> There's literally nothing to gain from doing so.

And yet KHTML was forked into WebKit which was forked into Blink, just to name one.

OSX has forked BSD code, as did windows.

This things happen all the time. MIT is a fine license, so is GPL, but we can't just say "oh nobody is going to fork this, it's dumb!" because it happens all the time.

> And yet KHTML was forked into WebKit which was forked into Blink, just to name one.

Bad example. KHTML was LGPL-licensed when Apple forked it (looking at https://invent.kde.org/frameworks/khtml, it may be dual LPGPLv2/GPLv3 or later licensed now). That probably is _the_ reason WebKit always was open source.

but I am not arguing that MIT is bad, I like it. I am arguing that "companies don't fork things" is false.

Maybe bad example, but it reinforces the position. Had KHTML been MIT WebKit would have been likely (more) closed source.

WINE is the best example, they switched after TransGaming's Cedega just released a commercial fork and contributed nothing back.

Why do you think this could happen to coreutils?

What could you possibly add to coreutils that would make a proprietary fork sustainable?

KHTML was LGPL. LLVM is Apache 2. Both are contributed to by companies.

The license matters less than you might think.

sure, but this is not what I am talking about, I am just saying that "there's literally nothing to gain from making forks" does not seem to prevent organizations from forking things.

You're talking about browsers. The parent is talking about grep.

Do you know of any place someone is maintaining a list of forks like this?

How can you maintain a list of private, closed source forks?

You maintain it as best you can. Many of the licenses require attribution. Sometimes things are announced and so on. Can it be definitive and complete? Obviously not. Could it be useful even so?

So this is great no? I’m happy windows and osx exist

I am not against licenses that don't require giving back, I am just saying it commonly happens.