Hacker News Reader

It says "copyright microsoft" in that license file. Just because THAT file is MIT is irrelevant. They didn't retain the original license file. They should have APPENDED to it, keeping the original copyright holder name, otherwise it's just blatant copyright infringement that coincidentally is released under the same license.

I am not a lawyer, but I imagine a lawyer would find it alright if they just restore the missing notice. I do not imagine there is much else that can be done here since he cannot really claim to have been significantly damaged by the absence of a single line, but these matters are best discussed with attorneys.

I imagine a lawyer sending them a settlement offer for the blatant copyright violation they committed would get them to settle for a five-digit amount, since just the cost of discovery (and potentially having the "let's just fork it" dirty email laundry aired in public court) would be much higher.

I doubt this developer has the money to burn on a lawsuit like that.

The letter is cheap, and there might be law firms willing to fund it for a cut of the profits, making the threat credible.

IANAL but my understanding from floating around open source licensing circles is that you'd have a hard time with the judge if you didn't just ask for the license to be put back as step 1. Microsoft willingly not restoring the license would be more problematic.

The forgiveness clause in GPL 3 is as much an acknowledgement of actual reality than anything else.

Just the absence of a license generally means the creator has all right reserved by default. You don’t need a license in every file because in much of the world copyright is given by default to the creator. A licensed file is permission to do something with that copyright material.

He had a top level license file that presumably applies to all files. He would not be the first to do that and will not be the last.

That said, if Microsoft had forked before the LICENSE was added or stated somewhere, they were reusing all-rights-reserved code, which is definitely copyright infringement. Again, I am not a lawyer.

> but you never put copyright notices in your files.

I thought having a LICENSE file in the project's root directory was sufficient. Is it not the case?

It is a fairly standard practice in at least some open source communities to add copyright notices to files that people have changed significantly, although there is no well defined minimum threshold for how much permits them to add a copyright notice. Thus, someone else can come along, fork the project, add copyright notices to all of the files and then give the impression that they wrote them, since there is no attribution aside from the one LICENSE file that you wrote. The git history might show the truth, but if they copy the files into a fresh git repository, that metadata will be lost. Projects take files from one another all the time, so there is no guarantee that they will preserve your commit history and then anyone curious who wrote the code needs to do digital archaeology.

That said, file level copyright notices are not perfect (since only the VCS shows who added what lines and that might not be preserved), but it is better than nothing and it is something that is guaranteed to persist as long as people are abiding by licenses. If they are not, that is copyright infringement and the copyright holder can do things like send cease and desist notices in response to the copyright notices being removed.

Also, I must emphasize that I am not a lawyer, but one might argue that it was not willful infringement if someone removed a copyright notice from 1 file by claiming it had been a mistake. However, if they remove it from all files, then nobody is going to believe it was not willful.

Thanks! I have some open source projects where I only have one LICENSE file (it is also in README), but I will consider adding it to all files, there are just too many files. :/ I am inconsistent, because I have projects that contain the copyright notice in all files.

Sufficient but a good idea to put copyright in all files.

Technically if there's no license found then it should be considered automatically copyrighted, with no permissions to copy. So leaving copyright license out actually makes it less open source.

The license does not necessarily need to be in the files. It could be a project level license in LICENSE, which is what the author here did.

Yes that's exactly what I said :)

It's not the license that's important, but the copyright notice.

It really should be in all files, not because it’s legally necessary, but because it’s the best way to ensure that the correct license/copyright follows the code to which it applies.

Consider future contributions; the contributor’s copyright should apply only to the files to which they contributed.

Similarly, consider any code that you incorporate from external sources; that code’s copyright and license should only apply to the files in which it has been incorporated.

Lastly, consider the case where the code is copied out of your project to be incorporated in a different project. The license and copyright should follow with those files (and if your files don’t include copyright and license at the top, it’s very likely the person doing that copying will insert it themselves for this same reason).

I was responding to this:

  Sufficient but a good idea to put copyright in all files.
  
  Technically if there's no license found then it should be considered automatically copyrighted, with no permissions to copy. So leaving copyright license out actually makes it less open source.

I said it is not legally necessary to have that in every file in the previous comment solely to contradict the idea that his project level LICENSE file did not constitute a license for the individual files. I did not at any time suggest that it was not a good idea to put notices in every file. In fact, if you look at my other comments, you will see that I said that he should have put copyright notices into every file multiple times. I even said it at the root of this thread:

https://news.ycombinator.com/item?id=43750850

I would say: absolutely no (ianal). But I've had stand up arguments with colleagues in the recent past that I was unable to win. They wouldn't even ask the legal team for an opinion. But it's nice to see some evidence here that I was correct.

It's not required, but it's generally safer to put a notice saying who owns the copyright and what license the file is released under at the top of each file. Some licenses like MIT, the BSD licenses, Zlib, etc are short enough that you can include the full license text in the notice, and others like GPL provide sample copyright header text to include. Here's an example of this from a random file in the SDL source code: https://github.com/libsdl-org/SDL/blob/main/src/video/SDL_bl...

