[swiftcoder]

> the stance of the 'Open Source' crowd

The original stance of the open source crowd was more along the lines of the GPL -> GPLv3 -> AGPL, which expressly prevents this kind of thing.

The proliferation of "give everything away for free" MIT/BSD/Apache licenses seems to me to have been an intentional campaign by corporate interests to undermine free software ideals

[jwr]

> The proliferation of "give everything away for free" MIT/BSD/Apache licenses seems to me to have been an intentional campaign by corporate interests to undermine free software ideals

As a counterpoint, when I make something open source, I really mean "freedom", which includes the freedom to build a commercial service using the software. I use the MIT license not because of "corporate interests to undermine free software ideals", but because I really want the software to be free as in freedom.

GPL, GPLv3, AGPL and similar license actually restrict the freedom to do anything you want with the software. I'm not saying there is anything wrong with it, just that "free software ideals" could mean different things to different people, and there might not be any "corporate interests".

[FuriouslyAdrift]

The GPL is meant to be viral... it infects other projects so as to open up software ecosystems. The open source movement came out of a time when nearly everything was proprietary and locked up.

Tivo-ization really woke up a lot of people to the dangers of proprietary lock in and abuse. https://en.wikipedia.org/wiki/Tivoization

[Jensson]

> The GPL is meant to be viral

But that also makes people not wanna bother with it. People in the middle who doesn't care either way will be very apprehensive of putting in a license virus in their project.

[FuriouslyAdrift]

You should always know what patents and licenses are part of your project/product.

[me-vs-cat]

This is both true and absurd.

True because that's how the system actually works today.

Absurd because monopolies on ideas -- specifically patents and software licenses as mentioned here -- have absurd durations.

[kmacdough]

But the point others make is, this type of libertarian free is wide open to the Embrace, Extend, Extinguish strategy. A very successful, repeatable business strategy to own supposedly "open" domains. There's a reason people have debated freedom to and freedom from for centuries. This is the core conundrum/challenge of freedom.

GPL style licenses provide some guarantee you're investing in an ecosystem that is resistant to EEE. Freedom from takeover in exchange for freedom to make any arbitrary business venture. It's a choice, but to conflate libertarian freedom as the only form of freedom is narrow and ignores this centuries old unsettled debate.

[rcxdude]

There are two freedoms of different people (or rather different roles) that are in conflict here: the freedom of developers to do whatever they would like to do with code that they have access to, and the freedom of users to be able to change and control the software they use. MIT/BSD/etc prioritise the former, while GPL prioritises the latter: free software advocates generally believe that proprietary software is immoral, and that _all_ software should be open to users to modify, even if that limits developers freedom to keep it secret. The GPL is an attempt to enforce this as much as can be achieved under current law, not a natural reflection of their wishes (which would compel all software to have source code available for modification).

(There's also a secondary motiviation for using the GPL which seems to be driving this kind of discussion, that of 'paying your part', but this is neither open source nor free software in origin IMO: the desired deal is 'I make the source code available, but I want a cut if you're making money from it'. You can often do this with a version of the GPL that is viewed as sufficiently anti-commercial, and then offering a paid proprietary license, but this is antithetical to the goals of those who wrote the GPL)

[spookie]

Corporate interests != GPL is not on the cards

[tsimionescu]

Amazon offers lots of AGPL software, and they fully respect the license in all cases. Ultimately the GPL is about protecting users' rights at the expense of developers' rights. So as long as AWS can offer a better/cheaper managed version of a software service, while still giving the users all details on how to run the same service if they chose to, then the AGPL is completely achieving its aims, even if the original company goes out of business.

[echelon]

> protecting users' rights at the expense of developers' rights.

Protecting the user's right to compete with the developer is not sustainable.

Protecting the user's right to run the software for free on their own or in their company so long as they don't resell it is perfectly salient and should be enough for anyone. That's really all the freedom a user needs.

If you're asking for more, it's because you want to take the developer's business. That's 100% unfair.

