LWN: Comments on "Redis modules and the Commons Clause"

Redis modules and the Commons Clause

pabs — Tue, 11 Sep 2018 02:23:29 +0000

"The Commons Clause doesn't help the commons"
https://mjg59.dreamwidth.org/51177.html

Redis modules and the Commons Clause

cladisch — Mon, 03 Sep 2018 06:30:00 +0000

VCs specialize on high-growth startups. Not everything is a high-growth/high-risk/take-over-the-world enterprise, and other forms of private equity or debt exist.

In other words: it is in a VC's interest to change an (Open-Source) company into a high-growth one, but not necessarily in the company's interest.

Redis modules and the Commons Clause

Cyberax — Sun, 02 Sep 2018 20:09:36 +0000

> He argues from a venture capitalist point of view, which is, frankly, not very relevant for anybody else.
Why is that? For individuals or small companies another option might exist - paid support, but this is really a thin gruel. It's unlikely you can grow your company just with that.

And for individuals perhaps Patreon or similar services might provide just enough money to be able to eat.

Redis modules and the Commons Clause

cladisch — Sun, 02 Sep 2018 19:02:24 +0000

> https://medium.com/@mattklein123/the-broken-economics-of-...

He argues from a venture capitalist point of view, which is, frankly, not very relevant for anybody else.
The most telling sentence:
> If we take consulting, services, and support off the table as an option for high-growth revenue generation (the only thing VCs care about), we are left with open core, SaaS, or some blurring of the two.

Yup, those off-the-table methods are how you do generate revenue from OSS.

Redis modules and the Commons Clause

pabs — Sun, 02 Sep 2018 06:35:03 +0000

Also Nadia Eghbal has been posting on these issues for a while:

https://nadiaeghbal.com/

Redis modules and the Commons Clause

pabs — Sun, 02 Sep 2018 06:33:24 +0000

Another followup on the economics of open source:

https://medium.com/@mattklein123/the-broken-economics-of-...

Redis modules and the Commons Clause

jrn — Sat, 25 Aug 2018 18:55:08 +0000

Can we *not* adopt this commenting style (name calling, etc), please?

AWS NIH

Cyberax — Sat, 25 Aug 2018 07:25:17 +0000

> Mellanox and Chelsio NICs have very similar functionality to the ENA, although I can imagine several legitimate reasons why AWS might have chosen to develop their own NIC.
ENAs have a lot of "secret sauce" inside. But to be fair, lots of larger instances just use a regular Intel network card using PCI pass-through.

AWS NIH

wmf — Sat, 25 Aug 2018 05:26:55 +0000

Mellanox and Chelsio NICs have very similar functionality to the ENA, although I can imagine several legitimate reasons why AWS might have chosen to develop their own NIC.

Redis modules and the Commons Clause

pabs — Sat, 25 Aug 2018 03:01:35 +0000

Redis modules and the Commons Clause

armijn — Fri, 24 Aug 2018 21:00:11 +0000

Yup, that's the one.

Redis modules and the Commons Clause

bkuhn — Fri, 24 Aug 2018 15:12:43 +0000

Your comments were interesting, kronat. My responses are below:

kronat wrote:

I believe there are cases in which maybe the freedom 2 and 3 gives too much power to who is just exploiting the work done by others.

This seems more of a general criticism of non-copyleft licensing, not of the four freedoms. I agree with you that licenses that fail to ensure freedom 2 and 3 in perpetuity generate other types of problematic power imbalances, but that particular issue seems orthogonal.

But I don't want to forbid the using of BSD license, or making defamatory statements against who is using it, just because I don't understand it or I do not believe in it. I will, more honestly, avoid contributing to projects which are distributed under BSD license.

Neither do I. I've licensed some of my own copyrights under 2-Clause BSD before. I have also not made defamatory statements about anyone; I criticized bad policy and pointed out manipulative strategies in use to advocate for that policy.

I believe who created the software and who maintains it has more power than who arrives later and make a contribution, but I don't see this reflected in any license.

