Projects general

What are you working on? What projects have you been putting off? What do you enjoy doing technology-wise? Have any ideas on how to fix or upgrade software that others might share and collaborate onto? I am creating a 4x4 macro keyboard with mechanical switches, the pcb for it at least, and have been in the process of cracking old captcha.... also have a desire to remake my local network to incorporate some openwrt repeater and possible VPN tunnel on that openwrt router. Please contribute :)

OP here Anyone doing anything cool or have tips for me?

>What are you working on? I'm finishing the final piece before I start work on implementing Pest: http://pestnet.io I'm also planning to begin work on the pieces of a simplified but compatible BitTorrent client. >What projects have you been putting off? I want to model the Latin language with a representation for human text I believe to be better than the character-by-character model. I'm still learning Latin, however, which has prevented me from beginning work beyond thinking about it; I don't want to end up in the situation where my model can't cope with certain aspects of the language unknown to me now; I'll soon reach the point where I can begin this, since I'm nearly done learning the language, sans expanding my vocabulary. >What do you enjoy doing technology-wise? I want to write correct software unlike other software; I want to create novel software. I want to program without the so-called text editors, and I want to write human language without them as well. I've already succeeded at an earlier goal related to this, but my work carries on. >Have any ideas on how to fix or upgrade software that others might share and collaborate onto? I work alone, but always enjoy discussing my work with others. Still, I don't want to overwhelm this discussion, so I'll link to a thread I made in /lambda/ if anyone's interested, or I could discuss it in some detail here. Most good work happens alone. >I am creating a 4x4 macro keyboard with mechanical switches I recommend buying a chording keyboard for an example of prior work; I can recommend a model if that would help.

I got a PC-98 recently, so I'm trying to service it with period-accurate parts. Fixed disk setup and installation has been a pain, but it's fun. Hardware is always the most interesting to me. (Mechanical) engineering seems so daunting when faced with creating novel solutions. I wish I had more free time.

Not much of a project, but I' want to build a linux system as minimal as possible, just enough to do some networking and run a lisp system. I decided I don't like emacs so I'll even ditch that for some flavour of vi*. I'm not sure whether to do LFS or some lightweight preexisting distro like either crux, smgl, or slack. These are my primary options after years of trying just about everything jnder the sun. My specific requirements are: a C compiler, Perl, and SBCL, plus vis, neovim or even nvi (the one with the least requirements). And perhaps irssi. A few other programs might be added later, but always centered around this simple workflow. For a shell zsh would be comfortable if it doesn't require a lot of external libraries or soykaf like dbus.

I'm writing a follow-up to a post on Urbit I wrote about a year ago: https://mayvaneday.org/blog/2022/august/urbit.html Some of my initial complaints have been fixed, but I'm still having a hard time seeing any actual value in the Urbit system. And it doesn't help that half of the apps on Urbit are broken or inaccessible. I got my Urbit key from a lainbro I was friends with five years or so and have since lost contact with. I wonder what he's up to these days. I'm also trying to mass-archive Twitter tweets in case Elongated Muskrat decides to actually put Twitter behind a paywall for real instead of just pontificating about it and never actually following through. And also going through the apps I use on F-Droid to see which ones I can compile directly on my phone via AndroidIDE. You know, for apocalypse purposes.

>>73083 Good guy non-clicky switches. >>73095 Hey, are you the anon who posted about Pest in the messenger thread and claimed to be working on an implementation? I've been meaning to bump that it to ask how your efforts were going. Especially since I've started dabbling with something adjacent. I myself have been playing around with ideas I had and still working on for p2p mechanisms that really I believe are close to a perfect variation where there really only a single "anchor node" is needed to have the entire system assemble itself autonomously and been trying to model it out in js (yes, I know but it's great for prototyping) and it's been important to me to base my design on a working model. Not sure how much Stanislav only dwelt in the theoretical of it, and I know there's already implementations. And while also eager to find out more about yours, it being focused on IRC and "station nodes", my interest resides more with approximating a feature set more akin to "modern" messenger clients and frankly I've been struggling to find solutions to a bunch of problems still, like how to efficiently get signed timestamps when there's no nodes online with sufficient credibility among other things. But I've got ideas I'll keep trying out. For now I've been rewriting the same components over and over to iron out the basic things and wrap my head around that domain. I remain optimistic however and if I end up coming up with a working system I'll try to put it online somewhere. What you see is a node failing to connect where each node tries to acquire exactly 2 other connections, which by itself is nonsensical but if everything works as intended it should form a single ribbon. The other is a stress test with 3-5 connections each if I recall correctly. >>73125 If it's one of these projects where messages are stored in a blockchain then, aside of being a terrible use of that technology, is going to be a dud sadly. Despite how much promise the whole web 3.0 thing had, thus far it has been a major waste of time (and money) aside from maybe a handful of projects. But I'm not one who desires to fuel the scoffing side of it who is always critical of crypto and all that and constantly decries the unavailability of graphics cards and what not. While the "blockchain bros" are overly euphoric over the theoretical possibilities while their attempts keep hitting the walls of scalability and critical mass, their opponents seemingly keep forgetting about incentives and while good will, desires and pure voluntarism can go a long way, it often just isn't enough. Not at a substantial level of adoption that is. A successful design needs to be mindful of both these "sides" or rather aspects. Another post in the aforementioned thread about messengers rightly stated that yes, running a node and providing a server has costs to it. But I wonder by how much that cost can be cut to reach a trivial amount, always.