The hyperscalers aren't giving back 1/1,000,000th of what they've taken. Yet we go after "source available" or "fair source" like it's some grave evil.

Where is there opportunity left for software outside of the major trillion dollar companies if we don't start giving developers the benefit of profiting on their work?

I make a point to cheer on every fair source, source available, or open core project I see. It's the sustainable path forward. We shouldn't be taking from each other - we should be finding out how to take back from the hyperscalers.

[chowells]

This is a very business-centric viewpoint. I publish a small number of open-source libraries. They are not a business. I have no interest in making them a business. In fact, the idea of making them a business is repellent. They're just code for doing some tasks more easily than starting from scratch.

I made some of them because I needed them, and had no reason to own them. I made some because I thought another library was poorly designed and I could demonstrate a better way. I didn't make any because I wanted money or recognition. I don't care who uses them, or how. It is literally impossible for a user to do anything with any of them that harms me.

I am deeply suspicious of any world view that declares it bad when people use code I have released for free. I released it so people would use it. Good for them!

[echelon]

To be clear: I am not talking this kind of open source.

Rather, full-time commitment to software of scale. Software that does have business use cases. Software where outages can cost money.

[FuriouslyAdrift]

All software meets this criteria to someone

[account42]

> Protecting the user's right to compete with the developer is not sustainable.

It's only unsustainable when you are interested in keeping "user" and "developer" as distinct sets.

> The hyperscalers aren't giving back 1/1,000,000th of what they've taken. Yet we go after "source available" or "fair source" like it's some grave evil.

No we are going after it when people try to pass it off as open source when it really isn't.

I like open source because it means I'm not beholden to the original developer in any way as long as I pay it forward. I'm OK if this means you can't find a profitable business model.

[echelon]

I can't comprehend this view at all.

I hate open source purism. It's not pragmatic and it's enabled us to be resold a world with ever disappearing rights.

This view is okay with hyperscalers. But it attacks the small developer.

The hyperscalers are removing our freedoms and privacy. Not small developers.

We need leverage against this.

[rpdillon]

Strongly agree with the view you're responding to. So maybe I can talk about it.

There's just tons of software that you expect people to re-host. Yunohost has a massive catalog of free and open source software that is specifically designed to be spun up in a matter of minutes on open source VMs. To do what you're suggesting for those pieces of software would destroy the ecosystem entirely. The goal is to have multiple providers that are interchangeable that can host the software you need. So if one provider goes down, you can switch.

Meanwhile, MBAs that wanted to make money on their open source software decided that a good way to do that was to host services in charge for them. I agree, but the challenge here is what do you do when Amazon decides to take your software and also make it available to host?

And that's the moment where people abandon free software because it's inconvenient for that particular business model. The bug is not in free software. The bug is in the business model of the companies wanting to claim that they're peddling open source software, while not actually doing so: they want to have a monopoly on providing that software as a service. I understand why, but it's not good for the customer.

A real example that's getting a little long in the tooth, but back in the mid-2010s, I wanted to buy elasticsearch for a geographic search for my startup, and turns out that elasticsearch hosting, which I preferred, didn't actually offer CPU intensive instances suitable for geo-hashing. And I ended up having to switch over to Amazon to get the kinds of memory and CPU allocations that were best for our use case.

I get that you're concerned about the sustainability of these businesses, but introducing a monopoly on hosting has other downsides.

[pabs3]

Open source means surrendering your monopoly over commercial exploitation:

https://drewdevault.com/2021/01/20/FOSS-is-to-surrender-your...

The hyperscalers are contributors to FOSS, both in code contributions and funding. They could easily far better though.

[anilakar]

> Protecting the user's right to run the software for free on their own or in their company so long as they don't resell it

One company uses the software internally to create more value for a customer of theirs.

A second company uses the software internally to provide a paid service to a customer.

A third company resells software they bought to their customer.

The end result is the same: Company makes money, customer exhanges money for value. Somehow only one or two of those use cases would be legal.

[tsimionescu]

> Protecting the user's right to compete with the developer is not sustainable.

