The terms of the AGPL are pretty easy to comply with

https://drewdevault.com/2020/07/27/Anti-AGPL-propaganda.html

by ssklash • 6 years ago

281 293 6 years ago

Yeah. People often resort to some fear-mongering when it comes to AGPL. When we released our software under AGPL, one developer of a somewhat competing app didn't find anything better but to claim that 'AGPL license makes it a useless toy'. Well, we don't agree.

woofie1112• 6 years ago

Awesome article! I was in an organization which had an AGPL ban. After a couple of months with lawyers, it's now a major AGPL supporter. AGPL lets you build ecosystems around your software, where everyone contributes, and no one can parasitically compete with you. It's absolutely the right tool for a lot of uses.

I'm no longer with the organization, but it had:

* Hundreds of open-source contributors

* Millions of users, mostly paying

* Zero direct competitors using the software. With equal ground for technology, it'd be almost impossible to overcome first-mover and branding advantage unless the organization really messed up

* Hundreds of major open source users doing interesting things with the software and using it in non-competing contexts. That translated into product enhancements.

With GPL or BSD, it almost certainly would have played out as lots of competitors doing an extend-and-proprietize, and instead of a successful open company, it would have been a dead organization.

So many misconceptions abound.

The flip side, though, isn't that everyone should use the AGPL. You should have the right license for the right context.

hpfr• 6 years ago

Anyone have guesses to the organization here, since they’re pointedly not mentioning it? I find that a bit confusing since their code is probably publicly available under the AGPL, which is really all this comment discusses.

debiandev3• 6 years ago

After a couple of months with lawyers, it's now a major AGPL supporter

I had a similar experience. I took exams on patent and copyright law with IP lawyers as teachers during my degree.

Then I worked in well known tech companies and had the opportunity to attend meetings with lawyers, business people, read contracts, etc. and oh boy!

Misconceptions abound terribly. The average techie overestimates his/her understanding of law and business.

"permissive" licenses are stupidly lenient. Companies never sign similar contracts between each other when exchanging goods or services. Contracts have hundreds of defensive clauses to mitigate risks or ensure fairness.

oefrha• 6 years ago

I find this comment puzzling. You're arguing from an IP owner's perspective. Whether you support using AGPL for your own software is completely orthogonal to whether you should use someone else's AGPL code in your proprietary product (assuming being proprietary is a done decision).

(Btw, one can argue that you have a vested interest in strengthening AGPL so that other companies would more readily choose your AGPL software and then be locked in to your ecosystem, so you're knowingly minimizing the associated risks on this thread.

You see what I did there? Yeah, attack on your integrity based on speculations, exactly like TFA.)

cycloptic5• 6 years ago

The simple solution there is to not make a proprietary product to begin with. Sorry, not trying to be snarky, but that (among other things) is the price you pay for that decision. It's baffling to me how some companies are so resistant to having to conduct legal reviews to use open source, which in a lot of cases will directly make them money, but at the same time these companies will gladly wave a giant NDA at someone just to start a negotiation.

lxe• 6 years ago

Any derivative works of AGPL-licensed software must also use the AGPL.

TBH, I'm interpreting this statement just like Google is:

Google states that if, for example, Google Maps used PostGIS as its data store, and PostGIS used the AGPL, Google would be required to release the Google Maps code.

What's the definition of 'derivative work' here?

woofie1112• 6 years ago

Eben Moglen talks about this at length (author of the license). The rough lines are drawn as per how closely the works couple and interdepend on each other.

For example, if I build an extension which works with Chrome and Firefox, over a well-defined API, that's an independent work. AGPL/GPL/LGPL does not apply.

If I have two pieces of code which mutually rely on each other and form a common system, and for example call back-and-forth, or have APIs specific to each other, that generally does form a derivative work.

There's a lot of literature out there about static versus dynamic versus microservice abstraction. That's mostly nonsense. Lawyers don't care about that. You can't get around the AGPL (or even the original GPL) that way, by isolating the GPLed code into a corner over a network service.

(disclaimer: not legal advice, get a laywer if you want legal advice).

p0llard4• 6 years ago

For example, if I build an extension which works with Chrome and Firefox, over a well-defined API, that's an independent work. AGPL/GPL/LGPL does not apply.
If I have two pieces of code which mutually rely on each other and form a common system, and for example call back-and-forth, or have APIs specific to each other, that generally does form a derivative work.

So why does the LGPL/Linking Exception exist at all? Software which links against libc doesn't have a mutual dependency; libc is quite happy to run crt0 as long as there's a main symbol somewhere, and yet apparently linking against libc would be enough to constitute a derivative work were it not for the exceptions in the LGPL.

apetresc• 6 years ago

A derivative work of PostGIS would be a fork or patches to PostGIS (i.e, a set database extensions for geographic/spatial queries). A mapping application that used PostGIS is not a derivative work of PostGIS, in the same way that a C program linked against glibc is not a derivative work of glibc.

At least, that's my layperson's understanding of it.

saagarjha13• 6 years ago

Note that there is a specific exception for that case: https://en.wikipedia.org/wiki/GPL_linking_exception

opus132• 6 years ago

in the same way that a C program linked against glibc is not a derivative work of glibc

I believe this is only as a result of the linking exception; I don't know if this has ever been tested, but my understanding was that linking (either statically or dynamically) against a library was generally considered enough to be a derivative work.

rbanffy• 6 years ago

A program using PostGIS via TCP is no more a derivative work of PostGIS than a program storing data in its underlying database is derived from PostgreSQL.

I can't imagine the tortuous logic one would need to arrive at a different conclusion.

JoshTriplett3• 6 years ago

You're conflating "talking to a server over a socket" with "linking to a library".

The former, as a general rule, does not create a derivative work. It's specifically called out in the GPL FAQ as something that does not create a derivative work. (Another case that doesn't is fork/exec of a separate program.)

The latter, linking a library, is why the distinction between the GPL and LGPL exists. As a general rule, if you link to a GPLed library, you do need to provide your source under a GPL-compatible license. (glibc is LGPLed.)

But people wouldn't use PostGIS by linking to a library, they'd use it by talking to a server.

taldo• 6 years ago

It's hard to draw a line. Considering that google3 is a massive monorepo, where linking a library into any binary (or library) is just a matter of a single line in the BUILD file, a lawyer could argue that such binary is a derivative work (or, it could be a lot of work to argue that it isn't).

Also considering the internal-only patches to OSS libraries, it'd be at least a pain, and quite possibly a wide avenue for trade secrets to leak (not to mention that most of those are so specific to Google's highly specialized production stack that are useless to anyone outside).

dathinab7• 6 years ago

But saying Goole Maps derives from a database extension which stores cordinates is kinda unrealistic.

It's like saying a public transportation company derives from a specific model of trains they use.

Sure the storage system enables Google Maps to work but any storage system with similar functionality would.

Ajedi32• 6 years ago

For what it's worth, the license doesn't use the term "Derivative Work". Rather, it explicitly prohibits bundling AGPL software with other software that you aren't able to license under AGPL:

You may convey a work based on the Program, or the modifications to produce it from the Program, in the form of source code under the terms of section 4, provided that you also meet all of these conditions:
[...]
c) You must license the entire work, as a whole, under this License to anyone who comes into possession of a copy. This License will therefore apply, along with any applicable section 7 additional terms, to the whole of the work, and all its parts, regardless of how they are packaged. This License gives no permission to license the work in any other way, but it does not invalidate such permission if you have separately received it.

This is a requirement which is common to many GPL-like licenses, and one which some developers object to (the later requirements for use over the network notwithstanding).

That said, I don't believe the above terms would apply to the hypothetical Google Maps example above, as the license's requirements for making the software controllable over a network seem to be significantly less restrictive:

Notwithstanding any other provision of this License, if you modify the Program, your modified version must prominently offer all users interacting with it remotely through a computer network (if your version supports such interaction) an opportunity to receive the Corresponding Source of your version by providing access to the Corresponding Source from a network server at no charge, through some standard or customary means of facilitating copying of software.

In that situation, the question of what constitutes a "derivative work" might come into play.

