Single vendor is the new proprietary | Hacker News Reader

Single vendor is the new proprietary

opensource.net

141 pointsyarapavan92 comments

> You should be on board if you want Open Source to win against proprietary software. But those companies are still doing what is, essentially, proprietary software: like the proprietary software companies of the 80s, they very much consider the software being produced as their exclusive property. They still intend to capture all the value that derives from it. And thanks to copyright aggregation or permissive licensing, they still can change the license any time they want. So it’s still proprietary: they just choose, for now, to release their software under an Open Source license.

This brings to mind two questions: why does open source need to "win" (why can't there be multiple options) and re: things being produced as their exclusive property, what is the issue with this? They did the work to make the thing, therefore it's their property (unless they choose to release it otherwise).

Sadly, a lot of the arguments I hear around OSS sound like the "you didn't build those roads" argument when they should be "thank you for making your work accessible to me." It's no surprise that most OSS work gets abandoned due to developer burnout when "open source" is often misinterpreted as "100% free for me to do literally anything I want, whenever I want, and you're evil if you disagree with my entitlement to your efforts."

The ideology around OSS has serious NPD vibes. It's worth people revisiting Rich Hickey's "Open Source is Not About You" [1].

[1] https://gist.github.com/richhickey/1563cddea1002958f96e7ba95...

The author is not so much against proprietary software, as bait-and-switch tactics that present software as open source.

> Anyone who truly thinks that software developed by a diverse set of actors working in an open collaboration is not better should just adopt the proprietary model. But they should be honest about it.

I don’t see how it’s bait and switch. You can still used and develop against the last open version. In fact, we’ve seen recent examples of that in Terraform and Redis.

Open source does not mean you are entitled to all future improvements unless the license says you are, like the AGPL.

Personally, I’m grateful these project made their source open for so long. The alternative would not be AGPL terraform, it would be completely closed terraform.

It probably feels like bait and switch for the folks that created or used lots of open source in the time period where there was less opportunity/temptation to move the license from GPL/BSD/MIT to something proprietary. In our minds, we attached some altruistic intent to those licenses. Things have changed now, of course.

Why do the actions of HashiCorp and company have any bearing whatsoever on your decision to release and use unrelated open source software?

I'm not trying to rationalize my emotional reactions and say they make logical sense. Just explaining why they are there.

But, if there was a long time period where almost no open source software went commercial...you would be conditioned to not worry about the risk of depending on it. Then when things change, you feel differently. I don't know why that's controversial.

I’d also go as far as saying most of the value added to these open core projects is contributed by employees of the companies themselves. The fact that they do it out in the open with the possibility of someone else forking it is a plus, we shouldn’t want this to be solely proprietary. Hashicorp paid for a lot Terraform’s development for instance. Microsoft is paying top computer scientists to develop an editor

The bait and switch is the community and social elements, not the code.

Inviting the local community to come and improve your space that you opened up for everyone to exist in, and then suddenly locking the doors is both entirely legally justified and a crappy thing to do.

This analogy falls flat because it makes it sound like the community is locked out of the improvements they made to your space. But there's no scarcity here: once HashiCorp locked the doors all of the improvements that the community made were still there, and the community could and did organize a new space for continued improvements.

It’s not perfect, I agree, because code is trivially copied (unlike material things). But like I said, this is not about the code.

The wider point about subverting expectations still stands. Yes, we re-organised into tofu/valkey/whatever but that’s come at a _cost_. It’s perfectly reasonable to say that this is open source working properly, the license lets them do this etc. I think it’s also perfectly reasonable to feel a little bit embittered towards a copyright owner pulling the shutters up and forcing everyone to move.

The term "bait and switch" infers intent as well though. The community was invited to come and improve the space, but then there's this group that unexpectedly starts to monopolize that space in a way that destroys its viability. So controls need to be put in place, or eventually the space has to be totally abandoned or shut down.

I think this point of view is greedy. FOSS is about sharing code, not enslaving other developers or organizations indefinitely into the future.

IMHO that is because the argument is generally about free vs open source, when really what people often want is more of an open and reciprocating community.

The ability for multiple vendors to coexist as peers is a strong signal that it is possible for anyone to come and both give and take to the level of their desires and needs (within generally fair ground rules).

Foundations like Apache will often push for incubated projects to be maintained by participants from more than 1-2 companies before they will accept them as full-fledge, "graduated" projects.

Commercial vendors may also change their terms from time to time, such as at contract renewal. However when you have a public-facing project, you may be affecting more than than your relationship with your individual customers, but relationships with the community as a whole.

The multi-vendor forks of Terraform and Redis do illustrate these other vendors see tremendous value in that community, to the point where they'll try to move the existing community over to a new project. And why not? Companies like Amazon doesn't rely on renting out key/value data stores as their primary revenue source, but a community around that data store makes the corresponding product much more valuable.

This is why creating a company around making an open source product is so hard - it is easy (via maintained control) to have a symbiotic relationship with the project you created and corresponding community turn parasitic.

> Personally, I’m grateful these project made their source open for so long. The alternative would not be AGPL terraform, it would be completely closed terraform.

