Just bought a new computer and I am using it as an excuse to finally move entirely off Windows, so I am setting up Arch and figuring out how to get games working and all that.
Trying to find Europe-based partnerships for Uruky and investors for a different product (not under the same company because it'll require machine learning), which needs a good amount of cash investment upfront. Also trying to find some part-time work to help fund the lack of a paying job.
Programs output one of two binary formats, one being pure structured object data based on CBOR and the other being a kind of XML markup (the latter is what you see in the top screenshot: charts and a table nested within flex layouts). Then end goal is a PowerShell/nushell-like environment where external programs, not just builtin commands, can output structured data and, uh, nicely formatted markdown documents without resorting to sixels or custom terminal protocols for things like drawing borders on cells.
Still a lot to work out, but I'm hoping to have a writeup soon which I might post here.
I am working on a PDF-merging program. This is to merge several pages from different PDF-files into one big page (not just concatenating the pages after each other). This is to align several maps produced from AutoCAD into one big map while removing overlapping sections.
While it's a small project I have learned some new things already about MuPDF and I'm using Raylib for the simple UI.
Working with PDFs like that is both a hassle and fun in a way for me. I’ve got a little script I knocked together at some point that takes 5.5x8.5in half-page PDFs as input and imposes them in the right order on 8.5x11in paper so that I can print on regular paper and make a single slice through the stack. It was a little tricky figuring out the right sequence, but not bad overall. We use it for making flight test cards and it’s saved me and my colleagues a surprising amount of hassle.
Working to perfect mddog, a terminal tool that lets you preview Markdown in the terminal as if you're on GitHub or Obsidian: headings at the correct sizes, images rendered, etc.
I look at Markdown files a lot for work and prefer to just stay inside the terminal without switching out of it so this has been a game changer.
It basically works by building out the Markdown into its real format using macOS built-in engines like CoreText and CoreGraphics, then send the outputs to the terminal as image strips. I'm enjoying learning about the terminal, Rust, and the kitty graphics protocol throughout this process.
If you give it a try, let me know how you like it!
Coincidentally, I was working on a similar scenario recently: I wanted to convert all Markdown files in a repo to HTML, preserving links between Markdown files (as well as to other resources), for later viewing in a terminal browser (like w3m). I had gotten tired of piping cmark-gfm into w3m, and I've now got a 500 KB self-contained proof-of-concept (I actually just hacked it into my luasmith static site generator).
The most similar tools I found prior to starting were grip (which depends on a GitHub endpoint) and Ghore (which is built on Node, which I've been successfully avoiding lately).
mddog looks nice and is much fancier, but I needed to support older terminal emulators... and NetBSD :)
I love when people's pain points coincide. I'll need to look at grip and Ghore -- they both sound like interesting projects. Good lock with building out your solution! Curious to see it if/when it goes public.
I intentionally narrowed mddog to macOS and newer terminal emulators to ensure its fundamentals are well thought out and work properly first. Thinking about expanding its support down the line...but we'll see.
Retaking my radio amateur exam... I almost passed the first time, aced 4 out of 5 sections, but alas the legal part tripped me up. I'm confident this time I'll have no issue.
Fighting an e-mail battle to return a faulty solar battery, that kept shutting itself down with an emergency error about voltage overloads. It's been three months since I opened my case and the company is really drawing things out.
At work I'm plugging holes. LLMs are relentless in finding issues.
I've been working on etoku, my Japanese language learning platform.
The idea is that it tracks your understanding of the language per concept (vocabulary word or grammar point) and feeds you sentences from real TV and movies that review those or teach you one new concept. Right now I'm refining the system for choosing review sentences because I find it picks sentences that are optimal (review as many concepts at once as possible), but are too challenging, especially for users who don't know a ton.
If you want to talk about language learning let me know at will@<the domain I linked to>! I love talking about this stuff.
regulator | 10 hours ago
Just bought a new computer and I am using it as an excuse to finally move entirely off Windows, so I am setting up Arch and figuring out how to get games working and all that.
brn | 12 hours ago
Trying to find Europe-based partnerships for Uruky and investors for a different product (not under the same company because it'll require machine learning), which needs a good amount of cash investment upfront. Also trying to find some part-time work to help fund the lack of a paying job.
technetium | 9 hours ago
Inspired by previous comments, was working on a new kind of terminal:
Programs output one of two binary formats, one being pure structured object data based on CBOR and the other being a kind of XML markup (the latter is what you see in the top screenshot: charts and a table nested within flex layouts). Then end goal is a PowerShell/nushell-like environment where external programs, not just builtin commands, can output structured data and, uh, nicely formatted markdown documents without resorting to sixels or custom terminal protocols for things like drawing borders on cells.