Disclaimer: I'm not a lawyer and this isn't legal advise, just my own interpretation based on my reading of the license.

anticensor• 6 years ago

GPL would be perfect if it had "and, you are not allowed to exercise any other permissions granted to for this work except for ones granted by this licence, even if you have separately received it or guaranteed by any higher authority, nor is the copyright holder allowed to grant such additional permissions to other such users." instead of "but it does not invalidate such permission if you have separately received it.".

dvt4• 6 years ago

Any derivative works of AGPL-licensed software must also use the AGPL.

That's why people don't use AGPL, and why I personally don't release any code under (A)GPL. If you want to make free software, make free software and release your code under Apache or MIT. And if you really care about attribution, use CC BY 3.0. Legally, you don't want to deal with the burden of constantly checking if you're complying with some license that a dependency of a dependency of a dependency is using.

By discouraging the use of AGPL in the broader community, Google hopes to create a larger set of free- and open-source software that they can take for their own needs without any obligations to upstream...

This is true, and I don't see much wrong with it. I don't like the idea of other code mandating how I should release my code. What if I just don't feel like up-streaming? Free is meant to be free, and (A)GPL does incur a cost -- at the bare minimum an ideological one.

A lot of reactionary down-voting going on; I'm quite open to debate. Why are you so passionate about AGPL?

saagarjha13• 6 years ago

Legally, you don't want to deal with the burden of constantly checking if you're complying with some license that a dependency of a dependency of a dependency is using.

If you're not modifying that dependency, you don't need to do anything since your dependencies are handling it. Especially with GPL, since it's likely that your dependency's dependency will make it GPL as well–but you should be doing this anyways, of course.

I don't like the idea of other code mandating how I should release my code. What if I just don't feel like up-streaming?

Then you're just not in agreement with how GPL works. It's meant to be an explicit guard against people taking projects, forking them, and never contributing back.

dvt4• 6 years ago

Then you're just not in agreement with how GPL works. It's meant to be an explicit guard against people taking projects, forking them, and never contributing back.

I'm definitely not in agreement with it, I just don't like that it tries to present itself as some kind of purist "free software" license, but in reality it mandates how code can be released. How is that free? To take a page out of the blog post, it sure sounds like propaganda to me.

pydry3• 6 years ago

I don't like the idea of other code mandating how I should release my code. What if I just don't feel like up-streaming?

Then don't use the software that's AGPL. Problem solved.

If somebody licences their code with AGPL then it probably means that they either don't want that or they want a cut of any commercial use of the code (dual license) - which is fair, no?

Free is meant to be free

There are two types of free (beer and speech) and it kind of feels like you're deliberately confusing them.

searchableguy3• 6 years ago

I think we should have a standardized MIT license that require paying above a certain market cap or something. An anti-sassfiy license too. That's what AGPL is used for currently but it focuses on contribution and openness rather than monetization.

Many of new database projects and infrastructure tools come up with their own licenses to protect themselves from such problems. A standardized approach would work better for everyone.

ddevault3• 6 years ago

Beware: the approach you describe qualifies as neither "open source" nor "free software". It might be worthwhile to develop such licenses and experiment with this distribution model - but please call it something else, like "source available" or "fair source".

tonyg• 6 years ago

Perhaps you're getting downvoted because it's 2020, and your analysis makes zero mention of what some would call the most important group of people involved in any software project: the end users. There's little left to debate - this particular dead horse is well kicked - but your comment omits any sign you're willing to acknowledge the "freedoms for whom?" aspect of the issue.

dvt4• 6 years ago

Perhaps you're getting downvoted because it's 2020, and your analysis makes zero mention of what some would call the most important group of people involved in any software project: the end users.

This is a fair point. I guess I think the protected class in the context of open source should be the developers -- because they're the ones putting in the actual work. So I'd always favor their freedom over the end users'. Even if the "developer" might be a Fortune 500 company.

Plenty of MIT-sourced projects blew up because Google or Facebook started using them. And, I'm no fan of huge Fortune 500 companies, but let's be fair: they often contribute back.

plus• 6 years ago

It sounds like you are arguing against a different point than Drew is making here. He is saying that the AGPL is no more viral than the GPL, and that its added restrictions aren't particularly onerous. You seem to be arguing against the whole GPL family of licenses, which is a perfectly valid opinion to have, but doesn't really speak to what Drew is talking about.

This is true, and I don't see much wrong with it. I don't like the idea of other code mandating how I should release my code.

I would suggest taking another look at Drew's argument for the AGPL:

The reason they spread these misconceptions is straightforward: they want to discourage people from using the AGPL, because they cannot productize such software effectively. Google wants to be able to incorporate FOSS software into their products and sell it to users without the obligation to release their derivative works. Google is an Internet company, and they offer Internet services. The original GPL doesn’t threaten their scheme because their software is accessed over the Internet, not distributed to end-users directly.

Whether or not this is true of Google in particular, it is true that if you are a company writing open-source software that is accessed over the internet, the AGPL makes it more difficult for competitors to benefit from your work while keeping their improvements secret. If it was GPL-licensed, competitors could use it however they want without releasing their modified source so long as the software isn't "distributed".

dvt4• 6 years ago

It sounds like you are arguing against a different point than Drew is making here. He is saying that the AGPL is no more viral than the GPL, and that its added restrictions aren't particularly onerous. You seem to be arguing against the whole GPL family of licenses, which is a perfectly valid opinion to have, but doesn't really speak to what Drew is talking about.

Maybe I misunderstood his post, but he specifically compares AGPL to MIT in the last paragraph, which actually motivated me to write my reply, as the difference between the two licences is night and day.

ddevault3• 6 years ago

This is true, and I don't see much wrong with it. I don't like the idea of other code mandating how I should release my code. What if I just don't feel like up-streaming? Free is meant to be free, and (A)GPL does incur a cost -- at the bare minimum an ideological one.

You don't need to go upstream - you just need to provide source to users. I explained this in the article, and it's not especially onerous. You've also fallen for a common misconception: the "free" in free software is not about cost, it's about "freedom". The GPL family of licenses are designed to ensure the freedom of the user, not the developer, to view and modify and improve and redistribute the source code for programs that they use.

tptacek7• 6 years ago

You write this as if the whole debate wasn't about what the definition of the "source" that needed to be provided was. The concern is that the AGPL is written in a way that admits to an expansive definition of that source.

detaro• 6 years ago

That's fine, everyone knows that many people don't like the (A)GPL. Not sure we need to have that discussion at every mention of it though? (I assume the latter is the source of at least some of the downvotes)

sokoloff• 6 years ago

I have significant responsibility for setting open-source policy at my company. We require specific review on a facts-and-circumstances basis of any AGPL usage.

Why? Because for most open-source licenses, it's clear how to avoid any license controversy entirely by simply never conveying the licensed work. If I don't ship you ("convey to you") any binaries, I have no obligations to you under GPL (or MIT or Apache, obviously). For the most part, we don't ship anyone outside the company binaries, so compliance (both in fact and in provability of that fact) is relatively easy. Not so with AGPL.

If there was critically useful code only licensed under AGPL, we'd consider it, of course. But it would get a lot more review than a typical Apache/MIT/GPL license situation. I have no philosophical objection to AGPL; I just find it a much less practical license than most of the other, much more commonly used, licenses.

woofie1112• 6 years ago

It's neither more nor less practical in general. It's less practical for your company.

It depends on what you're trying to accomplish. If your goal is to build a low-friction library for others to adopt (e.g. JQuery), it's a horrible, horrible license. If you're trying to build an ecosystem around an end-user tool (e.g. I'm making a web-based competitor to Adobe Illustrator), it can be an awesome license. Lots of open source tools aren't intended to be integrated into other systems, but used stand-alone. That's where licenses like the AGPL really shine.

q3k• 6 years ago

Yeah, making some of my recent projects AGPL truly scares away contributions from Googlers (as their internal, low friction patching process doesn't apply to AGPL software, and they have to go through explicit approval). Fun.

kelnos• 6 years ago

Agree that some of the examples cited are misrepresentations of what the AGPL requires, but I think, for many companies, just the requirement to publish changes (if any) would be a showstopper. Which is unfortunate, but I think is the reality of the situation.

jsnell• 6 years ago

Google states that if, for example, Google Maps used PostGIS as its data store, and PostGIS used the AGPL, Google would be required to release the Google Maps code.

