I'm surprised though, to find out it's only PR #1349. With the repository public for all these years, I expect more idiots than 1349. 😏

I used to select MIT primarily to keep my name associated with the code. The idea being if companies picked up my projects and used them in their products, it would be a good thing to have on my resume, and could land me a job.

I'm at a very difficult point in my career now, and the days of this xkcd are probably coming to an end. Nobody wants to use libraries anymore, they just vibe-code their own.

Now I'm reconsidering publishing my personal project source code at all, because I've rarely got contributions, and lets be real, a non-commercial license won't stop the code being stolen anyway.

I'm good at figuring things out. I don't know how to get code on GitHub even though I've lazily tried multiple times.

I get frustrated because I'm looking for something that I feel should be easily found (code on GitHub). Instead I'm given a puzzle I have to navigate through. I thought getting the code off GitHub was going to be the easiest part of my project. Nope.

Everyone wants your data so apps and websites are intentionally designed to keep people moving through them. I'm not saying that is what GitHub is doing here but it's the reason I have little curiosity to explore apps and websites anymore. It's flawed by design. Not because the opportunity to improve the end user experience hasn't occurred yet.

It was Ikey Doherty's of Solus and Intel Clear linux fame passion project; was, because he basically left yet another project behind. Now it's being carried over by solus contributors and other interested parties who are working on making it a stable base for power users and distribution makers to use of.

It's basically build atop of a declative Nix/Guix-like package manager if it was made for a more sensibly readable config language like kdl, essentially toml but good, and it only deals in packages, not configs, simplifying it a lot, thus, you can just use as if it was a normal imperative package manager (it isn't) and just:

sudo moss pkg add whatever

And it'll declaratively update the config file to add the package to it, which you can just copy to another machine to replicate the current system's state, all without having to edit a config file in a language that requires some knowledge In functional programming, nor a degree in compsci to create a package for the distro. "System model" they call it.

As far I know, solus is intending to eventually use it as base. They pretty notoriously banned AI to their distro codebase a while back, definitely not slop.

I forget how to find releases or download the code (instead of forking the project) just about every time I use github. He's got a point but god damn he also seems insufferable.

Ah, those python codes with requirements.txt containing libraries in versions that aren't available anymore, and the code won't run with the newer ones...

One more thing, and this could be a dealbreaker for some of you, the project is AI-assisted.

Maybe you should have gotten an AI to summarise the summary for you /s

Are people just not exploring websites anymore? Surely my household can't be the only one where if you don't know how to use a website or find something on it, we actively explore it. (Dodgy sites notwithstanding, I mean safe and legit sites). Am I strange that if I can't work out how to use something, I look for a manual. And barring that, I try different things, trouble shoot using logic, etc. to work it out for myself.... I'm not a programmer but I can't say as I had a great deal of difficulty working out how to use Github (as a consumer) for myself...

A user is not owed "ease of use"

I already covered this in the comment you just replied to.

At the end of the day, GitHub is a developer platform

This was also covered several comments up. If you don't have anything new to present, I'll bid you good day.

I think there is some pride and value in setting things up in a way that reduces friction.

I myself am of a similar opinion tho. Someone maintaining a thing as a hobby doesn't owe me shit. If it's too tedious to figure out, I'll move to another solution. It's really nice, when I get the option of a consumable bundle. But if they can't, don't like or simply don't care for ease of consumption — tough luck for me I guess.

tl;dr

Don’t look a gift horse in the mouth

Over on programming.dev, in this comment, you posted a chatbot's analysis of the flake. You then deleted it, perhaps after reading it and realizing that it was inane and uninsightful. At the same time, you're over here getting into a fight over the word "clanker", which you used to refer to the very same chatbots whose output you uncritically posted.

I understand that you have some inner turmoil, and that your behavior is merely an external facet of your emotional processing. However, I think that it's important to understand that the chatbot never has any ideas; indeed, it cannot ideate or cogitate. As such, your fencesitting is not gaining you any actual insights. To the current example, before you deleted the chatbot-output comments, I noted that the biggest and most insightful takeaway that the chatbot found was the ability to build a NixOS configuration locally, SSH to a remote host, copy the configuration to that host, and switch that host to the new configuration without rebooting; this is a standard part of NixOS which is in the man page, the official wiki, and the unofficial wiki. I get why you deleted the original comment, but I don't get why you posted it unless you genuinely believed that it was full of "hallucinate[d] useful ideas."

So do project that are under the GPL license force you to care how it’s used and force you to support them? This is a genuine question, not an argument, btw.

