Hacker News Reader

> It's almost like the FSF intended to trick the authors who choose AGPL into not realizing what terms they were licensing their work under.

It's the very opposite: it's to create clarity for all parties involved.

Alice publishes something under AGPL, Bob forks it and sneaks in an additional clause and Charlie gets a copy from Bob.

Now, is Charlie getting all the freedoms of the original AGPL or is he bound by the additional clause? The latter would allow Bob to trick Charlie into running closed source software masquerading as FOSS.

AGPL tries to prevent such sneaky behavior.

> A user might have wrongly thought that AGPL meant they could resell it

They 100% can.

That's what the court said which I agree with. But what the OSI is saying was the intent of the FSF and what the court case was about is very different.

The OSI, FSF and what the court case was trying to argue was that if Alice publishes something under AGPL with an additional clause. Charlie gets a copy from Alice, and decides to remove all additional clauses and to ignore them, then it is okay because the AGPL says that all additional clauses can be removed.

The judge said that this doesn't apply to the original licensor, that the original licensor can choose the terms of their own copyright, so if they want AGPL + additional clause then those additional clauses can't be ignored or removed. Only additional clause added by licensees can be removed due to the original terms they agreed too when using an AGPL library.

No, I think you're misreading the case. I think the intent of the clause, which the court reaffirmed, is:

- You develop didibusDB

- I enhance it. I add a clause. (note: as an enhancer, I'm the licensor)

- Alice grabs it from me. She can just remove the clause.

This clause avoids litigation for step 3. Otherwise, for Alice to do this, YOU would have to sue. This way, the power rests with Alice.

For the "original licensor" case, the license referred to here would have been AGPL+CC. I'm not sure of a legal argument to say I can't distribute something under a modified license. The only error was false advertising -- referring to it as an AGPL or open source or whatever other license, which the court found to be true.

Unless you have copyright on all the code, you're a licensee on at least parts of the code.

So while you're indeed the licensor of your enhancements, at the same time you're a licensee of the original code.

The nuance is a little bit more complex. Let's say I run an AGPL+CC project:

- I become a licensee of code under AGPL if I bring in AGPL code from an external project. Downstream can remove my clause.

- I don't become a licensee if I have an adequately strong CLA in place, and only accept contributions under the CLA.

- I become a licensee of the code if you make a PR to my project without a CLA. However, you're contributing code under AGPL+CC. Downstream cannot remove my clause.

That's my read on both the court ruling and the intent of the AGPL.

It'd be crazy for a court to rule that you can't modify a license. Once you have AGPL+CC, that's a new license (and it's not open source or compatible with most other sharealike licenses).

That's what I understood the court said yes, which I think makes sense. The original licensor is the one who dictates the rules of use, even to enhancers. But the OSI article implied this wasn't the intent of the FSF when they wrote the license, and that apparently the clause in the AGPL was intended to dictate the terms of the original licensor, somehow in that they couldn't add more terms to it.

Yeah, that argument makes no sense. The FSF is way smarter than that.

> It's almost like the FSF intended to trick the authors who choose AGPL into not realizing what terms they were licensing their work under.

Yeah I don't really understand the logic here. It's almost like they're considering themselves the root licensor for all projects that use the AGPL. Which (despite their intention) comes across as pretty hostile towards software authors.

If you don’t want to offer the freedoms of a licence then pick a different one. To choose one with an established meaning and then remove those freedoms undermines the purpose of having standardised licences. The authors are potentially setting users up for unintended infringement.

I see it as similar to organic food. Once you add non-organic ingredients then it’s not organic any more. You can do what you want but it’s fair enough if everyone else asks you to respect the labels they’ve built to let people make informed decisions.

> If you don’t want to offer the freedoms of a licence then pick a different one.

That's exactly what they did. Namely, they crafted a custom license that combined requirements from two other licenses -- granted in a confusing way.

> The authors are potentially setting users up for unintended infringement.

I do think the ambiguity in the contract language would certainly have bearing on any remedy the court would establish if they did find that the users infringed.

> I see it as similar to organic food. Once you add non-organic ingredients then it’s not organic any more. You can do what you want but it’s fair enough if everyone else asks you to respect the labels they’ve built to let people make informed decisions.

I'm not sure I see the similarities given that legal contracts are not a collection of labels. You don't sign a document that just says "lease" on it. If you're getting at the question of whether AGPLv3 + CC should be able to advertise itself as FOSS, then I agree, but that's not what this part of the conversation is about. If you're implying that they shouldn't be able to use the term "AGPLv3" in their license language, well I don't disagree but I don't know enough about the areas of law that would prevent that (is it trademarked? etc).

The FSF license is not just to protect the rights of the open source authors but the users too. If a software author doesn't understand this, they are the one who are to be blamed, as they don't understand the open source philosophy that the FSF advocates. One can ofcourse choose to disagree with some of their principles. In such cases, they are free to copy the FSF license an modify it as they want, but they can't call it an FSF license. Or they can choose a different open source license that aligns more with their philosophy.