I agree with you, actually - but Richard Stallman and the Free Software movement more generally really don't. They exactly and explicitly believe this right exists and should ideally be a legal right, and the AGPL quite explicitly maintains this right.

Ultimately the Free Software movement is predicated on the concept that ideas can't be owned. They generally oppose both copyright and patents, and not just for software. Their licenses are meant as a stop gap solution. Ideally to them, or at least to some of the more die-hard members, laws would be changed such that what the GPL grants would not be a license predicated on copyright, but instead a legal requirement for all software, while copyright would be entirely abolished.

In addition to their general opposition to copyright and patents, Free Software people also view software as having a special role in terms of privacy and control - that, even more so than books and other copyrightable works, you have a right to know what the software in your house and business is doing, and to modify and fix it if it's doing something you don't like. This is related to privacy rights on one hand, and also anti-monopoly, right to repair concepts on the other hand.

This is all very different from the Open Source movement, even though they basically use the same kinds of licenses. The OpenSource movement is more of an industry group that believes competing on building much foundational software is a waste of resources. Instead, they believe the best way to build this foundational software is in collaboration with other commercial or non-commercial entities, building it in the open such that all may benefit from contributions and add their own contributions. However, the Open Source movement is completely fine with, and even expects, then making a proprietary product on top of this open source base.

To them copyleft licenses are a tool to make sure others don't keep their improvements for themselves, but have the downside of making it harder to build your proprietary stuff on top. Conversely, software that takes your contributions but then doesn't allow you to use it in commercial offerings is completely unacceptable, since the whole goal of the movement is for different companies to build a common infra on which they can then build their own commercial products.

Ultimately, both the Free Software and the Open Source movements will agree that a core part of open source is that anyone should be able to compete on delivering the original software, even if for entirely different ideological reasons.

[LeFantome]

Agreed. The AGPL does not care at all about those writing the code. It is all about users.

[saubeidl]

The AGPL makes it so those users can be the same ones as the ones writing code.

[LeFantome]

As I recall, Open Source was about developers collaborating to make better software. It was a pro-developer philosophy vs the Free Software movement which was all about the rights of users (and developer hostile in my view). GPL and its children are from the Free Software Tradition.

Open Source provides the same “4 freedoms” as Free Software so most Open Source licenses qualify as Free Software as well.

If the goal is developer collaboration, permissive licenses are often the best choice. If you want maximum user entitlement, copyleft licenses limit developer freedom in exchange for a guarantee that future code will also be released as free software.

Cloud hosting was a challenge that did not exist when either philosophy first emerged.

With hosting, you are able to become the preferred source for software without adding much value to the code itself. This is what the author is complaining about.

The AGPL tries to address this in the GPL family but I don’t think it quite gets there. For permissive licenses, we see these “no hosting” exceptions.

If you read the early writings from the Free Software Foundation, they do not care if devs can make a living. The goal is user freedom. I think it is this philosophy that objects to the hosting exceptions.

Perhaps a better solution will be found in the future.

[sgc]

I often think the solution is to move away from crafting a perfect ideology to encapsulate in your license, and just throw out some numbers. If you make more than N* the median income of this or that place, you can't use this software for free (whether that means licensing fees, code contribution, etc can vary). Let the smaller fish grow. If they get big enough, they can give back.

[anktor]

How would that ever be enforced? Do you not run into WinRAR/Sublime problem of "ey you've been using this, pay us, please"?

[sgc]

At 50k USD / year revenue it would be impossible, probably a lot easier at 1m / year. Higher profile use cases are known, and companies tend to comply with licenses rather than pirate software. Just like now, you lose a lot of control as soon as your source is available.

[zimpenfish]

> The original stance of the open source crowd was more along the lines of the GPL -> GPLv3 -> AGPL, which expressly prevents this kind of thing.

Not wanting to further widen the schism but wasn't that the free software people rather than the open source people? cf [0], particularly the "not as strict" part.