The goal of copyleft is to give equal rights to the original author and all later contributors and users. No copyleft license is perfect, and if you have ideas for how to improve copyleft licenses to address some of its flaws, Fontana's coypleft-next project might be a good place to do that.

there are roughly two passages in which you state your opinion in the way Catholics politicians want to make abortion illegal for everyone, Catholics politicians want to make abortion illegal for everyone, regardless of the reasons, just because they believe that the right way is the only one they think.

The analogy is not apt nor accurate. I did not argue that it should be illegal for Redis or anyone else to publish software under the Commons Clause. I said that it is bad policy, and exploits a power relationship granted by existing copyright laws in a way that I find inappropriate and bad for society. I have never argued that proprietary software should be illegal. Even though I'm morally opposed to proprietary software, I don't think (e.g.) changing copyright law to only allow FOSS licensing would be successful in eradicating proprietary business models.

Redis modules and the Commons Clause

jubal — Fri, 24 Aug 2018 14:44:25 +0000

Bain Capital, isn't this the “investment” company Romney was involved with?

Redis modules and the Commons Clause

kronat — Fri, 24 Aug 2018 07:23:35 +0000

I respectfully disagree with Mr. Kuhn's position. You seem capable of guessing what is better for all other people, without actually caring to understand their position. In particular, in your blog entry, there are roughly two passages in which you state your opinion in the way Catholics politicians want to make abortion illegal for everyone, regardless of the reasons, just because they believe that the right way is the only one they think. Well, I believe there are cases in which maybe the freedom 2 and 3 gives too much power to who is just exploiting the work done by others.

More philosophically, I do not understand BSD license: its permissiveness allows a company to take/use/modify a project (and I'm thinking about small projects, done in the spare time, and things like that) without even the share-alike requirements of the GPL. But I don't want to forbid the using of BSD license, or making defamatory statements against who is using it, just because I don't understand it or I do not believe in it. I will, more honestly, avoid contributing to projects which are distributed under BSD license.

Going back to my first paragraph, I would like to ask if freedoms 2 and 3 make sense in today's environment. What I feel is that one of the most logical ways to develop is to distribute the software under the freedom 0 and 1. Then, anyone that wants to share their improvements should do it by passing the changes upstream (for free or by receiving compensation in doing so). I believe who created the software and who maintains it has more power than who arrives later and make a contribution, but I don't see this reflected in any license. The problem is that only in recent times big fishes (adapt the definition of "big fish" to the size of the project we are considering) have discovered plenty of software that they can use "for free", while their creators and maintainer have to do two jobs to pay bills (well, apparently it is not the case of Redis by looking at the amount of money they received these years).

Redis modules and the Commons Clause

jberkus — Fri, 24 Aug 2018 01:07:03 +0000

I believe that they just received another round of funding, from Goldman Sachs and Bain Capital, immediately before announcing this "Commons Clause" thing. Draw your own conclusions there.

The Commons Clause really isn't as important as we think it is

nilsmeyer — Thu, 23 Aug 2018 22:40:47 +0000

It's just an example of a contribution that only benefits Amazon - sorry for not making this clear. Or is it possible to buy a device compatible with the driver?

The Commons Clause really isn't as important as we think it is

Cyberax — Thu, 23 Aug 2018 22:38:10 +0000

As an Amazon employee who worked a fair bit with the ENA source code, what are the non-NIH-ed alternatives to it?

I'm seriously interested.

The Commons Clause really isn't as important as we think it is

nilsmeyer — Thu, 23 Aug 2018 22:18:26 +0000

I think the issue of the dangerous nature of the Cloud oligopoly is worthy of attention. Especially Amazon is a pitiful example of corporate citizenship, not only are their contributions mostly limited to supporting their NIH infrastructure (like elastic network adapters), they also tend to abuse their employees and try anything not to pay taxes. Endangering free software is just another addition to the list...

Ironically, most of the hosted, managed services perform worse than building something on the basis of virtual machines alone.

