Court finds calling stuff open source when it's not is false advertising
opensource.org
opensource.org
The court confirmed no such things. The decisions expressed in these two documents regarding the use of "open source" as a description of the product in question hinge upon the fact that someone else's software was released under a new license by Defendant, who had no authority to do so.
The court did not care to define open source, except to clarify that a license used previously by the Plaintiff is an open source license, and a license used subsequently is not. The court also did not consider any license-approving practices, let alone those of the Open Source Institute, of whom I find no mention in either document used to justify OSI's claim.
https://storage.courtlistener.com/recap/gov.uscourts.cand.33...
> Defendants' representation that ONgDB is a "free and open source" version of Neo4j® EE was literally false, because Section 7 of the Sweden Software License only permits a downstream license to remove "further restrictions" added by an upstream license to the original work.
It makes sense to me that if the court believes the defendants had no license to use the code at all, then describing it as any "licensed" version would be improper, whether their product was released under a "free and open source" license or even some other kind of license.
I don't fully understand the second part of the sentence about removing the restriction, but I think the court is just explaining why the defendants have no license.
> The court did not care to define open source, except to clarify that a license used previously by the Plaintiff is an open source license, and a license used subsequently is not.
In the Appeals Court decision I didn't see anything affirming that the Plaintiff's license (the text of which is AGPL + Commons Clause) was "open source". EDIT: There's more in the trial court decision.
> The parties agree that the truth or falsity of Defendants’ statements hinge on “the interpretation of Section 7 [of the Neo4j Sweden Software License], and GFI’s right to remove the Commons Clause from the Neo4j Sweden Software License.”
The court isn't deciding if ONgDB is "free and open source" according to a lay definition of the term, it's deciding if ONgDB has the right to license it in that matter, since that's what the parties are asking it to decide. In other words, if the statement is "ONgDB is a free and open source version of Neo4j", the operative words being challenged aren't so much "free and open source" as they are "version" (or more specifically, the implicit "validly licensed" phrase).
Because both parties agree that AGPL + Common Clause can't be classified as "free and open" the court can decide if ONgDB is "free and open" by determining if it's licensed under AGPL + Common Clause. The court doesn't have to have an opinion on what "free and open" means, because the involved parties are already in agreement.
Regardless of any historical context, in contemporary usage, "open source" has a commonly understand meaning and its what the OSI says it is. Other meanings make as much sense as pointing out that computers aren't machines, but a job title.
I would contest this, based on the many discussions there have been on HN about this, as well as many people (including many "normal" software developers who are not deeply invested in open source) I've met and talked to over the years having only a vague notion of what "open source" is exactly (often being something along the lines of "you have access to the source code").
I'd be willing to bet a substantial amount of money that if you were to ask 1,000 random software developers about Open Source, the OSI, and the OSD over half of them won't be able to give you a coherent answer on what these things are exactly.
This is the crux of the issue, really. Who did or did not invent the term isn't necessarily all that important: the thing is that in common understanding it doesn't really have the clear specific meaning in the minds of most, whereas for others – I would argue a minority – it has a very clear and very specific meaning, which inevitably leads to friction.
I guess the point of "Open Source was used before the OSI came along" is to demonstrate that it's a term that has a kind-of "obvious" meaning, and has been coined independently of the OSI's coining more than once. People don't hear "Open Source" and then consult a precise definition, they hear "Open Source" and assume it means what it says on the tin. The same applies to "Free Software", a term with 15 years more effort to explain what it means exactly, with even more limited success.
It seems to me that attempting to educate a landmass about a term with an "obvious meaning" is a futile effort; there have been many language pedantics over the years, and I don't know of a single attempt that has worked out. You're welcome to try as far as I'm concerned, but it seems to me that there are many more fruitful ways to spend time and effort.
I don't get why so many people are so obsessed with language. Over the last few decades terabytes of data has been used to discus this, and to what effect? I would actually argue it's been a net-negative since it has put many people off, and has been a distraction from far more pressing matters.
As far as "open source" goes, its in essence a brand of a cultural movement. And like real brands, some unscrupulous people want to enrich themselves off the brand by making a substandard product but pretending its the real thing. This is harmful to users who are swindled and it is harmful to open source devs whose reputation are stained by association. (To make the metaphor concrete, an example of a swindle would be someone makes open source software which isnt really, then sues all its users that use it in an open source fashion. This hurts the users, but also hurts people who make open source software because people cant trust the brand anymore)
I am unsure if you mean the OSI people or the people who disagree with the OSI definition here.
