The Commons Clause is an existential threat to open source

I would hope this was not an intentional choice to sow confusion.

Edit: obvious spelling error :)

By that I mean, Redis Labs are not the creators of Redis. They are a consulting agency that then hired the creator of Redis. In other words, they only exist because the original Redis was not Commons Clause. Now they want to build a moat so others can't do what they did.

Not a good precedent.

I can accept that you might not appreciate my work on the Commons Clause, but I'm glad to see you acknowledge why funding OSS is so difficult.

I think it's okay to fear the world of open source becoming a bit more closed. I do too. Whether that's a better world or not is up to debate, but I certainly don't think so.

I think where we might disagree is what the greatest threat to "open-ness" is. Sometimes people do bad things to open source projects, and force them to go closed. The license choice will inevitably follow -- and when they do, they can choose to close the source entirely or choose a licensing scheme that makes it available.

Licensing follows intent, and I certainly don't think the Clause inspires people to close their source. But sometimes people need to change their license -- and I think it's unrealistic to say that this doesn't happen.

But open source didn't depend on these funding models in the 90s/2000s did it? I did a whole post on the philosophy of open source a while back:

https://penguindreams.org/blog/the-philosophy-of-open-source...

..but one of the things I didn't find enough sources on where a lot of funding came from (And percentage of volunteer contributors).

This recent Redis things reminds me of the Caddy controversy, and maybe even more recently, how Win/Mac Docker binaries are behind a login wall.

What changed? Do developers no longer have the time to volunteer to work on this stuff? The desire? Is the demand for tech in The Valley/West Cost changing the dynamics of how much time people have to contribute. Were the earlier efforts way less commercial and new efforts moving much faster and requirement more maintenance/resources?

1. A group of people go off and, without involving their communities at all, try to solve a problem they think they have.

2. They spring it on their communities suddenly and without warning

3. Negative reaction ensues.

4. People decry death of open source.

5. Author of clause bemoans on twitter that the reaction of community should have been collaborative problem solving (and "starting a conversation"), despite completely failing to attempt collaborative problem solving with their community.

Is that about right?

If so, this is a great object lesson in how not to do this.

If you want the reaction to be collaborative problem solving, engage people in collaborative problem solving.

If you want the reaction to be to your solution, present people a solution. But expect people to do precisely that!

This is like when people drop completely designed/implemented megabombs of stuff on OSS projects and then complain when there is pushback.

Wow! That seems like a mistake. This casts too wide a net and catches more than it should.

> ... product or service whose value derives, entirely or substantially, from the functionality of the Software...

Ugh. Instead of unambiguous freedom-granting terms, this leaves it up to a court to help determine this vague notion of the product or service's "value".

Steer clear of these license terms!

EDIT: deleted request for OSI to clarify

One challenge is the name "Commons" is very misleading. It's not creating a commons, and it's also confusingly similar to the term "Creative Commons" which has been around for a really long time.

Also, the Commons website says that the Commons license is not OSS "at least not by the official definition set forth by the OSD". Yet that is the official generally-accepted definition.

Ultimately, though, open source has to incentivize independent creators and not big corporations. AGPL is somewhat successful here, because big corporations already ban its use internally, but it's far from addressing the problem. New restrictive licenses are needed.

@Antirez has confirmed the core project is remaining BSD. The "Common Clause" stuff applies to the modules that RedisLabs create.

Additional modules @Antirez creates himself, are also not under Common Clause, they're AGPL.

There is nothing to see here, folks; "free-with-source for non-commercial use" has been a thing in the proprietary world for decades. It's like "free trial" or "freemium" with source code.

It's a good thing for customers and would-be customers. Customers can track down bugs and security flaws if they have access to the code. Would-be customers can evaluate the product based not only performance of the executables but code quality.

It is a "better proprietary", in other words.

The guidelines/definitions/blah blah blah exist precisely because of attempts to do this in the past.

If people want to try again, great, let them. Experimentation is good. (I personally wish they'd stop experimenting with repeatedly proven-to-fail models, but c'est la vie)

In the end, it's the developers who are going to decide whether it's the right way for the world to be.

