Re: Notice of MIT License violation by Roberto Foglietta

Please don't feed the trolls,
I and most of us have better things to do than to get involved as specators to some sort of personal vendetta...

You guys should settle this outside of the mailing-lists.

Sent via BlackBerry Hub+ Inbox for Android


	  Original Message  	


From: [email protected]
Sent: 28 April 2026 00:17
To: [email protected]
Reply to: [email protected]
Cc: [email protected]
Subject: Re: Notice of MIT License violation by Roberto Foglietta


On Tue, Apr 28, 2026 at 3:41 AM Emmanuel Deloget via busybox
<[email protected]> wrote:
>
> 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 saying 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 in 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.

In the U.S. the law is DMCA section 1202(b). The catch here is judges
might interpret that the user is legal to do so under the MIT license.

Actually this Copyright Management Information (CMI) removal has been
claimed in many lawsuits in the U.S. that involve in generative AI
training.

As I studied these lawsuits I was forced to understand this law.

The AI companies like Microsoft and OpenAI are actively trying to
avoid CMI in their training. And d**n they had partial success in
arguing the cases (see, for example, Doe v. GitHub case that is
pending appeal).
_______________________________________________
busybox mailing list
[email protected]
https://lists.busybox.net/mailman/listinfo/busybox
_______________________________________________
busybox mailing list
[email protected]
https://lists.busybox.net/mailman/listinfo/busybox