You're basically arguing to turn 'OSS' into a meaningless blob of text that anyone can interpret as they like.
How about if you pick 1000 random devs involved in licensing? I don't see that it matters what a random dev knows, so long as the group using the term knows and has a consistent definition.
To put another way, how many people actually know the precise definition of the term "organic" (as used in grocery stores)? Probably not that many, at best they have a vauge notion, and the term has much wider used historical definition than open spurce does. I would still say its morally wrong for a company to use an alternative "definition" to attract customers to their product. It would be false advertising in exactly the same way misusing open source would be, regardless of what their customer base knows about the details.
We have the terms source available and open source which are well understood and has no ambiguity but RMS and the FSF want to push their free software term and have to change open source so their term isn't obsolete.
It's nothing to do with being a purist or ideal. Because the wider understanding is exactly as pure and ideal. They just aren't the words RMS wants us to use for those meanings. He is obsessive but not for any real gain.
This is not true.
https://www.gnu.org/philosophy/open-source-misses-the-point....
> by Richard Stallman
> The terms “free software” and “open source” stand for almost the same range of programs.
The article goes into some cases where RMS considers OSS to cover things that are not free software but these are about small details and certainly don't include anything close to "source available".
No they haven't. The disagreement over free software vs. open source is about the idea behind it not about what rights are granted by open source or free software licenses.
> However, the court held that it was improper for the defendants to remove the Commons Clause, and therefore the defendants’ claims in advertising that its ONgDB software was open source was false advertising.
> Even though the AGPL has a clause specifically allowing downstream recipients to remove "further restrictions" like the Commons Clause, the court stopped the defendant from doing just that.
The decision of the trial court reads:
> Section 7 of the Sweden Software License only permits a downstream licensee to remove "further restrictions" added by an upstream licensee to the original work.
"Sweden Software License" here refers to the AGPL+Commons Clause.
Archive link: https://web.archive.org/web/20220317154012/https://opensourc...
> All other non-permissive additional terms are considered "further restrictions" within the meaning of section 10. If the Program as you received it, or any part of it, contains a notice stating that it is governed by this License along with a term that is a further restriction, you may remove that term.
> "This License" refers to version 3 of the GNU Affero General Public License.
Thus I don't think that references to "this License" within the portion of their custom license that is the verbatim AGPL text would automatically expand to also encompass the tacked-on Commons Clause text.
It seems that this question was already decided by the court in an earlier case, Neo4j, Inc. v. Graph Found., Inc 2020.[3] They decided that only restrictions added by downstream licensees could be removed in this way. The original licensor can impose additional restrictions, and those cannot be removed, because it would be "contrary to the principles of contract and copyright law" to allow such a thing.
1. https://storage.courtlistener.com/recap/gov.uscourts.cand.33...
2. https://storage.courtlistener.com/recap/gov.uscourts.cand.33...
Could this be used as precedent if you release software and claim that it's open-source, but don't provide full source or even any source at all? Probably not IMHO.
In the case of AGPL, you have an original open source release A that is forked to create B, and restrictions are added to the license of B. The original license of A permits those restrictions to be removed by someone who forks B into C. C is under no obligation to abide by those additional restrictions added by B.
But for that rule to apply in this case, the original code must have been licensed without such restrictions. If the original code, as in the case of Neo4j EE, was licensed under AGPL+Commons Clause, then that combination IS the original license. Downstream forks cannot remove that Commons Clause restriction and claim that the fork is "Free," because the original license is still in force, restrictions and all.
That would be unfortunate, because there are companies that claim their products are "open source" even though they clearly aren't. For example pritunl markets their product as open source on their website, but while the source code is available on github, it is licensed with a proprietary license that doesn't let you do basically anything with that source code.
Absolutely they should control the term. Just look at all of the companies out there with "source-available" licenses masquerading as Open Source. Neo4j themselves here uses "Open Source" in all of their marketing materials and should be just as guilty as this fork of false advertising.
Not sure who else would be better.