I can’t see that statement, or anything like it, in the linked article. Am I missing something?

Doctor_Fegg• 6 years ago

You're not. The author writes "Google states" and then follows it with his own example/interpretation of Google's policy. That's really not "Google states" at all, it's "My example of Google's policy is...".

The nearest statement in TFA is the sentence beginning "Because Google’s core products are services that users interact with over a remote network interface (Search, Gmail, Maps, YouTube)...".

xyzzy_plugh• 6 years ago

Obligatory: I’m not a lawyer; this is for informational purposes only.

and

Google states that if, for example, Google Maps used PostGIS as its data store, and PostGIS used the AGPL, Google would be required to release the Google Maps code. This is not true. They would be required to release their PostGIS patches in this situation. AGPL does not extend the GPL in that it makes the Internet count as a form of linking which creates a derivative work, as Google implies, but rather that it makes anyone who uses the the software via the Internet entitled to its source code.

This is the problem: I've fought with IP lawyers we've had on retainer who always boil it down to: "It doesn't matter. We're not going to court to find out."

Is it FUD? Yes. But I am also not a lawyer, and I can only do so much to fight upwards as an IC. I've seen this exact mindset at every company I've ever worked for or with.

The real fear is that once you have an AGPL dependency, it's _possible_ that it makes it way into something you really don't want to release publicly, and now you have a huge mess of a problem. The easy solution: don't use software with that license.

(FWIW I hate this, and fully believe competent developers can not shoot themselves in the foot so extravagantly, but alas.)

woofie1112• 6 years ago

AGPL software isn't designed to be used as a dependency to a proprietary system. It's designed to be used in an open ecosystem. That's kind of the point.

* If you expect your code to be used as a dependency to a proprietary system, don't use the AGPL.

* If you are building a stand-alone system, or tools for an open ecosystem, use the AGPL.

That's a question of suitability-to-purpose. Although this use is permitted, AGPL simply isn't a good license for PostGIS (which is correctly licensed under the GPL). It is a great license for end-user applications. It's also nice for dual-licensing.

Licenses aren't something to get ideological about; it's more of a right-tool-for-the-right-job kind of thing. If you use a hammer with screws, you'll get hurt. Same thing with using BSD where AGPL fits, or AGPL where BSD fits.

SpicyLemonZest5• 6 years ago

I agree with what you're saying, but it contradicts the source article, which insists that Google is spreading FUD by saying it can't use AGPL dependencies in its proprietary systems.

cic485• 6 years ago

I’ve taken AGPL through two FAANG reviews. Both arrived at the same very-much-not-FUD legal conclusion.

Paragraph 1 of section 13 requires modifications to be disclosed and source code for them to be offered to remote users. The license uses the term of art Corresponding Source for this.

Corresponding Source is defined in section 1 in a crystal clear way. Two separate teams of lawyers concluded that they could coherently argue the Corresponding Source definition implied not only the modified AGPL software, but also stuff that merely _uses_ it, on the basis that “scripts to control” among other things implies the infrastructure most shops build around software, such as Borg configuration and possibly by extension Borg. After all, a modified version of PostGIS is only useful to run in context, and Corresponding Source requires the context.

AGPL is unchallenged in court. The risk to being wrong about it as huge. It’s risk aversion, not ideology, and it’s important to remember that identifying an argument as part of legal review does not call it the correct one. Anyone who’s ever worked with legal matters knows there is no such thing as “correct,” there are rulings. The existence of the argument condemns the license for FAANG, not its validity. Testing that validity against a claim is perilous.

Perhaps if random engineers stopped calling legal opinion FUD and falsehoods and took a moment to listen to the feedback from lawyers who didn’t write the license, we’d get somewhere with finding a palatable license for all parties. Instead, we get a holy war.

Strong disagree on developers not shooting themselves in the foot. If it’s on GitHub, it ships. If you think about licenses in your day to day engineering, you are the 1%.

an_opabinia• 6 years ago

Yes but on the other hand:

In truth, Google was never going to buy your software.

So what difference does it make? Stop making stuff for Google for free if they won’t contribute back. It’s that simple.

foobiekr3• 6 years ago

I have had the same experience and the same feedback.

It’s much, much, much worse in the embedded world. I think the author is being quite disingenuous on his take.

jcheng2• 6 years ago

I agree that random engineers should have the humility to listen to feedback from lawyers. And IANAL.

Here's the sentence with "scripts to control":

The "Corresponding Source" for a work in object code form means all the source code needed to generate, install, and (for an executable work) run the object code and to modify the work, including scripts to control those activities.

That being said, it feels like a stretch to me that the particular orchestration infrastructure your company uses could be considered part of this. Borg is clearly not "needed to... run the object code and to modify the work" (PostGIS), which you could easily demonstrate by running and modifying PostGIS without Borg--as every other user of PostGIS does.

pydry3• 6 years ago

The holy war could be avoided if Google simply paid authors of AGPL code they wanted to use instead of going on a tirade against the license.

I think half the reason it exists is to make it deliberately risky for FAANGs etc. because they're exactly the ones who SHOULD be ponying up to support the open source ecosystem they rely upon.

belorn3• 6 years ago

Under the same meeting ask the lawyer about the linux kernel and where the border is for its license. They will likely give a similar answer, through if the company relies on selling devices with a linux kernel the risk vs reward will result in a very different decision in the end.

debiandev3• 6 years ago

AGPL is unchallenged in court. The risk to being wrong about it as huge. It’s risk aversion, not ideology, and it’s important to remember that identifying an argument as part of legal review does not call it the correct one. Anyone who’s ever worked with legal matters knows there is no such thing as “correct,” there are rulings. The existence of the argument condemns the license for FAANG, not its validity.

Having worked with lawyers, this largely overstates the risk. Companies are happy to discuss, modify and sign new contracts every day.

All these contracts are "unchallenged in court", by definition, because they are entirely custom.

A lot of software licensing contracts for closed source have complex and restrictive clauses to prevent "renting" such software through SaaS or weakening limitations using legal loopholes.

Yet companies still sign such contracts.

Another type of custom and complex contract is employment.

Furthermore, companies sue each other every other day over contract violation around IP, copyright, trademarks, patents but also employment contracts, rent, building regulations, shipment delays, all of that.

The idea that a FLOSS license is some scary monster is propaganda.

The goal of such propaganda is to drive the FLOSS community to provide valuable software for free and with zero strings attached - aka free labor.

waisbrot2• 6 years ago

By this reasoning, if I run some AGPL software on GCP, wouldn't that place all of GCP under AGPL? I've written "scripts to control" the AGPL software, so my stuff is AGPL and Google's cloud offering is to run my code automatically so all of GCP is a "script to control".

ran38246924• 6 years ago

Corresponding source has the exact same definition in GPLv3 and almost exactly the same in GPLv2, so all this "its completely untested" thing is completely disingenuous. Google uses Borg to control gplv3 code that they also distribute, so, exactly the same case and its complete BS you are spreading. Lawyers are actually pretty good at spreading FUD about GPL, they always have been.

vslira• 6 years ago

AGPL is unchallenged in court

What would be the cheapest way to settle this? Can I found a SaaS company that uses AGPL code, grow to ~100 usd/month and basically ask the FSF to sue me? Even if the legal fees alone amount to tens of thousand dollars, that should be low enough that a small company or a motivated (and financially secure) individual could try.

tyrion• 6 years ago

Wouldn't it be a good idea to "fix" the AGPL in order to make it clearer?

