Your comment about writing a text editor library is precisely what got me started with amp ~10 years ago, where much of its core functionality was extracted here:
I also think NixOS is more targeted towards developers. It’s one thing to learn the syntax, APIs, and abstractions of Nix/NixOS. It’s another to stack all of that on top of learning programming in general.
This doesn't mirror my experience at all. I think the biggest challenge facing NixOS is the learning curve. There's a lot thrown at you from the start, and as you start to factor your configuration into separate modules, there's a lot of complexity you have to unpack.
I've since migrated to a flake-based setup with machine-based variations (for my laptop and desktop), including easily swappable desktop environments. At a whim, I can switch between sway, hyprland, and gnome. This was mostly a result of me exploring/tweaking these without wanting to discard the configs; I always end up coming back to re-explore tiling WMs.
My experience through all of this has been great. I've even done a full re-install on both machines when the xz vulnerability was discovered and the process was effortless. That includes lanzaboote for SecureBoot, LUKS, and out-of-tree git-based flake builds for custom applications I build from source.
Yep, same. Thoroughly enjoyed Arch, but slowly accruing implicit config state means you'll eventually have to re-install and configure everything again. You can mitigate some of that with dotfiles and some hand-rolled scripts, but NixOS is effectively that "done right".
Oh, Arch was rock solid and I was also able to keep my install going for years. However, in the few instances where I would get a new laptop or experience a drive failure, it was non-trivial to get it going again. If you're using a stock install + Gnome, it's not bad, but I had lots of things set up (e.g. Secure Boot, udev rules, usbguard, tiling WM, etc.) and those are all things you can declare in a NixOS + Home Manager config, have versioned, and re-establish in ~30 minutes instead of days.
I get the point you're making, but I'd argue that even when stated in a reductive fashion, the complex system that leads simple cellular automata to artistic expression is orders of magnitude more elegant and special than AI.
This is the crux of the problem. Where Prusa is openly sharing, you have companies that are benefiting from that without reciprocating. Part of the tax you're paying when buying an i3MK4 is the continued investment in the open source/hardware contributions of the company, not just the end product. Shelling out $1k for a Bambu is your prerogative, but it does cast a vote with your wallet for a company that is more predatory than collaborative.
Bambu Studio is open-source and available on GitHub. Prusa complains that it's hard to upstream, yes--but when you look at Prusa's GPL'd firmware for their printers, that's even harder to upstream. They don't even try to upstream to Marlin! But they don't complain about that. At least Creality, of all people, sponsors Marlin development--like, with money.
So what, exactly, is your point here? Bambu is compliant and is reciprocating; OrcaSlicer derives from Bambu Studio and works great. Their printer firmware is not as far as anyone can prove derived from GPL software (and I tend to think that by now somebody would've found it), so they keep that.
Let's get to the brassest of tacks: one can talk about "contributing" until one's blue (or orange) in the face, but Prusa can't or won't ship a functioning, assembled-before-you-get-it multimaterial CoreXY for less than $2,500. Bambu came out the gate with one, not reliant on Marlin or Klipper, and it actually works.
I genuinely can't believe I'm having to point this out, because I think Bambu does suck as a company and sniffing about patents well-and-truly sucks, but they remain, and are the only, such bastard-coated bastards that can actually ship something usable at a price somebody can afford. $2500 for a Prusa XL would be more expensive than my CNC. I actually do need multimaterial, not for models but for tooling, so continuing to limp along with my collection of Klipper printers isn't reasonable. So what's your actual solution for a quality-outcomes tool at a reasonable price?
I can say that the FoxAlien kits are pretty okay? But I really needed something 24" x 24", and I will not recommend this one (I think the company that makes it is gone, even).
Litmus' goal is to help email marketing teams send better email. To do that, we've built a suite of tools to help with the building process, allow performing QA with device/client screenshots, enable collaboration on designs with teammates, and analyze email performance post-send.
We're looking to hire a full-stack engineer to help add/improve features and build new products. Day to day, you can expect to write Ruby and JavaScript (Rails, vanilla JS, Vue.js, Ember, etc.). We practice CI/CD with great test coverage, are sincerely good about work/life balance, and have some exceptional benefits (e.g. 28 days of PTO + stat holidays). You can learn more about our team and how we work here:
Litmus' goal is to help email marketing teams send better email. To do that, we've built a suite of tools to help with the building process, allow performing QA with device/client screenshots, enable collaboration on designs with teammates, and analyze email performance post-send.
We're looking to hire two Rails developers to help add/improve features and build new products. Day to day, you can expect to write Ruby and JavaScript (vanilla, Vue.js, Ember). We practice CI/CD with great test coverage, are sincerely good about work/life balance, and have some exceptional benefits (e.g. 28 days of PTO + stat holidays).
https://crates.io/crates/scribe
Good luck with the project! Building a text editor to scratch your own itch is fun.
reply