2. There are organizations other than OSI which have their own opinions on what they consider a “valid” open source license.
3. Some licenses are functionally open source but not approved by OSI because of their limited use. Furthermore, public domain works aren’t licensed at all.
4. OSI has no trademark on “open source”.
OSI’s opinion is only as good as other people agree with it.
What other organizations have credible claims?
I’ve only heard of companies that have created non-open licenses and are trying to redefine the term (eg, open core, etc)
Personally, when Subway calls their sandwiches "good", I think they are actually making non-good sandwiches, and they're trying to redefine the term.
Both "good" and "open source" are descriptive terms that are subjective. The OSI recognizes this themselves:
> "Open Source" is not and cannot become a trademark.
> there is virtually no chance that the U.S. Patent and Trademark Office would register the mark "open source"; the mark is too descriptive
https://opensource.org/pressreleases/certified-open-source.p...
Credible to me, but I don’t get to decide what’s credible. We live in a society, etc etc. that’s why I asked.
Subway can’t call the cold cut combo sandwich a fried chicken sandwich even though I think they are a credible organization. They would be wrong. If there’s some other org that is credible and is credible using open source, I’d like to know.
Just like Subway would not be credible if they claimed their non-organic sandwich was organic.
The examples I’m familiar with are companies trying to benefit from open source as a positive marketing term and I don’t think are credible.
OSI might vehemently argue against people using “open source” to describe source available works, but… people do it. Someone does it in every comment section discussing this topic. People post their source code on GitHub without a license and call it open source. For several decades before the OSI was founded, the phrase “open source intelligence” was used to describe works that were public available, not those that were licensed in any particular way. It is, factually, not uncommon for people to use it differently than the OSI prescribes.
The fact that people have a disagreement on the use of the term is exactly why people have leeway in the use of the term.
To use food as an example, there are a ton of foods that are culinarily labeled “incorrectly”. The reason why they are legal to sell is because people used the words that way.
If you want to use “fried” as an example: consider refried beans. They are often not even fried once!
https://wearenotfoodies.com/culinary-misnomers-or-danish-pas...
“Organic” is a different story, it’s regulated in many places.
Their attempt to occupy the term is a little offensive.
Correction: They came up with the name, so it's theirs. I'll leave my public embarrassment here.
OS came to us from the OSD, written by the OSI and after hot debate in the early days of what should and shouldn't be Free Software. It was specifically a reaction to the FSF and a rejection of that group's activism, which didn't fit what they were trying to do.
[1]: https://en.wikipedia.org/wiki/Open-source_software (specifically the first section starting with End of the 90s)
https://lunduke.substack.com/p/who-really-coined-the-term-op...
Digging further is for extreme nitpickers.
Gutenberg didn't "invent" the printing press either. Movable type dates back to 11th century China if you really want accurate attribution. Except for the purposes of the history of the modern world, Gutenberg's invention is most relevant/accurate.
My internet usage started on university systems back in the 80s, along with BBS and usenet and I had never heard the term Open Source even once until 1998.
General awareness of the term and its meaning starts with the OSI.
I don't imagine the F/OSS ecosystem would be much different today had the OSD/OSI never come to be, unlike say the Apache Foundation or GNU projects.
Still, it could matter it it ever came down to legal proceedings. But in everyday, colloquial terms, I think the OSD just "is" the definition of Open Source.
Edit: Yeah, see @CogitoCogito's post for more on the "Open Source predates OSI" argument. The counter-argument, I suppose, is that they didn't "coin" the phrase in the strictest sense, but were possibly the first to rigorously define it and give it meaning within a specific context. Whether or not one accepts that argument, or a similar argument, is obviously open to debate.
They did not come up with the use of the word "open" in regard to computer software or standards. Any native or fluent English speaker could easily do that.
The term "open standard" was used in the 1980's. In the area of Unix, specifically, there existed The Open Group for Unix Systems (later called X/Open Company) and The Open Software Foundation. These two merged in 1996 to form The Open Group.
In the 1980's AT&T and Sun Microsystems collaborated on creating a standard GUI, and called it Open Look.
All these people were corporate cronies trying to create a monopolistic software ecosystem that they control; they used the word "open" insincerely: the obvious ingredient that was lacking was access to source code by anyone who isn't one of the oligarchs.