Obviously Microsoft is still committing copyright infringement and in the wrong here. However, if the author had copyright notices in each file and then Microsoft stripped them out or changed the copyright information, it would make it harder for them to brush it off with "oops, we forgot to commit the correct LICENSE file" like I'm sure they'll do here.

If they forked from before the author had a license, it’s worse. MS had no right to use it.

I’ve contributed to plenty of project that don’t have the per-file copyrights. It’s a choice not a mistake.

> If they forked from before the author had a license, it’s worse. MS had no right to use it.

You are right, provided he did not have a notice saying it was MIT licensed elsewhere.

> I’ve contributed to plenty of project that don’t have the per-file copyrights. It’s a choice not a mistake.

I would consider it to be both a choice and a mistake. The two are not mutually exclusive. There is no evidence in the fork that he is the copyright holder of the original code and it looks like Microsoft is. Part of that is Microsoft’s fault, but part of that is the original author’s fault for not including per file copyright notices, such that Microsoft could add theirs and be the sole one listed in every file.

I would not be surprised if Microsoft’s legal department doing a scan of public repositories for stolen code mistook him for infringing on “their code” given that they have no information that he authored it rather than their employee. It sounds absurd, but it has happened. I know for a fact the sg3 utils author added copyright notices to his code examples because he was getting contacted by companies, whose engineers incorporated his code into their projects without attribution, that thought he had stolen their code:

https://github.com/doug-gilbert/sg3_utils

I know that because he told me by email in 2013.

> There is no evidence in the fork that he is the copyright holder of the original code and it looks like Microsoft is. Part of that is Microsoft’s fault, but part of that is his fault for not including per file copyright notices, such that Microsoft could add theirs and be the sole one listed in every file.

Absolutely not! This is completely and only M$'s fault, whichever way you look at it. Copying a file and slapping your own license on it, without consideration of the original one, is never acceptable. Don't blaim the victim please.

As for incompetence - well maybe they (M$) need to get better at managing licenses? Accusing others of stealing when the reverse is true only makes everything worse. Let's not try to change the standard way of licensing because some developers can't be bothered to check the license (and even fix typos in comments, apparently).

As an aside, there is no need to add copyright / license to every file. I would even consider it an anti-pattern, because it pollutes the code with noise.

He made it extremely easy for a fork to make this mistake by not putting headers in every file stating the license and copyright.

If you look at any major project, you are likely to see license and copyright notices at the top. You even have the Linux foundation pushing to have standardized license notices at the top of every source file:

https://en.wikipedia.org/wiki/Software_Package_Data_Exchange...

That said, the engineer who did this is fairly young (7 years out of college) and does not appear to have any experience working on a fork previously:

https://github.com/avtakkar

https://www.linkedin.com/in/aviraltakkar/

He probably did not realize he had goofed. It does not take much imagination to guess the sequence of events:

  * He copies the source files into Visual Studio Code
  * Someone internal tells him he needs a license, so he grabs whatever Microsoft suggests while following internal guidance, without realizing that is only for Microsoft originated code.
  * Someone internal tells him to put it on GitHub, so it goes up with the wrong license.
  * He realizes that the files already had a license and tries to fix it. Presumably someone told him.
  * He misses the copyright notice that he failed to copy in the first place when fixing his earlier mistake.
  * People here hear about it.

Many inexperienced software engineers would do the same thing. This is not a licensed profession where you are expected to know all of the legal aspects before you may do anything.

> There is no evidence in the fork that he is the copyright holder of the original code and it looks like Microsoft is.

Only because they removed the license and copyright. If they were willing to do that in 1 file, they are willing to do it in many. It's not the authors mistake in any way shape or form.

It seems that only portions are from the original, so it is possible that they made the fork by copying files as they were needed. In that case, there was no removal of the notice. It just never was copied since they were using the project as an organ donor. That is a problem when using a single LICENSE file. It would have been handled had the original files had headers stating the license and copyright.

> but part of that is the original author’s fault

No: the original author could have made it easier to comply, and you could argue that he acted foolishly in not doing so, but that doesn't make it his fault.

Then who is responsible for his failure to add proper notices to the files? Microsoft?

If someone wanted to transplant only a subset of files from one project to another, there would have been no copyright notices to retain.

The copyright notice in the LICENSE file was still there. Bring it along for the ride -- either by putting all the files in one subdirectory, or by referencing it from each file, or by adding a note letting people know which files were under which copyright. They could even have also added copyright headers to each of the files in question.

There's more than one way for Microsoft to have retained the required attribution. The problem is that they didn't pick any of them. The author offered them valuable code, and the requirements of the license weren't onerous -- if they're unwilling to follow them, they should just not use the code.