It seems to me that every time there is a discussion around GNU licenses many people have no idea how to interpret them and then decide not to use them out of fear. It does not seem to me that this is in the interest of the GNU project. Why not fix this?

rmc• 6 years ago

“The company might have to release the source for everything!” isn't a downside for some people. The FSF's philosphy is that all source code should be opened. What you're describing as a bug (for a company), is (I think) a feature for the FSF.

asdgasdg• 6 years ago

This is the problem: I've fought with IP lawyers we've had on retainer who always boil it down to: "It doesn't matter. We're not going to court to find out."

That's not really an argument, the lawyers are just pulling rank on you.

api2• 6 years ago

I've seen the same phobia to a lesser degree around the plain old GPL. Lawyers think, and not unreasonably, in terms of risk. Is there a risk you might be dragged into court? If so, that's a very expensive risk. Is the library or application worth this risk? If not, then ditch it.

It's not an unreasonable fear either given past events like the SCO/Linux lawsuit. IBM had deep enough pockets to fight it and win. Anyone else would have folded or been ruined.

It makes it hard to be anything but 100% "libre" (almost public domain) or commercial and closed source.

Being 100% libre means someone big can (and sometimes does) just take your work and put their name on it and monetize it. This wasn't much of a problem years ago, but it's a growing issue now. Being 100% commercial is of course the opposite. We're excluding the middle, to great detriment, and we really need a workable legal solution that won't scare lawyers.

zrm• 6 years ago

But isn't the middle ground requiring copyright assignment and then dual licensing?

Oh dear, your company's lawyers are afraid of the AGPL? Well then pay the project for a commercial license.

nine_k3• 6 years ago

I agree with this analysis, and I think that dual licensing can help here. That is, something is free is you want to use it in a copyleft-compatible fashion, and costs serious money for a commercial license if you want to go proprietary.

Even SQLite, which is public domain, offers a commercial "warranty of title" [1] if your company requires a guarantee that there can't be a lawsuit regarding the use of this software.

[1] https://www.hwaci.com/cgi-bin/license-step1

guerby• 6 years ago

Lawyers run businesses, and drafting custom contracts/licence are a big part of their business.

GPL and standardized licences destroy this part of their business.

Also who pays a lawyers to read the thousands of page of licences for say Microsoft software (all the version of all the software deployed)? Where I work no one is able to know what Microsoft licence we need to buy, even when we ask licence sellers they mostly don't answer, and it changes all the time.

IIRC early Microsoft Visual C++ licences did not authorize the diffusion of debug compiled executables (don't know if it's still the case). How many violation of this?

bsder• 6 years ago

Embedded is generally quite serious about this.

It's not even whether you comply, the issue is that if your company is getting bought you do NOT want the deal to get hung up over software licenses.

skissane• 6 years ago

This is the problem: I've fought with IP lawyers we've had on retainer who always boil it down to:

And I think most engineers would rather be doing other things than arguing with their employer's legal team. Their legal opinions may or may not be correct, but it doesn't really matter either way, because you are unlikely to change their mind. If Legal says "AGPL is not allowed", you don't waste your time arguing with them or begging for an exception, you just go look for an alternative component with a license they'll accept.

(If, hypothetically, one's employer's legal team were saying things completely outside the norm of the industry, it might be worth making an issue of it – but, banning AGPL is certainly well within the norm of the industry.)

jpalomaki• 6 years ago

Back-in-the-days MySQL had the view of GPL that if your application strictly required MySQL database, the GPL license would extend to it. Even if the app would only connected to the database over network.

”A license is required if [...] You have a commercial application that ONLY works with MySQL and ships the application with the MySQL server. This is because we view this as linking even if it is done over the network.”[1]

Not sure if they ever actually enforced this rule.

I wrote this just to point out that different companies can view these licenses differently. If company chooses AGPL, they clearly want to block certain use cases. I would be cautious when trying to dance around the limitations.

[1] http://ww2.cs.lamar.edu/doc/mysql/manual_Licensing_and_Suppo...

marcinzm3• 6 years ago

Google states that if, for example, Google Maps used PostGIS as its data store, and PostGIS used the AGPL, Google would be required to release the Google Maps code. This is not true. They would be required to release their PostGIS patches in this situation. AGPL does not extend the GPL in that it makes the Internet count as a form of linking which creates a derivative work, as Google implies, but rather that it makes anyone who uses the the software via the Internet entitled to its source code.

Is there an existing national court case which supports this assertion? Otherwise it's a valid concern given that lawsuits and judges may have different interpretations than engineers. Just look at the Oracle vs. Google debacle over Java.

woofie1112• 6 years ago

Eben Moglen wrote the AGPL. There's a standard process courts use to figure this stuff out. Intent fits into it. Eben Moglen clearly states what he meant in many talks. Courts favor intended interpretation of a contract. That makes the risk pretty small.

SpicyLemonZest5• 6 years ago

I don't think this really supports your point? Eben Moglen has indeed clearly stated what he meant; he thinks free software ideals are very important and tried to write GPL-family licenses so that developers would be compelled to respect them.

marcinzm3• 6 years ago

Intent of whom? Eben Moglen is not the owner of the copyright but merely the writer of the license. If Oracle, for example, uses AGPL would it not be their intent in using the license which actually matters?

tzs4• 6 years ago

In the case of ambiguity in a contract, courts look at the intent of the parties to the contract. In almost all cases Moglen will not be a party, and at least one of the actual parties will not be aware of those talks or their content, so Moglen's talks on the matter won't really be of much use.

Courts will also look at the bargaining power of the parties. If the party offering the contract is a lot more powerful than the party accepting the contract, they will tend to favor the accepting party when it comes to ambiguous terms.

michaelcampbell• 6 years ago

Truth or falsehoods aside, the reason one of the places I have worked (large 10's of k's of employees, big legal staff) refused to let us use AGPL is that it had never been decided in court, and they didn't want to be the ones to foot that bill.

"No" is pretty cheap, and they were pretty good at it.

saagarjha13• 6 years ago

"No" is pretty cheap

Now, is it really? Paying a team of software engineers to recreate every AGPL project surely isn't…

evgen• 6 years ago

But using a non-AGPL project is a cheap and easy solution for just about every use-case out there. If there was something special that was AGPL licensed it might matter, but there isn't so it is easy to avoid.

DasIch• 6 years ago

Nobody is using every AGPL project and there aren't that many AGPL projects to begin with.

The problem is not just the risk associated with AGPL, it's also that the benefit is actually quite small. I think most companies would be a lot more willing to use AGPL, if there were any compelling projects using it.

hyperpape2• 6 years ago

"The Google page about the AGPL details inaccurate (but common1) misconceptions about the obligations of the AGPL that don’t follow from the text."

"The reason they spread these misconceptions is straightforward: they want to discourage people from using the AGPL, because they cannot productize such software effectively."

"Ask yourself: why is documentation of internal-facing decisions like what software licenses to use being published in a public place?"

To be perfectly clear, your claim is:

- Google internally knows they can comply with the AGPL without trouble

- Has consciously chosen to write policies prohibiting it based on misleading reasons

- Google adheres to that policy

- Most importantly: they are doing this because they believe that the influence this would have in discouraging use of the AGPL is valuable enough to substantially benefit Google

Have I understood correctly?

catern• 6 years ago

I don't know exactly what the author of this post means, but it doesn't have be a "conscious" decision on the part of Google to spread misleading information for their own benefit. It's very easy (I contend) for humans to believe things that are poorly supported by the evidence but which are beneficial for them.

So even if it's obvious that the AGPL doesn't have the effects that Google's policy suggests it does, it doesn't require active deception on the part of Google for them to believe it does. It just requires that they be deluding themselves for their own benefit - which I think is a very common practice for humans.

hyperpape2• 6 years ago

This is a good point. My two thoughts are:

1. I don't find it easy to read the original post as implying Google is honestly convinced of its viewpoint.

2. It's a huge difference in meaning. Assume for the sake of argument that Drew is right about the interpretation of AGPL.

How should we interpret Google's actions? If they're paranoid, and afraid that the AGPL might require them to open-source their entire code base, that's unfortunate. You might still call them greedy assholes making open source worse, but it is a real fear that most commercial entities would be concerned about if it's realistic.