In that environment it was very easy for any random developer or user to come up with a lament along the lines "these Open This and Open That people have open standards and open systems and even software with the word Open in the name; the only ingredient missing is, oops, the actual open source code".
Debian and Red Hat have their own teams determining whether something is Open Source enough to allow into their distributions, and obviously the Free Software Foundation keeps their own list of licenses which they consider Free Software.
If these groups of people disagree on whether a particular license is Free/Open Source software, it's not going to be considered an Open Source license by everyone.
That's about it.
I mean this is just factually not true. They have no such right. They tried to trademark it, but failed because it's just a generic term.
I'm going to label my company Open Source because we're open about how we source the laptops we programmed the software on. They're MacBook Pros, and you're welcome.
Edit: After reading the applicable sections of the trial court ruling, it is no different - the court simply accepted the plantiff and defendant's shared use of the term. It didn't reference OCI or FSF or any other authority.
Open Source is a pretty dumb term. Free Software is also a dumb term for other reasons. It's a pity.
- the plaintiff's software was under the AGPL+Commons, which makes it non-open-source
- the defendants copied the software, re-releasing it under bare AGPL, under the clause that says you can remove upstream restrictions
- that clause is actually about licensees, and the plaintiff isn't a licensee, they're a copyright holder. So the defendants had no right to remove it and their version can only be, and therefore is, subject to the Commons Clause too
- their declaration of being open source would have been valid if they had the right to license the work as they did, but they didn't, so it's not open-source
I assume that the plaintiff couldn't figure out a way to make a copyright license violation accusation stick, so they went after them for false advertising instead, and the judge judged it with the intensity of a copyright license violation because he knew the score too. This seems like a bizarre train for the OSI to leap on because it doesn't further anything about open-source and only furthers the interests of people who aren't.
The AGPL does not use that wording, but the court chose to interpret it that way because the alternative interpretation that a licence could restrict how the original licensor can convey their work was deemed to incompatible with the principles of contract and copyright law.
How does that work? It would seem like false advertising to claim I have to upgrade to remove their arbitrary limitations. I don't see how they legally enforce those limitations since they are licensing under a vanilla GPL v3.
I don't get how so many companies say they are "open source" but there's not a single line of uncompiled code available even on request.
The only thing it really addresses about the licensing, is that the license permits downstream users to remove additional terms added by upstream licensees. So original licensing of the licensor (adding the Commons Clause) is not allowed to be removed by a licensee as it doesn't fit the definition of an additional term.
Shouldn't Neo4j be sued as well for falsely claiming to be open source then?
Whoever sued these folks for false advertising should sue Neo4j as well.
How can it be legal to advertise a food products and the delivered results are nothing like what was advertised. Imagine a car sales person doing the same thing.
I'm thinking of launching one of our project in open source but don't want to end up in this kind or articles ^^
Other licenses may be open source, but you'd probably have to get lawyers involved to make sure. So it's better to just pick a license which the OSI considers Open Source.
If you don't want to rely on just the OSI, you can also check what the Free Software Foundation, Debian and Red Hat think of the license you've picked.
- Understanding Open Source and Free Software Licensing[1]
- Open Source Licensing[2]
And then Producing OSS[3] also contains a (very) brief section on choosing a license. It's worth reading though, for other reasons.
[1]: https://people.debian.org/~dktrkranz/legal/Understanding%20O...
https://www.oreilly.com/library/view/intellectual-property-a...
It will help you understand the relevant parts of the law (copyright, patent, trademark) as well as helping to differentiate between various types of Open Source licenses.
"this unlicensed project is closed source, source code is free to browse >here<"
"distributed unlicensed software (not in the public domain) is fully copyright protected, and therefore legally unusable... Examples of this are unauthorized software leaks or software projects which are placed on public software repositories like GitHub without a specified license" https://en.wikipedia.org/wiki/Software_license#Software_lice...
Deciding not to defend it means it's still copyright infringement to use the unlicensed code, and they could change their mind at any time.
I do however strongly disagree that "code without a license is not open source". publicly visible source code is extremely useful for research and education purposes, and I have in fact published code that is "free to view, research and experiment with" but contains no license and specifically states that you are not allowed to use the code. I still consider that code open source, but it is of course not "free" (as in speech) software. Please do not conflate the two.