> In the late 1990's Eric Raymond and others developed the term "open source" as a more business friendly term than "free software", with a more inclusive meaning where licenses that were not as strict about the passing on of modifications would also quality for the term.

[0] https://www.freeopensourcesoftware.org/index.php?title=Eric_...

[anikom15]

No, open source was definitely more leaning toward the GPLv2 than the BSD-style licenses.

[apgwoz]

…until businesses decided that the GPLv2 was legally risky and businesses started to avoid it.

[graemep]

Software businesses. Other businesses do not care.

[anikom15]

This is FUD. Some businesses may have a policy of not using GPL software, but all the major enterprises, including Microsoft and Apple, use GPL software.

[apgwoz]

My comment wasn’t written with enough precision.

Using GPL software: yes. Totally fine. The rise of Linux and all that jazz.

Incorporating any part of GPL software _into_ other products? Pretty much doesn’t happen. Every company I’ve ever worked for has said “do not bring LGPL or GPL software into the codebase.” When it comes to commercial software, be it cloud based, or downloadable, you’re not going to find much that tries to incorporate GPL stuff. You just won’t.

[anikom15]

Of course proprietary software products aren’t going to bring in the GPL. They literally can’t.

[hyperjeff]

Apple often goes to great lengths to avoid anything GPL.

[ksec]

>The original stance of the open source crowd.....

>The proliferation of "give everything away for free" MIT/BSD/Apache licenses...

Interesting how the world have changed. The so called GPL preference, or GPL > GPLv3 > AGPL among Open source crowd is a recent thing. Arguably in the last 15 to 20 years. Both BSD and MIT dates back before GPL. And you will see far more people prefer BSD and MIT in the 90s and 00s.

I have also long argued that the license preference among generation has somewhat a linkage to political shift in spectrum. Likely to do with Tech, now known as Big tech taking advantage. And it used to be very cool if your OSS project get used by a big company, until it is not.

[bitwize]

> Interesting how the world have changed. The so called GPL preference, or GPL > GPLv3 > AGPL among Open source crowd is a recent thing. Arguably in the last 15 to 20 years. Both BSD and MIT dates back before GPL. And you will see far more people prefer BSD and MIT in the 90s and 00s.

Back when I was getting started in the mid-90s, GPL and LGPL were kind of the default. BSD and MIT were used for certain projects, like the BSDs and X11 of course, but the goal back then was to build up a large library of open source (then, "free software") as viable alternatives to proprietary software, and the GPL was the easiest way to do that and ensure it remained free.

It was the rise of Rails, and the attracting of commercial programmers and startup bros to open source in the '00s, that motivated the historical preference for BSD and MIT licenses.

[frumplestlatz]

As a BSD developer from the 90s, that does not reflect my experience.

We wanted our code to be as widely used as possible. It’s really not any more complicated than that.

There was always tension between the folks that shipped GPL software and folks that shipped BSD/MIT software, but the dividing line was not whether or not we were “commercial programmers and startup bros”.

It has always come down to questions of what we believe freedom to mean, how we wanted to contribute utility to the world, and whether we saw the use of our software in commercial projects as a loss to ourselves in a zero sum game.

The rise of “software as a service” has changed that calculus for some, and disadvantaged those that sought to build commercial service entities around their open source software. In the areas that I work, it’s made no material difference.

As for Ruby on Rails, I think it’s outsized presence on hacker news might have given you an inaccurate picture of its influence on the broader open source ecosystem.

[orthoxerox]

> There was always tension between the folks that shipped GPL software and folks that shipped BSD/MIT software, but the dividing line was not whether or not we were “commercial programmers and startup bros”.

> It has always come down to questions of what we believe freedom to mean, how we wanted to contribute utility to the world, and whether we saw the use of our software in commercial projects as a loss to ourselves in a zero sum game.

I remember the shitstorm Zed Shaw caused when he built something on top of MIT/BSD-licensed libraries and released it under GPL. "B-but that's against the spirit of the license!" people said.