That's what Neo4j did, they took the AGPL and added a Commons Clause making it AGPL with Commons Clause.

They then say:

> Neo4j Enterprise consists of modules from Neo4j Community Edition and modules licensed under AGPLv3 with the Commons Clause in this repository, and other closed source components not present in this repository.

I actually don't see them claiming their enterprise edition was ever open source, they also have an open source community edition licensed under GPL V3.

This is their License file:

https://github.com/neo4j/neo4j/blob/3.4.13/enterprise/com/LI...

You can see the: "Commons Clause" License Condition at the end of it.

Also, the history of the project, their enterprise edition started as AGPL, and later versions they added the Commons Clause over it, and finally now they've just gone closed source, probably to avoid all these legal woes.

Advocates of FSF argue that if Neo4j changed the AGPL license in a way that invalidates the intent and principle behind AGPL, they can no longer refer to it as an AGPL license. And if they do so it is false advertising. The court has agreed with this assertion and also said due to the changes in the license, Neo4j was doing false advertising by claiming it was "free and opensource".

> This just doesn't doesn't make sense to me lol. If that's the case, why would the licensors add more clause? They're basically irrelevant.

Licensor can add extra clauses as long as it doesn't introduce any restrictions to the freedom granted in AGPL. If the new clauses clash with the freedoms granted in AGPL, the license allows users to treat these additional restrictive clauses as not applicable to them. The provision to add more clauses in AGPL is to only provide more protection of the freedom of the user (the main intent of the FSF license). For example, to plug some loophole that the FSF did not consider.

> A user might have wrongly thought that AGPL meant they could resell it, and ignored reading the added clause.

AGPL license doesn't restrict a user from commercialising an AGPL software and redistributing it. So a user cannot be faulted here if they chose to ignore any additional restrictive clause stating that user can only use it non-commercially.

> But they also said that the intended clause from the authors mattered, not the interpretation of the user,

That's the part of the ruling causing the debate - FSF supporters feel in such a case, it means the judge chose to ignore that it is an AGPL license from FSF (thus choosing to ignore the intention and principles behind it), and perhaps chose to treat it a derivative but new license by the author(s) of the software. Hence their ruling that it not "open source and free".

The FSF doesn't choose what license my work is under. If I decide to use one of their templates such as AGPL and then I add more terms and conditions, it is clear I'm trying to add restrictions. The FSF has no authority on my copyrighted code and what terms I want to license it under.

But the FSF seems to assert that in fact, it was trying to sneakily prevent a copyright author from adding additional clauses if they chose the AGPL as a template for their license, by tricking them into not realizing that any additional clauses they'd add could simply be removed by the users.

This is what I find very preposterous. Luckily the Judge ruled against that, and asserted the rights of the copyright authors to be in control of their own license terms.

It's one thing to licence my code in a way that says, no one else can add more conditions to the use of my code, which is fine and what the judge said applied here, but it is another to license my code in a way that says that my own additional terms are dismissabble and non enforceable, which sounds from the OSI article like this was the intent from FSF.

There could be a question of are you allowed to reuse the AGPL license template if you are to add additional terms? Are there copyright on license templates themselves? But that's a very different debate, and it would be pretty crazy to need a license for the use of a license text. No license template currently provide such a license to my knowledge.

> If I decide to use one of their templates such as AGPL and then I add more terms and conditions, it is clear I'm trying to add restrictions.

And you are free to do that. But you cannot call it an AGPL license any more, especially if it clashes with the original intent of the AGPL license - to provide users the right to the source of the software, to modify it, and to redistribute it commercially or non-commercially under the same license directly or indirectly (SaaS).

The court also partly endorsed this position by asserting that Neo4j was doing false advertising by claiming its software was "free and opensource" because the changes made to the license invalidated this.

> Are there copyright on license templates themselves?

Some commenters have claimed here that they do, but they are also "open source". I don't know about that. But I am sure FSF has valid grounds to argue that if one of their license is changed to invalidate their original intent, such license shouldn't be referred to as an FSF license (like GPL or AGPL). That's what Neo4j was doing - they were calling their license "AGPL with Commons Clause". And even the judge pointed out that this was false advertising.

Which makes sense - Firefox is opensource, but if you modify it to include malware and redistribute it as Firefox, Mozilla will obviously not like that as you are damaging their brand. (Ofcourse, that would be a trademark dispute, not a copyright one).

> perhaps chose to treat it a derivative but new license by the author(s) of the software.

... I mean yes, that's obviously what a license with extra clauses is. What else would it be?

> It's almost like the FSF intended to trick the authors who choose AGPL into not realizing what terms they were licensing their work under.

It is almost as if the FSF existed to push a certain ideology and uses licenses as one of their tools. Don't use a GPL license if you aren't completely aligned with their values, the or later part alone makes that risky.

>the or later part alone makes that risky.

You are entirely free to pick a GPL version and don't add the "or later" clause.

Many projects, including the linux kernel, do.