But if they knew better, and were lying, that's a completely different level of shitty behavior. If there's any ambiguity about that in Drew's post, it ought to be resolved.

waisbrot2• 6 years ago

Not the article author, but:

- Google internally knows they can comply with the AGPL without trouble

No, and I don't know where you got that. The article is talking about how start-ups can typically comply with AGPL without trouble. cic48 claims that all software running on Borg needs to be extensively modified and then can only be compiled using Google's proprietary tool-chain. If that's true, then they can't run AGPL on Borg without releasing the code for the tool-chain.

- Has consciously chosen to write policies prohibiting it based on misleading reasons

No. Has written policies prohibiting it for rational self-interest (GPL, MIT, etc licenses allow them to sell access via the Internet without any cost to them, AGPL does not). After writing these policies, they've provided an alternative justification.

- Google adheres to that policy

Outside of Google Legal, who would know whether or not they adhere to their own internal policies? If they used AGPL software internally, they would have no legal obligation to disclose that (because they'd be fully in compliance).

- Most importantly: they are doing this because they believe that the influence this would have in discouraging use of the AGPL is valuable enough to substantially benefit Google

Why "substantially"? I would imagine the cost of saying "don't use AGPL" and then publishing an article about it was very low. A couple person-weeks of work at the most? It's hard to imagine what wouldn't be worth that investment.

But I think you didn't understand the article, because the main thrust of it was to encourage businesses to not be afraid of the AGPL, not to determine whether Google is acting maliciously or not.

jgilias2• 6 years ago

So. From the POW of a total lay person as far as it comes to law.

Someone writes a blog post with an 'IANAL' disclaimer on top saying that what Google's army of lawyers have gathered from reading a legal document is false, and I should favor his interpretation instead. I don't know, I'm not exactly convinced.

simion3144• 6 years ago

Someone writes a blog post with an 'IANAL' disclaimer on top saying that what Google's army of lawyers have gathered from reading a legal document is false, and I should favor his interpretation instead. I don't know, I'm not exactly convinced.

Microsoft with their giant army of lawyers also said GPL is a cancer and they promoted this idea a lot but today Microsoft "loves" GPL. I think you can conclude that you should use your own brain to decide and not let a giant company decide for you, they might have a huge financial interest to promote a certain narrative. It is clear in this case that Google would prefer their employees won't work on their free time on AGPL code.

knorker11• 6 years ago

you should use your own brain to decide

I think you mean "your own lawyers". And my experience says that the AGPL opinion among these experts are about as one-sided as scientists and climate change.

But you're suggesting to armchair lawyer it?

jgilias2• 6 years ago

I was not actually talking about the totality of Google's verdict about AGPL. For some projects in some companies it might just be the right license to use. Rather, I was talking about the particular example of Google Maps and PostGIS.

Google's army of lawyers say: If PostGIS 'this' => Google Maps must be 'this'. And the author says, 'no, that's false'. Someone must be wrong here. And if I had to place a bet on which side is more likely to have gotten it the way a court of law might interpret it, I'd put my bet on a bunch of people with law degrees.

alex_young• 6 years ago

Companies like MongoDB release a 'free' version using AGPL and a commercial version under no such provision, and use this in marketing material to convince commercial users to buy licensing so they can incorporate their DB into web based products.

If this distinction is without merit for those simply using MongoDB as an unmodified DB, this seems like it wouldn't actually work so well as a sales tool.

mr_toad2• 6 years ago

Current versions of MongoDB are not available under an AGPL at all. It is effectively no longer open source.

Older, (unpatched) versions are still AGPL, but there were no commercial restrictions at that stage.

xtracto• 6 years ago

The current free license has this interesting snippet:

13. Offering the Program as a Service. If you make the functionality of the Program or a modified version available to third parties as a service, you must make the Service Source Code available via network download to everyone at no charge, under the terms of this License. Making the functionality of the Program or modified version available to third parties as a service includes, without limitation, enabling third parties to interact with the functionality of the Program or modified version remotely through a computer network, offering a service the value of which entirely or primarily derives from the value of the Program or modified version, or offering a service that accomplishes for users the primary purpose of the Program or modified version.

“Service Source Code” means the Corresponding Source for the Program or the modified version, and the Corresponding Source for all programs that you use to make the Program or modified version available as a service, including, without limitation, management software, user interfaces, application program interfaces, automation software, monitoring software, backup software, storage software and hosting software, all such that a user could run an instance of the service using the Service Source Code you make available.

I know a bunch of startups using Mongo... i wonder if i could make them give me their systems source, given this clause.

amadeuspagel• 6 years ago

Ask yourself: why is documentation of internal-facing decisions like what software licenses to use being published in a public place? The answer is straightforward: to influence the public. This is propaganda.

Or maybe it's just something that many people have asked about? And why wouldn't they publish it?

jeffbee7• 6 years ago

The blogger is reading way too much into the existence of this page. The site is clearly just a wholesale export of internal Google documentation. It is not generally intended for the public audience and I don't think you can read intent into specific pages being here. Look at this page for example. Nothing on it is relevant to you outside Google. You can't reach "go/" pages, you can't check out //third_party from their repo, etc.

https://opensource.google/docs/thirdparty/r/

dathinab7• 6 years ago

My problem with GPL is that the term "derived" is defined to vague or not in the way I like it, which in turn opens up all kind of legal uncertainty.

(I don't know if this also applies for AGPL.)

I would love to hear that AGPL legal reliably works like he describes (or there is a different license which does).

yarrel• 6 years ago

None of the V3 licenses say "derived".

Derivation is a specific concept in US copyright law, so in fact it doesn't need to be derived in the license. But it also doesn't apply elsewhere.

So the V3 licenses use the term "modify", which they define to specifically include the more internationally-understood concept of adaptation.

Where things get more complex is in each license's concept of linking. But that's only a problem if you want to be a free rider. Like Apple or Google.

dathinab7• 6 years ago

GPLv2 does use that wording.

It's still widely used.

AGPL is also originally based on GPLv2.

Through there is the GNU AGPL license which is often abbreviated as AGPLv3 which is based on GPLv3.

GPLv3 makes many things more clear but not necessary better. It also includes some cases I would explicitly not include as "derived work" (or more explicit wok which will need to be GPLv3 compatible as yes the terminology changed).

But that's only a problem if you want to be a free rider. Like Apple or Google.

No, it's a problem for many other use-cases which have nothing to do with free riding but protection of trade secrets and similar.

If you want to make sure your software is FOSS sure go ahead, but why force all the software directly around it to be FOSS, too?

I mean it makes sense for complete services and similar (like a database or the Linux kernel). But it IMHO is not very good for any kind of library.

In my opinion forceful open source is good for modifications and extensions of a given service (in the very general CS definition of a service which includes most kinds of libraries). But should be avoid to extend to any user of a service independent of the way the service is interacted with (static linkage, dynamic linkage, etc.) with an additional exception wrt. LTO.

Sure you can't sell such software well. But most FOSS software isn't meant to be sold. What is sold are products like support, service and maintenance around it and similar. There are very view companies which had success with a selling software based on a GPL vs. proprietary dual license schema as far as I know. I think, and might as well be wrong about it, that it's better to then have a license like mentioned above but restrict commercial use in some way with the option to pay to lift the restriction.

SpicyLemonZest5• 6 years ago

I think the author is misreading Google's policy (https://opensource.google/docs/using/agpl-policy/). They're not saying that communications over the internet constitute a linkage, they're saying that:

* Because of free and open development practices within Google, it's impractical for developers to be constantly aware of what libraries get linked into their binaries. (This is exacerbated because of their monorepo, but even in a multi-repo setup it's very easy to end up with random libraries in your dependency chain.)

* If an AGPL library did end up linked, because all Google's services are accessible through a remote network, the virality provisions kick in and they might have a legal duty to distribute the entire binary.

* Therefore, they have to ban AGPL libraries so that no core service developer accidentally depends on one.

p0llard4• 6 years ago

Google states that if, for example, Google Maps used PostGIS as its data store, and PostGIS used the AGPL, Google would be required to release the Google Maps code. This is not true. They would be required to release their PostGIS patches in this situation. AGPL does not extend the GPL in that it makes the Internet count as a form of linking which creates a derivative work, as Google implies, but rather that it makes anyone who uses the the software via the Internet entitled to its source code.

