It's how I've been organizing my tasks for some time, and I find it work for me excelently. The whole idea also prioritizes efficiency over everything else, and is created with the focus on personal projects, where one has the choice of what they do, and there are no real deadlines - though I suppose one of the two projects could be "coming from the top", while the other is a complimentary one chosen as a kinda-escape. I've never really shared this theory with mine with anyone, so I'd love some feedback from anyone to refine it and compare experiences. ** :PROPERTIES: :ID: 2025-09-22T18:28:19+02:00 :TAGS: org-social org-social-relay :CLIENT: org-social-rs :REPLY_TO: https://andros.dev/static/social.org#2025-09-22T17:00:26+0200 :MOOD: 🎉 :END: ** :PROPERTIES: :ID: 2025-09-24T07:39:13+02:00 :LANG: en :TAGS: linux kerneldev :CLIENT: org-social-rs :END: A curious idea for running multiple kernels on one machine: [[https://lore.kernel.org/lkml/20250918222607.186488-1-xiyou.wangcong@gmail.com/][Multikernel proposal on lkml]] tl;dr: Starting another kernel(s) on a separate cpu core(s), sharing physical resources. Allows for things like secondary RTOS or full handoff for seamless kernel updates. Seems kinda crazy to do something like that, but who knows, it might be another step in the fight for more efficiency for servers. ** :PROPERTIES: :ID: 2025-09-30T05:18:34+02:00 :LANG: en :TAGS: html org-mode js :CLIENT: org-social-rs :END: Modern browsers are so liberal when it comes to how you can massacre html... I've noticed it's possible to embed html in the middle of a plaintext file, and when opened in a browser (when it's expecting html) it just... mostly works. So how about abusing it for org-social? With an addition of one, commented-out according to org, line of squished html, and external scripts, an actual browser-compatible display is born. I guess if it was standard, any browser with js support could be a client by itself? Not an actual proposal, just a fun hack I want to share (compatible with current parsers* as a replacement to normal feed): https://adsan.dev/social.org.html ** :PROPERTIES: :ID: 2025-10-01T02:37:36+02:00 :CLIENT: org-social-rs :REPLY_TO: https://shom.dev/social.org#2025-09-29T23:14:07-0500 :END: It's worse than just rendering everything, they are happy to run arbitrary js (and even wasm in my case)! You could possibly even mask such behaviour if you played your cards right: if the html was down low in the text, and the css styling was one that mimicked how plaintext is rendered in most browsers, you could have a site *appear* as plaintext - if the user is expecting it, such as is with org-social feeds, it wouldn't raise suspicion. You could hovewer do stuff in the background, like fingerprinting, with the user having no idea if they didn't manually inspect the site. I guess there is a point in disabling js in browsers... ** :PROPERTIES: :ID: 2025-10-12T20:24:12+02:00 :TAGS: org-social :CLIENT: org-social-rs :REPLY_TO: https://andros.dev/static/social.org#2025-10-12T11:30:47+0200 :END: On the main network that is. There could in theory be a sefl-contained, niche sub-network of users that don't interact with us and don't advertise their feeds on main repositories. Beuties of decentralization! ** :PROPERTIES: :ID: 2025-11-01T13:36:58+01:00 :CLIENT: org-social-rs :REPLY_TO: https://omidmash.de/social.org#2025-11-01T13:32:26+0100 :END: #+BEGIN_QUOTE 6 museums in 5 days #+END_QUOTE Makes it sound like you robbed them or something... Though the louvre thing happened before, so it couldn't have been you. ** :PROPERTIES: :ID: 2025-11-01T13:41:55+01:00 :TAGS: org-social org-social-el :CLIENT: org-social-rs :REPLY_TO: https://andros.dev/static/social.org#2025-11-01T11:12:51+0100 :POLL_OPTION: Another one. Leave your suggestion! :END: I think the relay still needs some expansion. Of course the main problem now, it that there is only one (should I deploy one for myself?), because there is not that much incentive to do so. There could still be some other features though, dropped some issues your way if you want to give them a thought.