[frumplestlatz]

This is a very old argument, rooted in differences in how one believes OSS cooperation is best fostered; through social norms and practices, and/or through legal fiat.

BSD/MIT authors see most proprietary use as a feature — it can drive adoption and contributions that wouldn’t exist otherwise, while generally not directly competing with the original project in the open source commons.

It is considered an opportunity to leverage social mechanisms to garner support and contribution that would otherwise not be available.

GPL relicensing is different, in that it creates a direct rival open source commons with inescapable one-way asymmetry.

The license permits it, but since BSD/MIT authors tend to prefer social norms over legal fiat to sustain cooperation, they don’t see hypocrisy in objecting (even if GPL advocates do).

(I’ve tried to be as measured as possible, but obviously, I fall on one side of the debate, and I’m sure that my point of view leaks through.)

[bigstrat2003]

> It has always come down to questions of what we believe freedom to mean, how we wanted to contribute utility to the world, and whether we saw the use of our software in commercial projects as a loss to ourselves in a zero sum game.

Very well put! I personally believe that "freedom" must include the freedom to do anything you want with the software, even close off your fork if you choose. And I do not believe that a commercial project using my software harms anyone at all, as my project is still there, still available for all. Accordingly, I have always believed in and used permissive licenses. It has nothing to do with corporate profits, and I find it vexing that people make that bad faith assumption in discussions such as this.

[popalchemist]

Yep. This tinfoil conspiracy theory about these licenses being a scheme by VC backed startups is insane. Only a relative newbie to the scene could fall for that narrative. If you've been on the scene for 15-20 years, you know this is just the ethos, and these various licenses arose out of specific needs, but not opposing ideologies.

[keysdev]

> And it used to be very cool if your OSS project get used by a big company, until it is not.

Working for big companies used to cool, but now they are just all EvilCorps.

[sterlind]

I think the companies themselves changed. Google used to be much less evil.

[watwut]

> And it used to be very cool if your OSS project get used by a big company, until it is not.

Yup. Open source advocates would brag about adoption in enterprise. They would incessantly argue about how safe open source is compared to close source and would constantly complain about stupid managers not allowing open source.

Not all of them, but that stance was predominant when I was younger.

[nicoburns]

> The proliferation of "give everything away for free" MIT/BSD/Apache licenses seems to me to have been an intentional campaign by corporate interests to undermine free software ideals

Is it not because corporations started funding open source projects in a big way (multiple billions of dollars a year big), and they fund projects that have licenses that they can use in their commercial projects.

To me that's a sign of the success of Open Source rather than the opposite.

[tensor]

As someone in the MIT/BSD/Apache camp, no, for me it has nothing to do with corporate interests. When I release code for free I'm doing it altruistically, and to me MIT/BSD/Apache has the most impact as it can be used in the most places now and into the future.

[gr4vityWall]

> The original stance of the open source crowd was more along the lines of the GPL -> GPLv3 -> AGPL, which expressly prevents this kind of thing.

Expanding on this, the Free Software movement always focused on freedom for users - which, in a world where copyright applies to computer programs, ultimately leads to the licenses you listed to repurpose it.

The Open Source movement usually tries to advocate for open-source as the best development model. As in, writing it in the open and contributing with other people will result in objectively better software in the long term. Others treated it (when the term was coined) as a marketing term for Free Software, making it more palatable to businesses whose people running it don't want to talk about ethics too much.

https://www.gnu.org/philosophy/open-source-misses-the-point....

[jen20]

> The original stance of the open source crowd was more along the lines of the GPL -> GPLv3 -> AGPL, which expressly prevents this kind of thing.

s/open source/free software/

None of those licenses prevent Amazon-style freeloading though.

[account42]

Amazon isn't freeloading anymore than users hosting the software themselves. They are providing a service and getting paid for that. This is only a problem for the original creator because they would rather get paid for the service himself but that desired vendor lock in was never something compatible with an open source license that ensures user freedom.

[Zambyte]

The original stance of open source is to cater to "free-riding" businesses. That's like, why the term "open source" even exists. You're thinking of the "free software" crowd.