I don't really follow this.

The fact that the linking exception and the LGPL exist at all is enough to infer that the FSF/license authors consider linking against or otherwise using a piece of software as a component of a larger system is enough to make that larger system a derivative work of the smaller component, thus "tainting" (I use this work non-disparagingly) it with the copyleft provisions of the license.

If Google Maps (that is the suite of software running on Google's servers) uses PostGIS as its data store then it seems that without a judicial ruling on whether there's a legal difference between interfacing with a software component by linking against it and interfacing with a software component by using some other software interface, it's not possible to be so certain that there isn't an issue here.

knorker11• 6 years ago

Obligatory: I’m not a lawyer; this is for informational purposes only

That's the main point, though. You aren't, in fact, a lawyer. And this is not, in fact legal advice. You are presumably expert in a non-legal field, and you are giving expert opinion on something you are not an expert on.

The GPL is tested. The LGPL less so, but lawyers seem to be more comfortable with it.

There's the extra complexity that Google compiles its binaries statically.

You're calling Google liars. There's an alternative interpretation of events where a whole legal department, with great lawyers, and backed by great engineers to clarify the technical aspects for the lawyers, come to a different conclusion than yours.

And you just dismiss that as lies. I don't think that's fair at all. You wouldn't want a lawyer to come and say "bah, you should just make that inherently NP problem complete fast for all inputs. How hard could it be? If you say it's hard then you're lying.".

this is for informational purposes only

This is pretty arrogant. "I'm not a lawyer, but here are the real legal facts to ACTUALLY educate you".

You should say it's "for speculation purposes only", or "for entertainment purposes only".

Don’t be afraid to use the AGPL, and don’t be afraid to use software which uses the AGPL.

This is ignoring one big problem though. Agree or not, call them liars or not, but Google and it's employees will NOT touch your software. Not only will you not get Google as users (though if you dual-license yes in fact Google DOES buy software, if they can buy it as non-AGPL), you will not get Google employees as contributors.

saagarjha13• 6 years ago

You're calling Google liars. There's an alternative interpretation of events where a whole legal department, with great lawyers, and backed by great engineers to clarify the technical aspects for the lawyers, come to a different conclusion than yours.

Google claims that they care about your privacy. Entire teams of engineers and lawyers will say the same thing. But when you look at it, you can plainly see that they are probably just saying that because their main business is collecting data about you. Similarly, Google runs many projects with a closed-source server-side component, and using the AGPL would not let them do that. Do you see why it is not surprising they would come to a different conclusion?

knorker11• 6 years ago

So your legal interpretation as a lawyer (I assume) is an ad hominem on Google for unrelated reasons?

rando8322• 6 years ago

And Microsoft said all GPL was a cancer and unamerican and viral and disallowed it from about 50k engineer's computers due to "legal risk of GPL virality", and this was when they were by far the biggest software company. Wait, they lied? It was all FUD and really about threatening their business model. They changed their policies and actually said publicly they were wrong. Is it so hard to believe?

knorker11• 6 years ago

Right. So because MS used the wrong strategy that is evidence that Google's lawyers are deriving incorrect legal conclusions?

Is that your argument?

So because MS was wrong once, Google is wrong other times of your choosing?

n0n0n4t0r2• 6 years ago

Hello, While I'm employed to develop an agpl software, and I fond of this license, it's clear that with the wrong actors it can be a threat to some businesses.

I'll tell you a little story that happened around 10 years ago:

I got a call from a representative of Oracle, he asked me if we where using MySQL, and if I could described him how, because he wanted to help us make Better use of this tool.

We where pretty happy about MySQL at the time, and I went into deep details about how we used it. At the end, he told me point blank that the PHP's MySQL driver was licenced under the GPL and that we had to licence our whole codebase under the GPL since it was contaminating our code as a whole. (Even if we had encapsulated all the accesses to the driver around a single class)

The alternative was to pay the right to use it under a non GPL contaminating way.

Oracle then called and threatened us many times.

The argument that made them stop was when we told us that we where hosting the applications. This argument would not have been sufficient with the agpl.

There claim was unfonded but I can assure you that I didn't sleep well for a while!

djhaskin987• 6 years ago

Your story is a good reminder that these license questions aren't just about some holy war or ideology campaign. These are real people with real problems. We need to be more sensitive about how to structure licenses such as the AGPL.

I find I don't want even to use the GPL at times. Why? Not because I want to give away free stuff to Google. Rather, because I don't want to force others to use the same license that I use. You can use MIT code from a GPL code base, for example. I want to be nice. If Google uses it without paying me, that's fine, too. Honestly I'd probably feel pretty proud of myself at that point.

n0n0n4t0r2• 6 years ago

Yes and no: I currently work for the editor of iTop (an open source ITSM software).

Providing the software under an open source licence was a very early choice. But what if a rogue concurrent decided to fork and hire engineers to provide it's own, better version of the soft while keeping it closed source?

The only open source licence providing protection that I'm aware of is the AGPL.

When you throw on the project your lifetime savings, it is quite a reassuring licence.

Andrew_nenakhov2• 6 years ago

Stories like this are the reason I tell everyone to never touch MySQL and anything related to Oracle. If you need a relational DB, look no further than PostgreSQL.

bob1029• 6 years ago

Alternatively, if arbitrary, highly-litigious vendors start probing your private business activities, politely roll them to voicemail and your spam folder.

Unless you are already under contract or other legal order to do so, you do not have to tell anyone anything regarding the nature of your business. Every piece of information you give to one of those assholes is something that will be used against you in a lawsuit. Don't make things harder on yourself.

em-bee• 6 years ago

or MariaDB. or is oracle also going after MariaDB users, since they own a good amount of its code?

st1ck• 6 years ago

The argument that made them stop was when we told us that we where hosting the applications.

I didn't get that, can you please rephrase?

labcomputer• 6 years ago

The GPL only applies when you distribute software. The person you're distributing it to can request (and you must provide) source for GPL-derived works.

A service provider who hosts applications does not distribute software. They use it internally to provide the service. Thus, they do not trigger the distribution part of the GPL.

So Oracle can't say "you have to distribute source to your GPL-derived work or pay us lots of money" because you didn't distribute binaries for your GPL-derived work.

tzs4• 6 years ago

Any users of such software are entitled to the source code under the terms of the AGPL, including users accessing it over the network such as with their web browser or via an API or internet protocol

That doesn't seem to be quite right. It makes it sound like the source code entitlement applies to all users. The "including users accessing it over the network" is redundant then because they are a subset of all users.

What's actually in the text of the license is:

Notwithstanding any other provision of this License, if you modify the Program, your modified version must prominently offer all users interacting with it remotely through a computer network (if your version supports such interaction) an opportunity to receive the Corresponding Source of your version by providing access to the Corresponding Source from a network server at no charge, through some standard or customary means of facilitating copying of software.

The user source entitlement only applies to users remotely interacting with the software through a computer network.

saagarjha13• 6 years ago

Surely the user source entitlement also applies if I interact with the software without a network in between? Otherwise, that would make AGPL a strangely more-but-also-less permissive license than GPL…

tzs4• 6 years ago

It is almost literally GPLv3 with the interactive remote network source obligation added in.

If you diff GPLv3 and AGPLv3, the differences are (ignoring differences that are just in the name of the license):

1. The preamble has a few differences where they describe why they wrote each license and what ills they are trying to address.

2. In the actual terms and conditions, the first 13 sections and last 4 sections are identical (except for the name of the license). Only section 13 is different.

In AGPLv3, 13 is the section about remote network user source access, and about how AGPLv3 and GPLv3 interact.

In GPLv3, 13 is about how GPLv3 and AGPLv3 interact.

Finally, after the T&Cs is a section on how to apply the license to your programs. That has differences, such as including in the AGPLv3 version a reminder that you have to handle your section 13 obligations for remote network users.

e12e2• 6 years ago

That doesn't seem to be quite right. It makes it sound like the source code entitlement applies to all users. The "including users accessing it over the network" is redundant then because they are a subset of all users.

