Hacker News Reader

Comment by grabcocque | Hacker News Reader

I think it's important to separate the GPLv2 which is still vital and alive, and the GPL v3 which is now generally regarded to have been a failure.

View on HN

I am a huge supporter of GPLv3 and license most of my open source work with it. Still, the Linux kernel has been famously v2, and macos has removed as much v3 work as possible (to the point of using very outdated versions of many packages including bash).

Things like CLang/LLVM, and don't get me wrong, they're technologically amazing, still have a huge focus of not being in the GPL world.

Back in the early 2000s, the hardcore FSF people dreamed of a work with OSS replacements for Photoshop and every other major tool people felt they needed in the closed Win/Mac world. That never really happened:

http://penguindreams.org/blog/the-philosophy-of-open-source-...

> Things like CLang/LLVM, and don't get me wrong, they're technologically amazing, still have a huge focus of not being in the GPL world.

This is one of the key things.

Generally, I see that BSD/MIT/ISC/Apache licensed software is made by people who just want to make amazing things, and let other people make amazing things with what they've made. The license is just a way to make their code (which is the focus) available to anyone, without legal questions.

For GPL, it seems like a lot of people (including RMS) value the license more than absolute code quality. They would rather have worse software, under their idealogical license, than use better software under a less restrictive license.

> For GPL, it seems like a lot of people (including RMS) value the license more than absolute code quality. They would rather have worse software, under their idealogical license, than use better software under a less restrictive license.

RMS wants free software to be as good as non-free. He realizes that takes a lot of work, which he has dedicated his life to. We all realize it takes a lot of work, the difference is, some of us are willing to put in the work.

That said, sometimes the "ideology" behind a thing is important. Personally, there are certain products I don't buy, certain companies I don't endorse, becaues I don't want that blood on my hands.

And yes, BSD is a great license for people (companies) who just wants to profit off others' (work) and never give back.

Which is the perfect way of having an ideology. An ideology is about ideas. The fact that it is realistic comes second (and that's good, because most of the time we would debate on the impracticality of it until death, accomplishing nothing)

I find this schism very interesting. You have a lot of people now harming the GPLv2 communities by being nice. For instance there is a noticeable push for projects to go Apache which is GPLv2 incompatible.

OpenSSL is famously GPL incompatible but now it's only GPLv2 incompatible. Likewise all GPLv3 code is GPLv2 incompatible so you effectively have two worlds that both have similar goals but cannot share code work together. This is particularly striking in the Linux environment where the kernel is v2 licensed and the user land pushes really hard on GPLv3.

GPL 3 was about naking abuse and exploitation of GPL 2 harder. Seeing it as a failure shows our industry in a very sad light.

This might ve a bit melodramatic way to state it, but GPL 3 failing means evil thriumphing.

It boils down to a philosophical question about whether "tivoization" is good or bad. I am with Linus on this, I believe it is nothing wrong with it and I therefore doesn't support GPL v3.

There's nothing wrong with a world in which most citizens simply cannot control and modify most of the products they own? When everything runs software and communities of people can share ideas / updates that make the products better, you think there's nothing wrong with artificial restrictions that block people from doing this?

Think of Stallman's initial story: he could program the printer upstairs to tell users over the network when it was jammed. But with the new version of the printer, he couldn't get the source code. A Tivoized printer would give him the code and then refuse to run his modification. So, you think there's nothing wrong with "screw you, you can go all day and find out only hours later that all your prints failed, even though you know how to fix this"? That's what Tivoization does, and this mild example is just a very mild one. It could be hell of a lot worse.

Which is the reality as companies doing "tivotization" moved into MIT style licenses, which allowed them to do exactly that.

Even Apple and Google on their own products

I'm not sure GPL v3 is a failure, but it has created a really bad taste -- at least 3 times I've been unable to link GPL V2 and GPL v3 code. In one case I reimplemented one piece of code, in the others I went in search of a BSD library instead.

Take this with a grain of salt. If you can prove that you project can work without the GPL V2/V3 code being present (eg... the GPL code isn't a major contributing factor in the working of your program) then it stands that under copyright law (Here in Australia) that it isn't a Derivative work and thus not infringing on the copyright owner.

Playing it safe GPL V2, give credit to the source code in your credits page and push back any changes you've made to their original source. GPL V3, you're kind of screwed (even linking) that you need to adhere to their license agreement and have your source code open source.

Personally for open source projects I use GPL V3, and leave a comment in the license agreement the business can contact me for a difference license that comes with support contract.

To be clear, my problem isn't linking GPL and non GPL code, it is a GPL V2 project, wanting to include GPL v3 code! Same way the kernel can't merge GPL v3 libraries.

That's new. Do you have sources?

Reply on news.ycombinator.com