That is, it's not going to destroy open source or anything else unless the people using/developing open source want it to.

Past that, Kevin's refrain is kind of tired, at least to me. Open source has always been a community that basically eats its own over time (nothing is ever good enough, free enough, open enough, etc). Solving that seems more important than solving this.

Saying "The OSS project we tried to build as a business sinks tons of money into open source and gets no help" as if it was a new problem is just silly. This is not a new problem at all. The fact that they believe doing literally the same thing that hasn't worked in the past will suddenly work to solve this problem in any meaningful way is disappointing.

For a person whose twitter bio likes to tell me how impressive they are, this is not impressive at all. I would have hoped for something better.

Because if this is all they've got, then yeah, that model of open source is likely screwed.

I will paraphrase a colleague:

The underlying issue they are trying to solve is that developers believe they should be able to extract some of the wealth they feel they created. In open source, that wealth is often only created because the software was free. Otherwise, people would have used something else that was free but worked well enough. Developers like to often argue this isn't true, but history shows it to be true basically always :) So saying you should be able to extract this wealth is probably wrong. Saying you should be able to get paid a reasonable amount of money is not.

It's a dirty practice to build something open source and after a certain point, try to close it off. But, it's not unheard of. Any project can change their license, I believe Google did this with Android. Older versions will get forked and others will do with those forks what they choose.

I'm hoping doing this would cause so much backlash that it'll force projects to revert back to OSS. But, it could just push people out of contributing to OSS -- then again, there might be some people more likely to contribute if they knew no one was going to make money off their work.

Let the market sort it out.

Software: [name software]

License: [i.e. Apache 2.0]

Licensor: [ABC company]

If they actually used the Apache 2.0 license anyone would have the right to sell the software. Hence putting that clause and then putting "Apache 2.0" as the license makes no sense. The licensing terms would be contradictory.

> If the Commons clause were to be adopted by all open source projects, they would cease to be open source, and therefore the Commons clause is trying to destroy open source

I’m new to the matter, but I do see in the Commons Clause an intent to manipulate and mislead. If the goal is to provide an option for companies who would otherwise take their software closed-source, why does that option have to sound super-duper open sourcy while not having terms recognizable as open source? I’m not even a rabid free software guy or anything, but just compare the name to the terms. There is no possible way that “Commons Clause” is not intended to be confused with, or somehow gain a positive spin that has nothing to do with its content from, CC or Apache Commons.

People who care about things like Creative Commons or about open source as anything more than “source available” will label this as “evil” or at best a marketing ploy.

In the case of Redis, I think that it's a really deserving project and I fully trust the integrity of the people behind it.

As someone who created a popular OSS project outside of Silicon Valley, I can relate to how difficult it is to be an outsider in this industry and I understand why Redis Labs would feel entitled to capture a tiny fraction of the enormous value that they created.

I think that if the Commons Clause is just related to optional things like plugins or addons then I honestly don't see any harm; at least the plugins/addons are still available for small companies to use for free so it's still a step up from the typical OSS freemium model where they sell the plugins.

But it's not true, it's complete, self-admitted clickbait. An asteroid will destroy most life on Earth tomorrow. If a sufficiently large asteroid strikes Earth tomorrow. It's an overwrought premise leading to predictably overwrought handwringing.

> For purposes of the foregoing, “Sell” means...fees for hosting or consulting/ support services related to the Software...

Why anyone would ever use this license is beyond me. No paid hosted version. No paid consulting. No paid support.

I mean this is great for folks who have already made their money but for those of us with bills to pay, it's not exactly realistic.

[0] https://commonsclause.com/

Redislabs made a good point, cloud providers are making huge amounts of money off these products and often contribute little back. Take Linode and their Block Storage service for example. They use Ceph, but contribute nothing to the Ceph open source. Or QEMU, the entire basis of their business, and not a single commit from anyone there.

The Commons Clause addressed the AGPL question you ask in your comment.