Misleading/confusing text is the real problem

david.a.wheeler — Thu, 23 Aug 2018 18:44:15 +0000

I agree that the misleading/confusing text is in many ways the real problem. There are millions of proprietary licenses, and many organizations use "open core" business models. But calling it "Commons" makes it sound like this is from the Creative Commons Corporation, and the way they refer to Apache makes it sound like this is from the Apache Foundation. In addition, a lot of discussion about OSS refers to "the commons" - yet this "Commons" license isn't OSS. If they'd make things clear, including renaming things to make things clear, it would be <i>much</i> better.

Redis modules and the Commons Clause

jberkus — Thu, 23 Aug 2018 18:38:48 +0000

The bizarre thing about Neo4J adopting the Commons Clause for some modules is that those modules were already proprietary. That makes the Clause inherently meaningless for them; it's like pure grandstanding.

Odds that it was forced on them by their funders?

Redis modules and the Commons Clause

bengen — Thu, 23 Aug 2018 13:26:18 +0000

For the last 3 years we see... 2015-06: +$15m, 2016-07: +$14m, 2017-08: +$44m.

The timing would be about right for negotiations about further funding. Perhaps some early investors really want to extract some profit from their investment, right now?

Redis modules and the Commons Clause

aigarius — Thu, 23 Aug 2018 10:19:33 +0000

Any such restrictions make the software non-free. It is plain and simple. It is sad to see Redis Labs moving to non-free software, but that is the fact. Trying to mislead the public with naming and FAQ and statements will not change the fact. "Commons Clause" added to *any* license makes it non-free. That's it.

Distributors will likely fork the last free version of the modules and move on from there without Redis Labs, like it has happened numerous times in the past.

Redis modules and the Commons Clause

nim-nim — Thu, 23 Aug 2018 08:50:33 +0000

Nothing surprising about influxdata, they went open core quite a long time ago, surrendering their market lead to actual free software alternatives (Prometheus).

Basically, they want the community, without surrendering control, and can't stand that anyone else can “take over” their project by funding a greater dev team (that's the real sticking point, not that they can't get funding, but that someone else may get more of it).

Redis modules and the Commons Clause

nhippi — Thu, 23 Aug 2018 07:03:40 +0000

Well all those companies contribute massively back to Linux. Well, perhaps except Amazon, whose contributions back to upstream open sources projects is pitiful.

Redis modules and the Commons Clause

pabs — Thu, 23 Aug 2018 05:33:46 +0000

Some more coverage:

https://www.influxdata.com/blog/its-time-for-the-open-sou...
https://news.ycombinator.com/item?id=17822257

http://antirez.com/news/120

Redis modules and the Commons Clause

sml — Thu, 23 Aug 2018 03:37:47 +0000

I think the worst part of this is the deliberate obfuscation caused by the naming that Commons Clause encourages.

Plenty of companies have relicensed their previous Free Software product with a proprietary license. Anyone who signed the CLA would know that they are signing the rights to their code over to Redis Labs. And Redis Labs wanting to make money from their product is not a bad thing.

However Apache 2.0 is a well known and well regarded Free Software license. "Apache 2.0 modified with Commons Clause" is a dirty trick to take advantage of the good reputation of Apache 2.0. If Redis Labs renamed their license to something more honest like "Redis Labs Proprietary License" it would fix all the problems I have with this.

And FOSSA should be ashamed for promoting such weasel wording.

Redis modules and the Commons Clause

bkuhn — Thu, 23 Aug 2018 03:04:21 +0000

IANAL either but I don't think one needs to be a lawyer to figure out what's going on with that. The main issue is that it's hard to know without digging into the code, but when I read that text that mentions combination of Commons Clause and AGPL'd code, I suspect what's happening is that Commons Clause parts are coming only as part of the Enterprise edition, which is not the demo-ware version that's AGPL'd on GitHub.