Still a lot to work out, but I'm hoping to have a writeup soon which I might post here.
mcksp | 5 hours ago
I was thinking about something similar for a while, i will definitely read your writeup :D
ThinkChaos | 11 minutes ago
I'm getting 404s on those screenshots
vincent-uden | 12 hours ago
I am working on a PDF-merging program. This is to merge several pages from different PDF-files into one big page (not just concatenating the pages after each other). This is to align several maps produced from AutoCAD into one big map while removing overlapping sections.
While it's a small project I have learned some new things already about MuPDF and I'm using Raylib for the simple UI.
tonyarkles | 8 hours ago
Working with PDFs like that is both a hassle and fun in a way for me. I’ve got a little script I knocked together at some point that takes 5.5x8.5in half-page PDFs as input and imposes them in the right order on 8.5x11in paper so that I can print on regular paper and make a single slice through the stack. It was a little tricky figuring out the right sequence, but not bad overall. We use it for making flight test cards and it’s saved me and my colleagues a surprising amount of hassle.
quan | 9 hours ago
Working to perfect mddog, a terminal tool that lets you preview Markdown in the terminal as if you're on GitHub or Obsidian: headings at the correct sizes, images rendered, etc.
I look at Markdown files a lot for work and prefer to just stay inside the terminal without switching out of it so this has been a game changer.
It basically works by building out the Markdown into its real format using macOS built-in engines like CoreText and CoreGraphics, then send the outputs to the terminal as image strips. I'm enjoying learning about the terminal, Rust, and the kitty graphics protocol throughout this process.
If you give it a try, let me know how you like it!
jaredkrinke | 8 hours ago
Coincidentally, I was working on a similar scenario recently: I wanted to convert all Markdown files in a repo to HTML, preserving links between Markdown files (as well as to other resources), for later viewing in a terminal browser (like w3m). I had gotten tired of piping
cmark-gfmintow3m, and I've now got a 500 KB self-contained proof-of-concept (I actually just hacked it into my luasmith static site generator).The most similar tools I found prior to starting were grip (which depends on a GitHub endpoint) and Ghore (which is built on Node, which I've been successfully avoiding lately).
mddog looks nice and is much fancier, but I needed to support older terminal emulators... and NetBSD :)
quan | 8 hours ago
I love when people's pain points coincide. I'll need to look at grip and Ghore -- they both sound like interesting projects. Good lock with building out your solution! Curious to see it if/when it goes public.
I intentionally narrowed mddog to macOS and newer terminal emulators to ensure its fundamentals are well thought out and work properly first. Thinking about expanding its support down the line...but we'll see.
nemin | 8 hours ago
travisgriggs | 4 hours ago
Working with Nordic Thingy-91s. Power profiling with modern batteries is definitely a black magic (to me).
andrewsledge | 3 hours ago
Work: AI context layer management/ontology management for shared knowledge layers
Home: building an axe throwing target
willmorrison | 3 hours ago
I've been working on etoku, my Japanese language learning platform.
The idea is that it tracks your understanding of the language per concept (vocabulary word or grammar point) and feeds you sentences from real TV and movies that review those or teach you one new concept. Right now I'm refining the system for choosing review sentences because I find it picks sentences that are optimal (review as many concepts at once as possible), but are too challenging, especially for users who don't know a ton.
If you want to talk about language learning let me know at will@<the domain I linked to>! I love talking about this stuff.
rs86 | 2 hours ago
I am having fun learning F* - although I wish it had a more Googleable name. Also in a rabbit hole about the Neolithic.
marcecoll | 10 hours ago
I've been building a DST platform (similar to Antithesis) and I'm onboarding a couple of friends so they can test their own systems.
ThinkChaos | 8 minutes ago
I'd be quite interested in reading articles about your experience building it and "key takeaways" if you write it someday!
fbegyn | 6 hours ago
Small stuff at work since I'll be at OSMC for the next few days.
I like events, but I'm a nervous "leaver" for travels. If I'm on the way I'm generally OK, but I always stress the leaving part :D
heavyrain266 | 8 hours ago
A place for long-form technical writing (mostly essays, not a blog), most likely going to be made with Astro.