[jibal]

The MIT and BSD licenses predated the GPL. People have a choice as to which ethic to follow ... it's not the result of a corporate conspiracy. (And I'm a social democrat, not a corporate simp.)

[swiftcoder]

I'm not so clear the choice was made consciously. There's a big swing away from the GPL and towards MIT/BSD around the time that Apple starts adopting a bunch of open-source projects for inclusion in MacOS X, and it accelerates when various big companies announced that they would be forbidding GPLv3 adoption. Fast forward to the cloud provider era, and basically no new software is being placed under the GPL (at least in part because Amazon/Google/Facebook/etc are predicating contributions on being GPL-free)

[drob518]

The problem with GPL was "tainting." It was never clear in what cases you could use GPL without it dragging all your code into being freely available. LGPL was supposed to help with that; AGPL made it even worse. The lawyers were terribly confused and recommended you just not use anything with "GPL" in the license.

The reason why MIT/BSD licenses flourished is that they were easy to understand. As long as you didn't sue the original author or try to claim the code was written by you, you were free to do almost anything with it, including mixing it in with other for-profit code.

Whether that's an abomination or a blessing depends on your corporate vs. free software politics.

[duskwuff]

GPL also gets incredibly ambiguous when it's applied to anything that isn't software written in a language like C for a personal computer. (What does "linking" even mean with respect to a Javascript library, for instance?)

[account42]

It's only unclear if you are trying to skirt the spirit of the GPL.

People who aren't trying to get freebies from he commons without paying back never had a problem with the GPL.

[pabs3]

Just think about the goals of the GPL (software freedom for users), and its easy to see what you should do. Make it practical for a user to obtain and modify the source form (non-minified, non-transpiled, non-concatenated TypeScript/JavaScript/etc) of the library, build the version ran by the web browser, and replace the original with their modified version. Source maps can make part of that easier too. Progressive enhancement helps. Clean frontend/backend separation helps.

[duskwuff]

The "goals" of the license are irrelevant. What matters from a legal perspective is the text of the license, and every time that text sets legal conditions which reference specific technologies like "linking", "object code", or "interface definition files" which don't apply to all programming languages, a lawyer's blood pressure rises.

[notpushkin]

> For example, Corresponding Source includes interface definition files associated with source files for the work, and the source code for shared libraries and dynamically linked subprograms that the work is specifically designed to require, such as by intimate data communication or control flow between those subprograms and other parts of the work.

So I think it would cover pretty much any JS library usage, except maybe if you just drop in a widget and don’t interact with it in any way.

[muragekibicho]

Exactly. I was building a WASM lib using Emscripten and this was the exact question I could not answer

[rincebrain]

(All personal opinion, etc.)

I'm not actually sure what a better way to square the circle of not making the large entities that have developed a weird patronage relationship with open source projects run away while also avoiding the kinds of problem that the GPLv3 and AGPL are hoping to deal with, would be. Limiting the virality scope might be beneficial there, but I'm not sure how you would word that in a way that's not gameable.

It feels like we've wound up in a weird position where because so many GPLv2 projects moved to GPLv3, companies were startled into paying attention to the risks involved in a new license with open questions about how it would shake out in actual courts, as well as being jolted to the very real possibility it could happen again, and took the path of risk reduction by moving toward platforms where that couldn't happen.

You might compare it to everyone pointing to Solaris's source closing as a reason to not trust Oracle about MySQL's license remaining GPLv2. (As it turned out, so far, they haven't changed the license, but there was certainly a lot of fearmongering about that at the time.)

So I think I agree that it's not so much a coordinated effort to steer anything as the direct effects of companies avoiding funding that space, as well as the knock-on effect that anyone whose goals involve large companies using their product and leveraging that avoids picking a license that precludes that in turn.

[pferde]

The creation of those licenses, maybe. Their mass popularization, and the pooh-poohing of GPL licenses that often goes with it in related discussions, is much more recent.