>>73125 Hey Lain, I've spend the better part of the last hour going through the posts on your site and it's good stuff. I hope to have put together something as interesting as that in the future. Keep doing what you're doing.

>>73125 here again. I couldn't find the "scrape everything without consideration" thread in the catalog, so I'm posting here since this also counts as a personal project. I'm trying to reduce my time spent online, so I'm stocking up on Kiwix files. But Kiwix's catalog of parts of Wikipedia and Stack Overflow or whatever only go so far, so I'm writing my own Bash scripts to turn sites I like into the .zim files Kiwix likes. https://codeberg.org/lethe/ZimBuildScripts Hitting random Breezewiki instances seems to be a good way to get copies of fandom.com wikis offline, but recursive Wget doesn't pull images if they're hosted on a separate subdomain, so I haven't been able to get images for a few of the sites.

Been thinking of projects to make money from lately, but everything has a 10+ open source solutions, or even 10+ free tier hosted solutions that I wouldn't be able to compete with. Email read receipt service? There are 5 browser integrated extensions for gmail, hotmail, etc. that do everything you can think of related to email snooping. AI anime character detection and recognition for anime videos? Open sourced already. CRM? Of course that's already open sourced. It's like there's nothing else left that we need. Everything else invented is just cluttering the market with more gimmicks.

>>73946 Always like seeing someone on this board with a real goal to give real advice to. Lainon, if your goal is money, you're making the classic mistake of working the process of writing a great program exactly backwards. You shouldn't start with looking for a technology space you like to work in like you're going to submit a grant request to a singular university research board, and then somehow spin that into a product, you need to start by finding people with a problem they will pay you to fix, and then develop the technology that will best solve their problem. The project that became Google didn't start out of trying to "create something new" with linear algebra, it came from noticing how hard locating things on the internet was and deciding to apply whatever would fix it. Also, don't be too particular about how "innovative" your product has to be. A small amount of technological change can mean a huge productivity boost for your users; they're not necessarily correlated. As long as you've made something that can't be copied in days to weeks most larger organizations will be too ossified to try beating you until it's too late. And even if you are cloned, many businesses today beat out cloners for no better reason than that hearing they're getting a clone makes customers uneasy. It can be a little difficult to get into, but my advice is to find a semi-traditional industry you can get reasonable access to, and spend a while learning as much as you can about how things are done there. Know the business' goals, compare the products, learn the production process, and so on. Identify as many cost, time, and difficulty pain points as you can, and spend time thinking about all sorts of different tech that might reduce, circumvent, or solve enough of them to be worth paying more than what it will cost you to develop and run them. That can be your project that hasn't gotten ten working open source solutions yet. And since you're serving businesses, it'll be easier to sell to the ones desperate to get a market edge than to often fickle consumers just looking for entertainment and trends to hop on.

>>73946 Your product can give experience or features not existing in those open source or free products. UI/UX is a lot of issue with autistic open source people. Also, most people don't know about free and open source product, they just know the first click of the link that SEO provides, or marketing. Be faster, be better, and that usually beats the "be first" notion.

Open source concept of what a cool metaverse would look like based on O3DE with an NPC running as a finetuned LLM, a cool web-view so that one can interact with pages within O3DE and a decentralized chat system based on either Session, Veilid or Briar (hoping that Veilid is mature enough when I get to implement the messaging stuff). Hopefully one day we will be able to replace humans and real life. Will be on github with MIT license and some documentation, so that people can see what O3DE is truly capable of.

>>74030 >my advice is to find a semi-traditional industry you can... This is great advice, but I have no ideas on how to apply it. I used to be closer to the manufacturing industry with more opportunities to learn business demands, but now I'm in some lousy, but good paying, gov job, cut off from virtually everything My email read receipts prototype works with several email providers now, but it's probably gonna be a while until I make my $50 back I spent on this project

Are there any more lainanons who'd like to share some projects their working on? I'm hella curious since you usually impress me with your niche knowledge

>>73187 >Hey, are you the anon who posted about Pest in the messenger thread and claimed to be working on an implementation? Yes. Excuse me for taking so long to reply. >I've been meaning to bump that it to ask how your efforts were going. Especially since I've started dabbling with something adjacent. As of now, I only have part of the core loop finished. I've finished all libraries I need, however. >where there really only a single "anchor node" is needed to have the entire system assemble itself autonomously Theoretically, Pest can scan the entire IPv4 Internet to find nodes whose addresses have been lost. >And while also eager to find out more about yours, it being focused on IRC and "station nodes", my interest resides more with approximating a feature set more akin to "modern" messenger clients and frankly I've been struggling to find solutions to a bunch of problems still, like how to efficiently get signed timestamps when there's no nodes online with sufficient credibility among other things. Some of these problems may very well be intractable, which means one quality or another must be sacrificed. >But I've got ideas I'll keep trying out. Tell me how that's been going.