The AGPL is written in the context of the gpl, where user is someone who runs the software. The AGPL extends this to someone who uses the software over the network.

You're right that that kind of means "all users" - but let's take the example of a GPS appliance - if you buy the device, and run software on in, you're a "gpl user". If the appliance is a server that runs a web application - the person operating the appliance is a "gpl user", the one interacting with it via the web is not - but they are an "AGPL user". If you relay directions to someone (say as a passenger/navigator in a car) - you could argue the driver also "uses" the software - but only the passenger would be considered an "AGPL user".

All this must be seen in light of the four freedoms, and who it applies to (who is a user).

The AGPL extends freedom zero (the freedom to run the software) - to those accessing it over the network - in short, it grants the right to self-host and run the code.

victorkab• 6 years ago

Disclaimer - As a CTO of a company I have to take a stance on these issues. As a matter of fact, as you raise money, part of the due diligence is to audit the libraries that you use to make sure that you don't use libraries that can jeopardize the future of the company.

Using an AGPL library for a Saas company will be flagged as a risk by lawyers during the due diligence process.

At Truework, we love open source and try to contribute when we can, but the reality is that if a library is AGPL, I'll ask my team to not use it. It's just not worth it to risk so much for a single library. Yes, it's unlikely that you will have to go to court because you use that one library, but if you do... well that's bad news for you.

If you build a business, it's about making sure that there there is a reasonable ROI for risks that you take. APGL tilts the balance towards "we shouldn't use this, it's too risky". Make your own decisions, but keep that in mind.

marcinzm3• 6 years ago

And if you rely on large companies as clients they will have a line asking about GPL/AGPL library use during contracting for similar risk reasons.

KingOfCoders3• 6 years ago

It's the same when you go through M&A.

hitekker• 6 years ago

The new title is misleading. It is a charitable interpretation of a half-screed/half conspiracy theory.

The reason they spread these misconceptions is straightforward: they want to discourage people from using the AGPL, because they cannot productize such software effectively.

Statements like this can be encompassed by its original title, as the author intended.

mark_l_watson• 6 years ago

I think that the AGPL is a fine language in cases where the code being released is a complete system, that you want many people/organizations to use and improve it, but you don't want anyone cloning it as their own standalone product.

I understand why Google would not want to use it, and also other large companies.

For individuals and smaller companies I think the AGPL can make good sense, especially for complete web applications.

knorker11• 6 years ago

I think that the AGPL is a fine language in cases where the code being released is a complete system, that you want many people/organizations to use and improve it, but you don't want anyone cloning it as their own standalone product.

Yup. In the same way that you could have a license that requires people to publicly criticize <Politician> to use your software.

You'll get some. But most will just not agree to your terms.

emptyparadise• 6 years ago

Big corporations avoiding AGPL like the plague makes it a much better "don't be evil" license than any other attempts at making one of those. If anything, I'd love to see some dual license scheme where a different license is available only to small developers or non-profits, rather rather than corporations with bags of cash. I wonder how that can be enforced - perhaps some sort of a CLA that only gives limited rights to relicense instead of just handing over the copyright entirely.

And yeah, this goes entirely counter to making profit, but hey, I think we're privileged enough to afford to try ideas like this.

labster2• 6 years ago

It reminds me of the story I heard around a campfire once, about a programmer who decided to fix a bug in a single AGPL module, and they were forced to make their entire code open source. Even today, long after the bankruptcy, they say you can still hear the screams in the shuttered, decaying boardroom. And the bug somehow got unfixed. It’s out there now, waiting ... for its next victim.

exhilaration• 6 years ago

Is this a story lawyers tell around the campfire to scare new law school grads?

labster2• 6 years ago

Not quite. That version ends with the line: “and their lawyers ... never got paid.”

svnpenn• 6 years ago

Network use is distribution

I use the Open Software License for this same reason, I dont think companies should be able to get the benefit of open source libraries with the cost of publishing any modifications, regardless if its a webapp or desktop app. I am curious the difference though:

https://choosealicense.com/appendix

seems to only be different with regard to trademarks?

ThePhysicist2• 6 years ago

After much consideration we finally released our core software (https://github.com/kiprotect/kiprotect - a privacy & security engineering toolkit) under the AGPL. In the past we've released other software under more permissive licenses like the BSD-3 or MIT licenses, but we've decidedly picked the AGPL for our toolkit, for the following reasons:

- We want to encourage people and organizations to use the software as a data processing & protection tool, just like they use other Linux/Unix tools.

- Using the software does not require integrating it into a software project as a library (though you can do that).

- We want to ensure that any extensions and modifications of the software (e.g. new anonymization or pseudonymization methods that people would implement) will make it back into the main software as open-source, so that everyone can benefit from them.

- We want to keep potential competitors from being able to just resell our software as a SaaS offering or integrate it into a commercial, closed-source product.

For me, the AGPL fits this bill quite nicely, and that it deters organizations to use the software for free in their closed-source projects is actually a bonus. We offer dual-licensing by the way, so organizations can just buy a commercial license if they want to integrate our tool into their closed-source software. That at least should allow us to refuse such a license to a potential competitor and the funding we (hopefully) will generate via the commercial licenses will help us fund the development for everyone. Win win.

CodesInChaos• 6 years ago

- We want to ensure that any extensions and modifications of the software (will make it back into the main software as open-source, so that everyone can benefit from them.

isn't that a contradiction with

We offer dual-licensing by the way

since dual licensing requires every contributor to agree to a CLA and not merely contributing the code under the terms of the AGPL?

ThePhysicist2• 6 years ago

Maybe, but I don't think so. In general we plan to have every contributor sign a CLA as we want to make sure we can keep control over the development of the project, in case we will have to change its license later.

Some companies don't want to contribute back to open-source software and that's fine, if they pay a license fee instead we can use that money to pay ourselves and build new features for the open-source version, so everyone benefits. I'd even say that maybe open-source software is licensed too liberally these days. Right now for-profit companies make boatloads of money on the back of open-source software, without giving back much. I think if more projects were to use contagious licenses like the (A)GPL it might actually benefit the whole open-source ecosystem. Imagine what we could build if just 1-5 % of the largest companies would pay a small license fees for the open-source libraries they use.

e12e2• 6 years ago

Rather heavily editorialized title? Original says: "The falsehoods of anti-AGPL propaganda"

pgcj_poster• 6 years ago

I release my software under AGPL whenever possible—even if it couldn't conceivably be used to provide services over a network—because it means that Googlers aren't allowed to use it on their work machines. If Google ever changes their policy on AGPL, then I'll either switch to the JSON license or a custom one that will arbitrarily exclude companies and people that I don't like.

enriquto• 6 years ago

I love the AGPL, and I love that it is more and more used by larger and larger projects. For example, the widely used Overleaf. This is a good world.

evmar• 6 years ago

I can't comment on the "true" motivation for releasing these documents, but I noticed the same site includes https://opensource.google/docs/thirdparty/oneversion/ which is a very Google-internal-only kind of problem, complete with broken links to Google-internal-only tools and processes (rosie and lsc). That page at least is definitely just a dump of the internal documentation.

DannyBee2• 6 years ago

I haven't commented on HN in a while, but this article actually pretty much complains about me, since I wrote the text in question and enabled the policy to be released.

So i'm just going to say:

1. The author claims the Google states something, then doesn't actually quote anything google stated, but instead writes their own interpretation of the words and a made up example. That's not a good start.

2. Having set up their own example, the author then proceeds to say it's wrong because of an appeal to authority, asserting no reasonable disagreement is possible - when not only is it possible, but large numbers of IP lawyers disagree on what the AGPL requires. It's true that someone wrote it, and there are FAQ's, but depending when and where, their opinion on what it means is not as relevant as one might think.

In truth, lots of people disagree over interpretation of the AGPL - many more than disagree about GPL interpretation or LGPL interpretation (which are fairly settled at this point). Searches over any legal licensing list in existence will you this.