They are the Open Source Initiative, stewards of the Open Source Definition.
The closest I see is a rule then when you're using the OSI trademark or logo you need to follow their definition of "Open Source".
If his definition has shifted, it might be entirely correct to say he's wrong.
You can also accept his general definition but validly disagree with how exactly to apply it.
Inventing a term gives you a lot of authority but not total control over what it means.
Had they not claimed they were “100% free and open source”, surely they’d have been able to change the license as they saw fit? Yes?
Open source is a prerequisite for truly free (as in speech) software, but the opposite is not.
OSI is claiming that the term "open source" is what is false advertising. But both the trial and appellate courts are specifically calling out the phrase "open source and free" as false advertising. Neo4j EE and ONgDB may be "open source," but they are definitely not "free."
If the defendants had only claimed that ONgDB was "open source" they might not have had any issues. But because they removed the Commons Clause and also declared it "free," they are in violation of the original license of Neo4J EE, a license which is still in force, and a license which is not "free."
If that statement is wrong, then 'open source just means source-available' is also wrong, for the same reasons. People who release OSI 'open source' software call it open source, people like GitHub who talk about licenses use OSI's definition, and people who release closed-source-available software clarify that it's not 'open source', using those words. The only people who treat 'open source' as meaning 'source available' are people who say 'free software' and want you to too.
And when you say 'free and open-source', the 'free' in that phrase does mean $0 - you wouldn't need 'open-source' if free meant FSF 'free' and open-source meant source-available, because FSF 'free' software requires source availability, so 'and open source' would be redundant.
Therefore you can either interpret FOSS as meaning $0 and source-available, or $0 and OSI 'open source'. The former would not have been false advertising, as it is in fact $0 and source-available. Therefore the interpretation the court is going with almost certainly involves OSI 'open source'.
But what I cannot do is take a piece of software that is not "free" and advertise that it is free.
That's what's going on here.
It's not really about any of that. If you read the court opinion, it's basically copyright protection infringement on the basis that someone copied a product and changed the licensing in violation of the original license. The false advertising claim is on the basis that the duplicate software is free and open source, because it can't be free and open source if the original license forbade that.
Can you quote the ruling where it defines open source, or Open Source? There are several definitions in the dictionary and OSI is only one of them. There is still debate in many communities about whether open source and a noncommercial clause are compatible or not. That's not something the court needs to look at in this case.
That's "source available". "Open Source" is widely understood to mean a bunch more, captured in the "Open Source Definition": https://opensource.org/osd
However, the courts ruled upon the composite term "open source and free". Under the OSI definition, open source implies free. But no court has yet so ruled.
There are also a handful of orgs (most prominently a few VC-funded companies) who want to freeload on the goodwill of "Open Source" to market their proprietary software. And there are some Free Software advocates who dislike the very term "Open Source". (Not that "Free Software" has court-endorsed clarity either.)
A few companies have tried it and backed off.
But everybody kinda knows that CC0 is more or less in the spirit of the OSD (so long as there are no patents involved) and OSI doesn't go around hollering about it, because it's not that important.
As opposed to the Commons Clause, and other historical attempts to undermine "open source" as being nothing more than "source available" — those are battles worth fighting. The OSD is more important than OSI. If OSI were ever to change the OSD in a substantive way, there would be a huge revolt against them.
From https://opensource.org/faq#cc-zero
> CC0 was not explicitly rejected, but the License Review Committee was unable to reach consensus that it should be approved, and Creative Commons eventually withdrew the application. The most serious of the concerns raised had to do with the effects of clause 4(a), which reads: "No ... patent rights held by Affirmer are waived, abandoned, surrendered, licensed or otherwise affected by this document.". While many open source licenses simply do not mention patents, it is exceedingly rare for open source licenses to explicitly disclaim any conveyance of patent rights, and the Committee felt that approving such a license would set a dangerous precedent, and possibly even weaken patent infringement defenses available to users of software released under CC0.
So it's more that there are concerns and nobody has cared enough to fully resolve them to arrive at a clear answer. It's also not like the CC licenses are commonly used for software so this is not surprising.
For example: https://opensource.org/node/878
> it’s impossible to make a globally applicable statement that a certain piece of software is in the public domain
Not that something truly in the public domain is not open source.