Corporate Open Source is Dead
Open link in next tab
Corporate Open Source is Dead | Jeff Geerling
https://www.jeffgeerling.com/blog/2024/corporate-open-source-dead
https://www.jeffgeerling.com/blog/2024/corporate-open-source-dead
It's always been possible for these companies to pull the proverbial license rug from under the community's feet. It was just a matter of time before they did it.
Point is, you can't trust one powerful entity, especially not when money is involved.
It's not really a rug-pull in the usual sense though - of "all of a sudden you cannot use this product anymore"
You can still use it up to the commit where they changed the license. And then people just make a fork from there and the community moves away from the initial project to the fork
exactly. Forking for any reason is the essence of FOSS.
Scenarios like OPs were taken care of right from the start. That's just the legal side, tho. But someone still needs to do the actual work which is why it sometimes fails.
Another opinion about how it's OK for mega-corps to reap the benefits of opensource without contributing back and how we should just suck it up because of principles.
I understand he's a big contributor to ansible and other IaC, and most (all?) of his stuff is opensource. It's cool that he has the money to sustain himself and that he can stick to his principles with the money he has, but IMO either he lacks empathy or understanding for those in a place trying to make a living from opensource or in his opinion companies just can't be trusted with opensource if they have a CLA.
He's entitled to his opinions, but I don't share them.
I don't really understand the connection between the blog post and your comment. Could you expand on the connection between his stance against CLAs and your paraphrase about mega-corps and how we should "suck it up because of principles"?
It's the freeloading part of his post.
From a company's perspective, if they want CLAs or if they want to use an anti-open-source license, they do not care about your freedoms. They're protecting revenue streams. They'll often talk about freeloaders, whether it's Amazon building a competing hosted solution, or some startup that found a way to monetize support.
But in the end, even if you have GPL code and you charge people to get it, it's not truly free as in freedom, if the company restricts how you can use, modify, and share the code.
[...]
Freeloaders are part of open source—whether they're running homelab or a competing business.
I agree, I don't like CLAs, but the handwaving of the distinction between freeloaders is what gets me. There's quite a big difference between somebody running your opensource service in private and not making any money from it vs a business actively competing against you with your own code, creating support tickets, and not contributing back code (or minimally at best).
The argument is basically one of principle. "It's not libre / free as in freedom so it's wrong".
I guess I read his point more as being that it's effectively impossible for a license or CLA to distinguish "good" freeloaders from "bad" freeloaders, so it was inevitable that businesses would start doing license "rug-pulls" like the examples he gives.
"It's not libre / free as in freedom so it's wrong".
I think it's more "It's not libre / free as in freedom so it's not open source, don't pretend it is".
The "wrong" part would be derived from claiming its something that it isn't to gain some advantage. I'm this case community contributions.
There's not a handwaving distinction between open source and not, there are pretty clear guidelines.
"Only X is opensource because OSI says so!". It's like believers referring to the bible for stuff without thinking themselves. "The bible says it's so, so it must be right!", while disregarding that circumstances have changed.
I mean, yes? That's a good summation.
The part where you get to call something "open source" by OSI standards (which I'm pretty sure is the accepted standard set) but only if you adhere to those standards.
Don't want to adhere, no problem, but nobody who does accept that standard will agree with you if you try and assign that label to something that doesn't adhere, because that's how commonly accepted standards work, socially.
Want to make an "open source 2 : electric boogaloo" licence , still no problem.
Want to try and get the existing open source standards changed, still good, difficult, but doable.
Relevant to this discussion, trying to convince people that someone claiming something doesn't adhere to the current, socially accepted open source standards, when anybody can go look those standards up and check, is the longest of shots.
To address the bible example, plenty of variations exist, with smaller or larger deviations from each other, and they each have their own set of believers, some are even compatible with each other.
Much like the "true" ^1^ open source licences and the other, "closely related, but not quite legit" ^2^ variations.
^1^ As defined by the existing, community accepted standards set forth by the OSI
^2^ Any other set of standards that isn't compatible with ^1^
edit: clarified that last sentence, it was borderline unparseable
What many open source softwares do is write 5 lines of code, use autocode generators to convert 5 lines to 1000 lines and then automate code to have 200 commits in git repo with different datetimes. Basically, it is shown to the outside world that software is very complex.
We should all shift to Suckless
This is not even the topic of discussion. (Disclaimer: The rest of the discussion here is tangential to the original post.)
While I have no qualms with you using Suckless. As a person who has used suckless tools (dwm, st, tabbed, etc) for 2 years straight in the past. There are simply more performative and easy to work with tools and programs. I spent a few weeks worth of time in total over the 2 years trying to configure dwm alone to match with my workflow, applying and maintaining over 40 patches from the website along with some of my own written patches. When I switched to awesomewm, I easily reproduced my design in a single day and extended it further w/o much effort. Sure, the codebase is much bigger, but it is a fair trade for me if it reduces the time I need for configuration as a user while giving me the same performance.
I had, more or less, the same experience if I talk about st which I later replaced with Alacritty. The socket mode also helps me save RAM space when using Alacritty as I tend to have multiple terminal emulator instances open.
The stock distributions of suckless tools are always unusable unless you patch them to hell at which point they are definitely going to stop being suckless unless you really want to brainstorm and spend more time on it. (Read the Suckless Principles.) You can argue that criticizing Suckless Tools is in essence criticizing yourself because they are intended as canvases to paint on by the user. For me, it's just a waste of time and I would rather go for a program that is less hassle to configure & maintain.
Honorable Mention: Suckless Devs endorse Neo-Nazis in their team.
I agree that suckless should be simpler and should have a bit more code to ease user's life but they've deliberately kept it so so that the user understands what's going on like Arch or Gentoo.
The elitist group that thinks C99 is the best programming language?