BartyDeCanter

Same here. I have been moving everything I can to self hosted FOSS, contributing to FOSS projects, and rehabbing old hardware. It’s been fun, I’ve met people from around the world and I’m getting tools I like to be even better.

Locally, I’m working with the library to start Linux days, where we help fix old computers and move them to Linux. There’s been a lot of interest due to Win11.

Oh yeah. I’ve had to do a small amount of it on much simpler systems for work from time to time, and it’s always been damn hard. Often rewarding in a weird way, but very difficult.

I had a problem on my work laptop with them about five years ago, but rolling back fixed it. Never on my personal machines.

Edit: TBF, I’ve never had a personal laptop with an nVidia card. I generally prefer to build my own desktops, though I do have a laptop. It has an AMD GPU, also with no problems.

What do you think Project Support is?

My partner bought a Skylight screen a month ago. I put it up, but it’s basically been unused since.

For me, there was this very early health tracking watch I got, which was so fragile that it would reset and lose all data if I did anything more active than walking.

Some Google TV that was well reviewed, but at some point shortly after I got it had a software update that made the UI unusablly slow. Like, 5-10 seconds to respond to every button click.

Eh, that looks like typical take home for a staff level engineer in a big city.

Edit: Assuming they get paid every two weeks, that’s an annual take home of $161,122. Depending on state taxes, insurance coverage, 401k contributions, dependents, etc, that’s a base salary of $200-250k. Which, yeah, that’s what I budget for a staff salary.

For consumer hardware supported by stock kernels? No advantage at all. At most you may want to switch kernels, but most distros have a handy tool for that.

The only time I’ve compiled my own kernel in the last 15 years has been for work on very specialized embedded systems.

Generally, yes, adapting to the size of the window is pretty easy. Most of those libraries have a layout engine, so you define the size of things relative to the size of the window and each other, then the layout engine takes care of the rest.

As for discoverability, well, it’s a hard issue in all UI schemes, and tends to be particularly difficult in CLI/TUI applications due to layout and input constraints. In a GUI application users can click on buttons, scroll, and generally figure things out. For a tui, there are probably going to be a large number of (possibly reconfigurable) keyboard shortcuts and maybe some sort of command system. How you let the user know about all of them without having to memorize a giant table can be difficult. The common options I’ve seen are a “?” popup (lazygit), a context-aware small popup during multi-key commands (helix), a command palette with search (lots), a top menu bar with accelerator keys (old school WordPerfect), or a bottom bar with context available options (lots). They all have their respective tradeoffs, and can make something go from “useable after hours of practice and reading” to “oh, this is intuitive!”.

Most languages have a fancy TUI library nowadays that makes it pretty easy. Here are a few:

• C/C++ - Ncurses, notcurses, ftxui
• Python - Textual
• Go - BubbleTea
• Rust - Ratatui

The difficult parts are generally:

1. Designing what it should look like
2. Dealing with all the interaction options
3. Making it all discoverable for your users

There are dozens of us!

Helix when I can install things, vi when I can’t.