Redislabs states the goal is not to close source redis core, that will remain free forever, and that the use of the priority modules will be free for most use cases, but when companies makes money from packaging it all up as a service, they should have to pay a license fee, which seems completely fair.

I’ve worked for a large cloud providers and some small ones, it’s pretty common at these places to not care what the License is because as long as it’s running on a server, it doesn’t matter. And selling service based off them, just means profits without giving back in most cases.

The final point of the Commons Clause says very clearly that this was not meant to be slapped on every OSS license and that it’s not a final solution but meant to spark conversation.

In that way, it’s working exactly as intended.

Not every business can be Red Hat.

  - A person or company creates an open-source library, at great cost of time and effort.
  - A big company adds that library to their cloud offerings, at arguably far lesser cost, and monetizes it.
  - The library author receives no compensation.

Commons Clause is misleading in the context of Open source licenses.

On a separate note, it is important to address volunteer open source contribution and development. I think as a community we are missing a model that guides how :

a) developers,

b) their hosting organizations

c) financial contributors / donors

can all benefit from their activities.

Right now, there are different models and experimentation are being used.

But I think we, as a community, could come up with 3-4 models that make it worthwile to the above groups.

If I can choose to apply whichever license I want, then it would seem that The Commons Clause could only possibly harm those that choose it.

Anyway, there are plenty of unprofitable licenses out there, and the reality is, for all of them, if you’ve already given something away for free, once the cat’s out of the bag, you probably won’t make any money, even if your license doesn’t say you’re a bad person for even thinking about trying.

That said, The Commons Clause is probably fine for foundations, where lots of people are sharing an array of technologies, and not really asking for money. Meanwhile, it’s been several decades of HTTP over TCP/IP, and last time I checked, there isn’t a whole lot to lock down, and the real money is in adding reach to flip switches from the other side of a network connection. JSON isn’t going away, so, there really isn’t much to apply licenses to, in this area, without being silly.

But hey, databases, operating systems, desktop applications, peripheral drivers, all these things still happen, so I guess there’s stuff to license. But really all the fringe licenses mostly get applied to JavaScript frameworks, and by the time you actually read a JS library’s license, three other projects have replaced it, so why bother.

https://github.com/dlang/dmd/blob/master/LICENSE.txt

But any attempt to make software open source, but still make sure that I get licensing fees or whatnot if someone makes a profit from it, is in my opinion seriously misguided.

With Open Source, our business model should be that lots of people work together to build an ecosystem; no-one profits directly from selling the code, but everyone profits from using the code. Some people contribute to the project because they offer consulting services, some people profit because they use it for their job, etc.

Open source should be just about taking, if million dollar companies, forget supporting projects they rely on, can't even tell the authors something has gone wrong.

Ideally those who can afford it should throw some support to the projects they use. But clearly this is not happening, if the project becomes too important then they simply acquire the project with downsides for everyone.

There has to be a middle way that has upsides for all parties and untill that is figured out this can work in the interim.

Open source developers are building software for free that is then used by extremely large corporations to make vast amounts of money without compensation to the original author. Even worse much of this revenue comes from business practices that the original authors may find highly objectionable including mass surveillance, "surveillance capitalism," and data-driven propaganda, con artistry, and manipulation.

If you create open source today you have no choice but to indirectly offer free labor toward the construction of a global surveillance state. The commons clause offers a way out while still encouraging people to use your software for personal, private, academic, and other purposes and retaining the openness of open source.

The "free as in beer" part of OSS is a relic of the 1990s when the prevailing enemy was "embrace-extend-extinguish" at the hands of (mostly) Microsoft. It was sort of a growth hack to enable open source software to gain critical mass and challenge closed source proprietary platforms.

Times have changed. Today the big threat is SAAS walled gardens, surveillance capitalism, and other business models that in many cases run on and depend upon open source. In some cases SAAS can verge on piracy-- not necessarily in the legal sense since nothing is being done to violate any license but in the moral sense of profiting handsomely from the work of others without compensating them.

Edit: I will provide a concrete example of the last point: RethinkDB.

