Re: Notice of MIT License violation by Roberto Foglietta
Emmanuel Deloget via busybox <[email protected]> Mon, 27 Apr 2026 21:40:10 +0200
| Newsgroups | gmane.linux.busybox |
|---|---|
| Message-ID | <CAKh=nwGT_gfa8J5BXK3zc=mYZWERbfFpn5Jt27ud9rCrqoM49Q@mail.gmail.com> |
Le lun. 27 avr. 2026 à 20:33, Aaro Koskinen via busybox <[email protected]> a écrit : > > Hi, > > On Mon, Apr 27, 2026 at 03:53:15PM +0200, Ellie via busybox wrote: > > On 4/27/26 7:54 AM, Kang-Che Sung via busybox wrote: > > > I'm not a lawyer, but I think it's good to inform you that the MIT > > > license does not require attribution in generated binaries. > > "The above copyright notice and this permission notice shall be included in > > all copies or substantial portions of the Software." > > > > That's the quote in question, isn't it? Where does that say that it no > > longer applies to generated binaries? I am not a lawyer, and some points below might be wrong, so you should not assume that I'm sayinbg the truth. At the same time, I have been in charge of licensing issues for a few years now in the company I work for, so I worked on the subject with lawyers and paralegals, and I believe that while I may not be entirely right, I have at least some knowledge on how this works. If the license shown by the binary differs from the license ins the source code, then there is a problem. Given the state of the law (at least in France and in Europe, but I believe some part of this is also true in the U.S.), an open source license is a copyright scheme, as it defines how the copyright works for the source code (it's not only that ; it's also a contract on how you can use, distribute... the software, but that's not the point here). The fact is that the code was originally copyrighted by a person, and this copyright has been removed. The fact that the copyright in the source code does not change has no impact here, because most users will not get the information from the source code, as they get binaries. If the binary lies about the copyright, then there is an issue. Otherwise you could take linux, compile it, have a /proc entry that would says that Linux is distributed under the terms of the BSD licence, argue that you did not change the text of the LICENCE file, and everyone would be cool with it. No. The license text displayed by the binary shall be the license text itself. If a copyright line is removed, this is Bad and it is certainly not legal. In other words: you are not allowed to remove a copyright line. And this is really not useful here, because the MIT license allows you to *add your own*. There is no problem in having 274 lines of copyright in a MIT licence text. > > (When I include MIT dependencies myself I tend to interpret it as notice > > both in the source and at runtime, since I personally can't make much sense > > of it. So I tend to go with what I hope will avoid me trouble, as a > > completely uninformed person. Needless to say, please don't interpret > > anything I say as legal advice.) The license says that you can do nearly whatever you want with the code. It does not grant you ownership of the code, and it is still copyrighted code, so the original copyright cannot be removed.In fact, the license text itself explicitely says that the copyright line is of termendous importance ("The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software."). Best regards, -- Emmanuel Deloget _______________________________________________ busybox mailing list [email protected] https://lists.busybox.net/mailman/listinfo/busybox