To read all comments associated with this story, please click here.
...various wireless network drivers written for OpenBSD (easily the best wireless support in the open source world) have been ported ...
Best?! Well, not if you talk to OpenBSD developers, it isn't. Is WPA even supported? Yes, I know that OpenBSD people does not believe in encrypting Wireless stuff since they think VPN is the way to go. Apparently the rest of the people disagrees by their actions. While that works nicely when you are using you own setup, it gives you immense problems accessing other open access points.
On BSD, I find Linus strange. While he agrees that sharing is the way to go, he feels that he needs to tell people what to do by forcing them (through GPL). Luckily, people that believe in the BSD principle, share without restrictions, does not and the OS-world is a much better place than it would've been without BSD. I see no shortage of code merging into FreeBSD so obviously people act differently from what Linus fears. It is a bit like freedom of speech, it is the right to speak those words you disapprove of that needs to be protected, not the opinions you agree with. I believe in completely free and open software so I will defend your right to freely branch and modify it. I will of course encourage you to share the modifications but I will not force you to do so as it is your code (only the modifications, the original source is as free as before, an incredibly common misunderstanding by some "GPL-only" pundits). I do see a place for GPL software, but for me it is an extremely poor choice for an OS-license. Feel free to disagree
Edited 2007-06-16 00:17
"""
On BSD, I find Linus strange. While he agrees that sharing is the way to go, he feels that he needs to tell people what to do by forcing them (through GPL). Luckily, people that believe in the BSD principle, share without restrictions, does not and the OS-world is a much better place than it would've been without BSD.
"""
Linus is not strange about BSD. He feels that if one makes their hard work available for others to use, he has the right to expect "tit for tat".
I sometimes find Theo a bit strange about BSD because he gives people the freedom not to give back and moans when some choose not to.
But don't take that as a criticism of BSD. Because copyleft licenses have the nasty side effect of sucking up OSS code from other free projects without giving back.
"Compatibility" in the GPL world, means "we can take from them". Usually, they cannot take from us. But few people of the copyleft mentality seem to seem a problem with that.
In the insterest of full disclosure, I should say that I like copyleft. I even like the new extensions that come with the GPLv3. (Though, as always, I'm more concerned with the GPLv3's divisive aspects than I am with the refinements in it.)
But I do respect the BSD philosophy, even if I do sometimes call it the "rape me" license.
In the end, I think that we are stronger for having different classes of licenses available. And note that I say "classes of licenses". I do not think that rampant license fragmentation is good. But BSD and Copyleft represent two substantively different views of freedom. And I believe that we are stronger for having a choice at that level.
I find Linus' arguments about the BSD license vs. the GPL to be a little misleading. How many people are running a "vanilla" official Linux kernel without some sort of patching? [...] I consider this to be "branching".
Linus explicitly states that branching is a prerequisite for merging. He never claims that branching is undesirable. He merely suggests that merging is more interesting and more productive than branching.
Frankly, I don't think the differences between GPLv2 and the BSD license really affect the likelihood of branching and merging very much at all in real life.
The reason why developers and projects use branching is to take existing code in a different direction. With BSD code, the possibilities are almost endless. But with GPL code, there are rules for what you can do with the code. This makes branching less attractive in the GPL world than in the BSD world.
That's not true. There is no rules in GPL in regard to what you can do with the code.
You can do anything you want. Even use it in proprietary products. The rules only kick in in regard to distribution, but that is irrelevant in regard to what you can do with the code.
From a philosophical point of view I prefer the BSD- or MIT-license - from a pragmatic view I prefer GPL. The latter one is a requirement as long as fundamental rights are not protected.
[i]Frankly, I don't think the differences between GPLv2 and the BSD license really affect the likelihood of branching and merging very much at all in real life.[i]
With GPL2 you must make your improvements available if you distribute your code, whereas with the BSD license you need not (although you certainly may)







Member since:
2005-07-07
I find Linus' arguments about the BSD license vs. the GPL to be a little misleading. How many people are running a "vanilla" official Linux kernel without some sort of patching? If you're running *buntu, (open)SuSE, Red Hat, or Fedora, you can put your hands down now. That's just the kernel, I am not counting distributions. I consider this to be "branching".
Now, it is true that on the BSD side of things we have FreeBSD (PC-BSD, DesktopBSD), NetBSD, OpenBSD, Darwin (Mac OS X). However, there is an immense amount of code sharing and cooperation between the members of the BSD family. For example, various wireless network drivers written for OpenBSD (easily the best wireless support in the open source world) have been ported to FreeBSD and NetBSD with relative ease. This would be a great example of "merging".
Frankly, I don't think the differences between GPLv2 and the BSD license really affect the likelihood of branching and merging very much at all in real life.