RethinkDB was probably the best NoSQL database, at least before Mongo was improved enough to challenge it on robustness and versatility. It still has features Mongo and others lack.

Unfortunately RethinkDB the company failed because... well... nobody paid them. Instead they paid services like compose.io to run RethinkDb for them in the cloud. Cloud SAAS database services were making more money off RethinkDB than the database's own developers were. In the end they failed and the project is now almost abandoned.

There are numerous similar examples. AFIAK the Commons Clause would have prevented this, allowing either RethinkDB to capture their own SAAS market or charge compose.io something for making money off their software.

1. The existence of the Commons Clause. 2. The application of the Commons Clause to software that previously was not subject to it.

To the original author: How would your view be different if a company released a new product or module that was Commons Clause-licensed from day 1? Would that still be a threat to open source? (What if the alternative to it being Commons Clause-licensed was that it wouldn't exist at all?)

The problem is when I want to, say, copy a few functions into my own project which I'm publishing under completely unrestricted terms (say, Unlicense). Or use as a dependency, etc.

So, I'd say it's okay-ish for "product-like" stuff, for user-facing applications. But please NEVER publish libraries everyone depends upon under these terms.

Cloud providers do redistribute. It runs on their servers. They do not have to honor copyright restrictions. They can make proprietary modifications and keep them, and they do. They might be honoring the letter of the license, but certainly not the spirit of it.

What's the difference between the Commons Clause and the creative commons "nc" (non-commercial) license?

I'm having trouble seeing how the Commons Clause is a threat to open source, while the GPL and Creative Commons are not.

This Commons Clause is pure blarney and Drew is exactly right.

The best open source monetization case I've seen yet may be SQLite3, which is in the public domain. SQLite3 has a consortium, has some proprietary products (e.g., the more extensive test suite) not available to the public at large (in source form or otherwise), and is so ubiquitous that there are a number of large companies that are so utterly dependent on it that they are practically obligated to be members of the consortium. I've no idea just how profitable SQLite3 has been for D.R. Hipp and friends, and I don't want to know, but I assume it's been good enough that they are happy to continue working on it. SQLite consortium members probably (I imagine anyways) get to influence SQLite3's development, and that's one generic way to monetize source code regardless of license: get paid to do further (specific) development / bug fixes / support, but SQLite3's proprietary test suite makes it very difficult for third parties to be able to have contributions welcomed by the core, which tends to limit third party commercial support.

Choice of license doesn't seem to have that much to do with whether something can be monetized. Consider the Linux kernel, which is GPL'ed: prolific contributors profit mostly by being well-employed to continue being prolific contributors, but almost certainly no one is making a dime from commercial Linux kernel usage without source code distribution. There might be some GPL licensed codebases where the copyright owners are making a few bucks from commercial usage by third parties who don't want to distribute their changes, but probably not a lot. There are almost certainly no BSD/MIT licensed codebases where the owners are making any money from commercial usage by third parties -- the license simply allows third parties to use such code outright with absolutely minimal obligations on the third parties.

You can have open source and monetization for some value of "open source", but not one that most people would really consider "open" -- getting to see the source code != open source.

Getting paid to support open source (e.g., add features, fix bugs) can be very difficult when it is easy for anyone to do that. SQLite3 has a mechanism for limiting third party contributions (a proprietary test suite) which limits third-party commercial support though in general that would also increase the risk of forks (but SQLite3 is so widely used that forks are extremely unlikely). The SQLite3 approach won't work generically for very many open source projects as it depends on SQLite3's popularity and very widespread use.

The fact is that for many of us open source is just a way to build and maintain a reputation, and we then monetize the reputation.

I myself have been offered payment for development on open source codebases that I maintain, and I always reject such offers, mostly because accepting them would interfere with my main responsibilities, and in some cases also because it's not my right (e.g., I don't think it would be fair to take money to work on jq without giving @stedolan the right of first refusal). If I were to lose clients/employment, I might reconsider this.

What kills open source software is greed, and very few people are immune from that.