Hacker News Reader

Comment by geofft | Hacker News Reader

Over a decade ago, I remember that we were debating the merits of CLAs that got all of OpenOffice.org owned by Oracle (cf. https://lwn.net/Articles/443989/ as a random starting point).

I really think this was the FSF's fault - they insisted on copyright assignment for any contributions to GNU, because if the FSF held the entire copyright, they'd be in a better place to pursue legal action against violators. But history has shown that this clearly hasn't been required: Linux doesn't have this policy, and no Linux enforcement case has ever failed because the Linux Foundation wasn't able to demonstrate clear standing. And the FSF basically legitimized everyone else asking for CLAs. Had they stood firm and said that giving up ownership of your code to a central entity is antithetical to the spirit of free software, and the license is all that people can rely on, I think we would have been in a much better place today.

View on HN

Unlike MIT/Apache/BSD-style Open Source licenses, Free Software can't be relicensed by anyone but the copyright owner. It's a completely different situation.

The current problems stem from pretending that Open Source has any more of a relationship with Free Software than it does to proprietary software - an expectation that reifies (and honestly necessitates) things like "the Spirit of Open Source" in the minds of developers. The Spirit of Open Source is that you're working for massive companies who can ignore you.

The only relationship between Open Source and Free Software is the fact that Open Source code can be arbitrarily relicensed and therefore can be relicensed as Free Software as easily as it can be as proprietary software.

Please stop spreading incorrect information. Re-licensing is an exclusive right granted to the copyright holder; it has a specific meaning under copyright law. There is nothing in the Apache, BSD, or MIT license that grants re-licensing. Using source code that has been licensed under a permissive license in a larger work that is licensed differently (including more restrictions) isn't re-licensing.

Absolutely true. There's a consequence of this that people tend to ignore: You cannot remove the (e.g.) MIT license text from an MIT-licensed work if you redistribute it as GPL.

The MIT license gives you "Permission ... subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software."

You can abide by that condition by distributing your software under the GPL and retaining the MIT license text for portions. You cannot abide by that condition by "relicensing" the software and removing the MIT license: you are in violation of the license.

Fair distinction in MIT's case, and other attribution-ware.

> Using source code that has been licensed under a permissive license in a larger work that is licensed differently (including more restrictions) isn't re-licensing.

Whatever you want to call it is fine. I prefer to think that GPL'd software that incorporates Open Source takes a copy and makes it GPL (especially because if I modify the Open Source at all, my changes are not available under an Open Source license.) Other people who are using that piece of Open Source are not using my copy, which is GPL.

MIT, Apache, and BSD are "GPL-Compatible Free Software Licenses" according to the FSF: https://www.gnu.org/licenses/license-list.html#GPLCompatible...

I gather you have some meaning of Free Software that is different from what the FSF means by it, and also different from what groups like Debian and Fedora mean by it. Can you expand on what you mean by "Free Software" and what relationship it has with the "Free Software" movement as defined by the FSF, Debian, Fedora, etc.?

Do you think that the FSF, therefore, works within "the Spirit of Open Source", that contributors to glibc are working for the FSF, which can ignore them?

> "GPL-Compatible"

Of course they are. They can be placed into GPL software. They can be placed into any software.

> Do you think that [...] contributors to glibc are working for the FSF, which can ignore them?

Yes.

You're selectively quoting. The important part wasn't that the FSF claims that those licenses are "GPL-Compatible," which nobody disputes; the important part was that the FSF claims that those licenses are "Free Software Licenses." Do you dispute that?

Elsewhere in this thread you imply that you believe that the GPL is not an Open Source license. Can you clarify that, too? Because, again, it sounds like you're using terms that don't match what the Free Software and Open Source communities mean by those terms. (In particular, GNU would disagree with this claim - https://www.gnu.org/philosophy/free-open-overlap.en.html - as would the OSI, of course.) So you should define them.

Reply on news.ycombinator.com