Working on a low-level interactively compiled language. Goal is to target roughly the same niche as various simple flavors of forth, being designed for REPL use, compiled, and (kinda) composable, but with some nice features out of the box like a (mildly) optimizing compiler. I realized today that it made a lot of sense to actually write a bytecode interpreter, and the bytecode interpreter should have instructions to output machine code. This actually substantially simplified the design and made it a lot easier to implement macros.

> What are you working on? An OS[0] although most focus have been on userland, which have been given use in a linux distro[1], and a programming language[2] [0]: https://eltaninos.org [1]: https://eltaninos.org/en/download/distros/glacies [2]: https://eltaninos.org/en/download/tools/vesper About the OS userland: * safe libc (tertium) with a POSIX compat layer (simia) * posix utils (ecore) * make alternative (redo) * reproducible build system (arbor) similar to ports but using redo * functional package manager (venus) - with archiver (venus-ar) - with crypto hash (venus-cksum [blake2]) - with configuration format (venus-conf) which works recursively About Glacies: * arcan as display server * durden as wm * s6+s6-rc as init system * sndio to deal with audio

>>73083 Working on an 2016 Discord Clone, because the current one is a mess and I like challenge If anyone cares https://oldground.haydar.dev

Spent two hours earlier today adding a feature to NomadNet to let me hide all the spammy peer/propagation notifications from the announce list. This way I can more easily see when new nodes that might have content on them come online. https://codeberg.org/lethe/nomadnet

>what are you working on? I'm currently working on writing an IRC server in C++ I'm trying to write it in a functional style. >what projects are you putting off? I really want to write a semi-3d raycasting game which has NPC's controlled by AI. Bit of a vague idea, but I think it could be interesting. >what do you enjoy doing technology-wise? I'm currently learning some Lisp and Rust and enjoying it a lot.

I've posted about it before in the desktop thread a while ago, but I'm working on a X11 GUI C11 library that implements my design GUI preferences. This thing is primarily written with the help of XCB for anything X related and Cairo for drawing. The end goal of this project is to become a building block for DEs / shells targeting all of my devices, desktop, laptop and (linux) smartphone. And I finally pushed it to its first public release ! Thought currently its still hardly usable for most programs because it lacks widgets (for this first release I only implemented the simplest ones) but that's should be coming this year (assuming I have enough free time to work on it) and the engine that manages the layout, windows, events, inputs and provides the API to build widget is up and running. https://codeberg.org/fraawlen/Derelict_Graphics

>>75008 i like the glassmorphism and the alternating vidya backgrounds on the login screen a lot. gives me a lot of nostalgia for my old internet friends i met on there back in 2018 (almost 6 years since then, wow) (´• ω •`)ﾉ discord’s been downhill ever since they removed their tag system, plus now there’s too much bloat with custom profiles and the profile themes. do you think you’ll make a way for people to find different servers? anyways, i dig this project lain, good work. i like it a lot.

>>75169 Really happy for your first public release! Awesome that you are still working on it :)

>>75169 That is seriously fuarrrking cool. I would wanna use this for all kinds of embedded devices. I can see audio/video systems using this.

>>75172 >>75173 Thanks! >I would wanna use this for all kinds of embedded devices These are also part of this project's target (as long as they can run xorg), which is why the the config lets you customize how each key or mouse button is interpreted, and all common shortcuts are user-configurable too. Finally, custom application specific shortcuts have 12 dedicated key (or button binds), by default mapped from F1 to F12, and appear in the windows X properties, so that a third-party desktop / shell component can show them, kinda like the bottom row of the screen of pic related.

>>75170 Thanks lain, that means a lot!! Nostalgia motivated me Yeah I plan to add some kind of discovery thing for people to find different server, and keep it simple instead of UI bloat, maybe I'll switch to an 2018 client

released the cross platform window creation library used in my booru viewer/music player etc since the wayland backend is mostly ready https://gitlab.com/chino-chan/ikafb

I have been working on this imageboard software: https://codeberg.org/cirefl/yutai-invites It only has three dependencies (OpenSSL, SQLite, ImageMagick) besides the Zig compiler, so it should be easy to set up if anyone wants to try it.