When I wrote the interpretation you see here, it was the best available info at the time, guided not just by own views, but by listening to a lot of smart lawyers, counterparts at other companies, etc.

Even if you ignore all the lawyers and whatever as useless, the author has the huge problem that There are people who make AGPL software that take the view listed in the google policy

So it's not just Google or lawyers, it's software authors. Not just a few, either.

While companies are generally happy to ignore one or two people whose interpretation is outside the norm (and just not use their software), that's much harder when there is such widespread difference in view of the AGPL.

3. For no particular reason at all, the author then decides to assert a tremendous amount of bad faith in interpretation of the AGPL and reasons for publishing such policies.

We published our policies because over the years I (and others) were repeatedly asked by counterparts at other companies, various communities, and others to understand what our policies look like, for a variety of reasons (to understand for themselves, to use as a template, etc)

Period. That's the whole reason. It even says this: https://opensource.google/docs/

The author here has decided that's clearly all lies or deliberate misinformation, without even bothering to even ask anybody.

jkaplowitz6• 6 years ago

Thanks for having publiished these policies publicly! You gave a pretty clear summary of Google's motivations for banning AGPL on HN five and three years ago two of the previous times this came up, and having worked at Google 2011-2015 (including some interactions there with you), your public explanations of why from back then are extremely credible and rational given the logistics of what's easy and hard inside Google.

https://news.ycombinator.com/item?id=9956542

https://news.ycombinator.com/item?id=13979443

Not that I'm surprised, as you've always been pretty forthright and honest. Thanks for continuing to be that way.

DannyBee2• 6 years ago

Of course. I appreciate the thoughts. I try my best to be transparent and straightforward.

I actually do think it's reasonable to expect companies to re-evaluate every few years (at a minimum) because the world changes. At some point, the reasoning from X years ago will not make sense anymore.

One of the top rules of sane policy making is to write into the policy the conditions that should cause you to re-evaluate it, whether it's time, or change in evidence you relied on, or ...

You want people to understand your rationale and what might change your mind, otherwise you shouldn't expect them to go along with it.

Even though such a rationale exists in various emails, etc, I actually failed to do that in the resulting police here, unfortunately (the internal version is not any better in this respect). That one is on me and I think would be a more legitimate complaint here.

I don't think it would have changed yet (if anything, the AGPL has become more contentious as the economy has gone south), but that's just an at-a-glance view.

raphlinus• 6 years ago

I can confirm what Danny says here.

It's a shame the original article has such flaws in its argumentation, because at the core of it I think there are some good points. Notably, other people don't have to blindly follow Google's example and reasoning. In fact, as discussed here, there are good reasons not to. For one, because Google is so highly integrated (with Borg and its monorepo, not to mention custom approaches to security), it's harder to use AGPL software without risking linking. Most other users outside Google would just apt-get a package or the like. And of course the business risk to Google is large simply because of their scale.

You can even disagree with Google's decision not to use AGPL, but they have openly explained their reasoning, and I think that adds to the discourse.

Disclosure: I used to work at Google, and was fairly actively involved with open-source activities in different capacities while I was there. I am also on record as sharply criticizing Google when I believe that is warranted. This is just not one of those times.

young_unixer• 6 years ago

Even as an individual user, I try to avoid using software licensed under GPL or its derivatives.

I'm against copyright laws, and any second that I have to spend thinking about how to comply with them is a second wasted to me.

BSD and MIT make me waste the minimum practically possible amount of time.

I like knowing that I have the freedom to not care about infringing on anyone's copyright as long as I only develop software using permissive licenses.

jkaplowitz6• 6 years ago

BSD and MIT maximize freedom for developers, which seems to be your primary concern so that's totally understandable. You're right that they're much easier for developers to comply with than the GPL family.

GPL and AGPL maximize freedom for both direct and indirect downstream users, such that users of any software derived from that GPL/AGPL software are guaranteed strong rights to be able to understand / modify / pay others to modify their software. BSD/MIT allow proprietary derivatives which withhold this freedom from users.

Both are pretty valid concerns, even if they're opposed to each other. But for cases where you're a pure user and not a developer, there's no compliance hassle involved with either the BSD/MIT family of licenses or the GPL family.

minieggs• 6 years ago

For fans of AGPL I suggest checking out EUPL. Half the length of AGPL. Changing/removing portions of the license is allowed.

More info: you may have seen this on Hacker News recently: https://www.arp242.net/license.html.

saagarjha13• 6 years ago

Changing/removing portions of the license is allowed.

Note that licenses that can have random bits changed or removed have the issue that you need to read them in their entirety each time you encounter them.

purpleidea• 6 years ago

Great explanation, and I hope people get the message. I've tried to explain this to people and companies before, but Drew does it better.

Thanks!

aazaa• 6 years ago

The author makes the following claims:

In truth, the terms of the AGPL are pretty easy to comply with. The basic obligations of the AGPL which set it apart from other licenses are as follows:
- Any derivative works of AGPL-licensed software must also use the AGPL.
- Any users of such software are entitled to the source code under the terms of the AGPL, including users accessing it over the network such as with their web browser or via an API or internet protocol.

That first claim is debatable at best. Nowhere in the entire document will you find the term "derivative work." Instead, the AGPL forces on the reader a bunch of other terms that may or may not add up to "derivative works."

Other licenses have no problem coming out and using the term. Doing so connects those licenses to a well-understood body of intellectual property law. Not so with AGPL. See the discussion in the book by Rosen (an actual lawyer) for more [1].

So right out of the gate, the AGPL is not easy to comply with.

[1] https://www.rosenlaw.com/oslbook.htm

wwarner• 6 years ago

Sorry, this isn't really the case. The AGPL can be used in a predatory fashion. Projects that are AGPL'd are often dual licensed, with a commercial license for commercial applications and AGPL for experiments. I once negotiated for a couple of weeks to pay for an open source package that was dual licensed under the AGPL. I wanted to make small modifications to the source. The commercial terms were available on request, and I requested and got sort of tangential answers until finally the company's lawyer told me that I'd be liable for 15% of all revenue connected to the library in any way. If I had used the library first and asked for the terms later, I could have been in a very serious situation.

zzo38computer• 6 years ago

I use software with GPL and AGPL, and may even contribute, but when I write my own software, I make it to be the public domain instead, because I don't like copyright, and because public domain will make it we don't have to worry about the license conditions, in case it may be confusing.

I have a program which links with AGPL software. My program is itself public domain, and is released only as source code, but I would think if anyone distributes it, or makes it available to access over the network, that they would have to do so in a way which is not contrary to AGPL. If you know how it is work, then hopefully you can comment too, to say what parts I did wrong, I suppose.

gridlockd• 6 years ago

Why should I listen to the cigarette salesman talking about the harmlessness of smoking?

dekhn• 6 years ago

Back when I was a postdoc I talked to my advisor about GPL and BSD. The advisor (who was a supremely intelligent person) said the simplest, clearest possible thing I could imagine: "I wouldn't license software under GPL because I read the GPL license and couldn't understand what it was saying. I read the BSD license and it was totally clear. I want to license something using a license that makes technical sense to me."

The fact that we see pages and pages of hackers arguing over what the implications of a license family (GPL and AGPL) are based on the legal text suggests that many people don't understand the GPL.

GordonS• 6 years ago

Yep, this is precisely why I refuse to use GPL or AGPL, but am completely comfortable with BSD.

If I'm going to be bound by T's and C's, I need to be able to understand the implications - there is just too much ambiguity and room for bad actors (e.g. Oracle) with GPL and AGPL.

telamohn• 6 years ago

I see a lot of reflections from a business point of view.

I guess if your sole purpose is to conjure money out of thinware then yes, AGPL is going to be poison to your wallet. Capitalism and AGPL simply does not mix well.

But we tend to forget, that the end of the day we're all users and have to consume what we've cooked.

As a user I tend to lean towards AGPL alternatives, It's comfty and let's me relax my legal muscle after the hard workout of summarizing this thread.

The terms of the AGPL are pretty easy to comply with

Related Stories