The thing that makes me most concerned about these situations is that none of the people that get trapped into one of these predatory business practices of “try the AGPL version and if you like it get the Enterprise one” have enough background knowledge on what's happening to even begin considering the creative solutions you're suggesting. Companies and individuals with the resources to research and understand these things just avoid codebases like this. Their target market are, as other poster have said in this thread, small to medium businesses who just won't know what they're getting into until their stuck.

This is a problem, BTW, with any copyleft relicensing business model. The Commons Clause is almost a complete red herring for the companies that already have proprietary relicensing going on. I suspect they like it mainly because they sense solidarity with other businesses that are trying to build a block of businesses that can give “something akin but not really Open Source is good enough” legitimacy. What we've learned here that if one company or a small group of companies hold all the copyrights on software that you rely on, be worried, even if the license seems ok.

The Commons Clause really isn't as important as we think it is

bkuhn — Thu, 23 Aug 2018 02:55:07 +0000

So many people were writing to me and Conservancy to ask for an opinion on this Commons Clause thing, that I wrote a blog post so I wouldn't have to keep responding individually. The one point I didn't make in the blog post is that I think we're all (including me) giving this more attention than it deserves. This is one silly non-Open-Source licensing thing drafted by one lawyer who is promoting multiple “sounds like Open Source but really isn't” licensing schemes. In turn, a very small group of companies think is coolest licensing thing they've ever seen, presumably because it sits better with their VCs than actual FOSS. There's not much more to the story than that.

While Jake's article is well written and covers all the angles, I worry that giving it this much attention (including the attention I'm admittedly giving it by writing a blog post and by posting here) actually is helping these anti-Free-Software folks promote this thing, and that our response subtly legitimizes something that really is just “yet another creative way to do proprietary software”. That's the trap of press coups like this: someone proposes something outrageous, everyone feels compelled to respond because they fear that something this outrageous might get traction, then it looks like it was a legitimate and widely held position from the start (instead of the fringe activity that it is), and then suddenly, the position is granted legitimacy because the response in outrage gave the position legitimacy it wouldn't otherwise have.

The best outcome, which fortunately is still pretty likely for the moment, that everyone will have trouble even remembering what the Commons Clause was in five years, and the codebases that chose it will fall to obscurity, or have since been licensed another way, or both.

Redis modules and the Commons Clause

Cyberax — Wed, 22 Aug 2018 23:58:45 +0000

Linux is used in core systems by Google, Facebook, Amazon, IBM and other huge companies. It's business-critical for them.

Redis? Not so much. It's used in a lot of small and mid-range companies, and it's not a good place to be.

FreeBSD suffers from the same issues - it's not used a lot, so it's moving at fraction of Linux's speed.

Redis modules and the Commons Clause

gerdesj — Wed, 22 Aug 2018 23:54:30 +0000

"I feel for Redis. They are widely used " - replace Redis with Linux or *BSD or whatever in that comment and have another go mate.

I respect you as a commentard - have another go.

Redis modules and the Commons Clause

armijn — Wed, 22 Aug 2018 22:15:29 +0000

*cough*

https://www.crunchbase.com/organization/redis-labs#sectio...

Redis modules and the Commons Clause

armijn — Wed, 22 Aug 2018 21:20:42 +0000

The GPL family of licenses has always explicitly allowed selling the software. AGPL3 for example says in section 4:

"You may charge any price or no price for each copy that you convey,
and you may offer support or warranty protection for a fee."

which the Commons Clause is now trying to take away. Possibly it is not compatible with section 7 of AGPL 3, and then people could just remove Commons Clause for Neo4J Enterprise and use plain AGPL3:

"If the Program as you received it, or any part of it, contains a notice stating that it is
governed by this License along with a term that is a further restriction, you may remove that term."

and all that would be left is a bitter taste.

PS: I am not a lawyer, this is not legal advise, etc. etc.

Redis modules and the Commons Clause

Cyberax — Wed, 22 Aug 2018 21:08:11 +0000

I feel for Redis. They are widely used but large players like Google or Amazon are not really interested in supporting it, they all have their own in-house databases that are used in place of Redis.

And smaller players don't have resources (or want to) support such a basic component as the database.