Copyleft only works because of copyright.

Does this use the same config performance settings able CachyOS? I don't use Cachy anymore but when I did, the minimal performance improvement didn't just come from using the kernel. There were configurations that also helped with this.

I never tried SELinux before and I don't know all that much about NixOS but apparently it has to do with how NixOS works under the hood which makes it hard and there's no incentive to actually struggle to implement support.

A big shame because SELinux model is something I would love to have on my system as I don't want everything to have access to everything by default.

I release code under MIT when I don't care how it's used. It's as simple as that. Especially for simple projects, projects I don't really want to support, and projects like libraries.

I build using GPL when I want community involvement.

read up on Secure Boot and consider turning it off if you can. makes life easier generally

Actually websites like Softonic would be neat for this kind of stuff.

Well some are not actually developers, they’re more like systems integrators who adapt a new component to a well established technology stack using the manufacturers documentation and files.

There’s also a legitimate security and safety argument to being allowed to have closed source. Haters will regurgitate what I was taught thirty years ago, that you can just diff the patch binaries, check the section that changed in a hex editor, cross that to resident memory and reverse engineer what’s going on from that point and bear in mind this was before ghidra, let alone ai, but that’s doing the jetpack backflip through the ring part of the meme as compared to the “look at source, walk across the mat, locate vulnerability, taa-daa” part.

The truth is that there is security in obscurity. If there weren’t we wouldn’t suddenly be seeing so many obscure vulnerabilities getting fixed due to ai scrutiny in the past few years.

Another reason a company might like mit is there isn’t any license wrangling work to do. Someone has to figure out what constraints in what jurisdictions the company’s product is under and what remediation is appropriate. Even if they don’t pay out a lawyer to figure all that out they had to have somebody do it and that costs money in the form of hours of compensated work.

A part no one likes to talk about is the fact that the mit license comes from mit, an institution with long running and deep ties to the intelligence apparatus of the United States and the mit license greasing the skids of adoption would make it a great way to increase uptake of compromised code. Which has never happened. Do not research this statement.

But yeah a lot of them are literally leaching off the commons. If a community doesn’t defend its well someone’s bound to come in and start charging for it.

Careful, you might anger the omarchy oligarchs

I don't see what rights were given up when protecting against tivoisation though? I don't think companies should have the right to exploit people's labor in the first place, especially if they lock-down whatever work they derive from it. I think Linux should be (A)GPLv3 and Linus is usually very wrong about many things.

But I see your point if you want to be explicit and careful about the future. But given the track record of GPL license improvements, I'm inclined to think that the "or later" suffix is actually a good move, particularly if your project has multiple contributors and you don't want to worry about asking permission later to relicense to whatever improved GPL may exist in the future.

Sounds like a shell script not a distro

Well, I admit I'm being a little facetious.

But that fact that "normie" users have zero tolerance for friction is a real problem. If you can't do anything that's not spoonfed to you, then you are prey for corporate abuse. The only people who really care about minimizing friction are the people trying to exploit "normies" for profit. Most open-source software is built to solve problems that are more important than "how can we exploit the general population a little more effectively?"

I don't think the solution here is for every single open-source project to devote a large percentage of their time and resources to eliminating friction. That would be a huge drain on teams that have more important things to do.

Rather, the solution is for users to be willing to spend a few seconds here and there to understand what they're doing and what their options are.

A user is not owed "ease of use" for every random thing on GitHub. It's appreciated, but unless something is given in return for the software, it's on the user to figure their shit out.

I'm not against helping users. If they're civil, they should be addressed. But anyone coming with anger and entitlement, like the original meme, should absolutely be shown the door.

At the end of the day, GitHub is a developer platform. And release distribution is a feature to make the developer experience better. Take it or leave it.

The main reason is that you don't know what rights you might be giving up in the future.

There was controversy when v3 was released that it closed the loop hole for companies to technically be in compliance of GPL but locked people out by using encryption keys. FOSS people, of course, said this was needed due to companies like TiVo at the time using hardware keys to prevent people from boot loading their own firmware when they were using Linux.

Torvalds supported the version but did not want to relicense Linux to force companies to stop the practice. And look were we are now: tivoisation is everywhere from Pixel to routers to tractors.

Your point that FSF is likely to make improvements of the license is valid, but you always want to make sure that you agree with the new license. They could, in future versions, say that by using the license, you transfer your rights to FSF.

Better to be explicit with the version you agree with and then relicense if needed.