Projects like Terraform do not work without massive community involvement in building providers and providing guidance and samples around manifests. In some cases, products have even built API around the need to write a Terraform provider. That sort of community growth and involvement around a closed source platform is exceedingly rare. The alternative to an open source Terraform would have most likely have been a failed Terraform project.

For me, neither side needs to 'win' but people should understand the facts on ground when they adopt a project for their needs. "Open Source" isn't the full story, you need to consider the motivations or potential motivations of a project's backers.

I think there's a real grievance to be had by projects that were formerly open source, solicited contributions from non-paid volunteers, and then change licensing models. Maybe those people were ill-informed, and shouldn't have contributed in the first place, and this article is trying to publicize the ramifications of a projects stewardship.

> I think there's a real grievance to be had by projects that were formerly open source, solicited contributions from non-paid volunteers, and then change licensing models.

It's definitely possible to take advantage of volunteerism, but in this case I don't see the real issue. The project remains just as open as it ever was, even upon licence change. It can be forked, and the project might change name, but that's all. It's not the nicest outcome, but it's the same as if a sole owner just decides to stop working on a project. The only thing to do is fork and move on.

> I think there's a real grievance to be had by projects that were formerly open source, solicited contributions from non-paid volunteers, and then change licensing models.

I think that's the crux of the frustration and fair. But at the same time, the previous version before the license change can still be forked/the contributions are intact so is there a tangible loss or just an emotional one?

This can be a very real loss. Third party contributions are driven by specific needs. When the license changes too much, those third parties can either pay for a fork (with money/time investment) and improve their own modifications or get the enhancements in the upstream product without any improvements of their own. There is a middle ground with paid source code access, but I don't see a lot of commercial offerings do that.

Wrt winning: Open is the least worst way of doing things.

With the xz kerfuffle people say "aha! So much for your security now!" when the alternative is someone tampering with code on an unsecured ftp serve i.e. it's the devil you can see

For sure, I think that last part is underestimated but powerful. When the SolarWinds or similar attacks occured, we had to rely on investigation reports put out via PR teams and such about how such an attack occured, what processes missed it, etc. With the xz attack, every step of the attack was out in the open for everyone to analyze. Or at least the vast majority, if there were private emails exchanged between "Jia Tan" and the original maintainer, those we can't recover. But we can see pretty much every other aspect of the attack.

> re: things being produced as their exclusive property, what is the issue with this? They did the work to make the thing, therefore it's their property (unless they choose to release it otherwise).

the issue is that they also want the benefit of not being ignored, so they claim to be less proprietary than they actually are while trying to build mindshare

> the issue is that they also want the benefit of not being ignored

So in order to get any interest in your project, you have to be open source (and align with the ideology absolutely)?

I don’t think this was intended to be a normative statement. Just an observation that “open source as marketing” seems to work, so people use it despite it not actually being true of their project.

you don't have to; you could seek interest among fans of proprietary software instead, using whatever marketing techniques work there. apple has had a fair amount of success with that approach e.g.

In order to get interest in your product among people who believe in the mission of OSS, you do actually have to be OSS.

I could agree if we were talking about something unrelated that OSS folks also happen to like but yeah, your product does have to not have meat if you want buy-in from vegetarians.

not gonna lie this sounds like a borderline cultish view on things (only OSI-approved open source projects should get recognition, rest should be damned to obscurity and ridicule)

having the source available is a huge benefit compared to proprietary software as you can view, introspect and patch your tool - having a permissive licence is a "nice-to-have", not an entitlement.

i didn't say anything about OSI approval, or even about "should". the fact is that a lot of people (myself included) have significantly less interest in a project if it is proprietary, and especially if it looks like the proprietor is going to exert control over it as a means of extracting value. and other people are fine with it, and might even prefer a project controlled by a company they trust to do a good job. and there's nothing wrong with either one! live and let live, with nothing to ridicule either way.

what is wrong is when companies entice the people who would otherwise just have ignored them by claiming to be open, and then turn around and try to exert control once they already have mindshare. that makes people feel baited and switched and leads to a very natural amount of resentment.

> why does open source need to "win"

Open source does not need to win.

But your ability to be in control of your computer needs to be preserved (restored?). A proprietary fridge cannot control your diet, while a proprietary App Store can control what software you install on YOUR phone (unless you live in EU, hello DMA!). The tail wagging the dog, so to speak. Proprietary software has also been shown to break user workflows or remove functions in an update while leaving users with no choice whatsoever.

One alternative to having open source win is to ensure software comes with a robust warranty and other assurances you expect from the things you buy (like the ability to resell it etc.). EU's CRA will make software vulnerabilities in WiFi routers covered by warranty, for example. But that's just a first step.

You can also ensure robust and interoperable data storage options. For example, https://obsidian.md/ stores all notes in Markdown, not holding the data hostage in case users will not like how future versions will work. GDPR actually has a provision for data portability (Art. 20), but it does not seem to have a requisite effect on the industry yet.

And until the above issues are solved, open source remains the best way to ensure that a software tail cannot be wagging your computer dog.

From the OSI's old history page [0]:

> The conferees decided it was time to dump the moralizing and confrontational attitude that had been associated with "free software" in the past and sell the idea strictly on the same pragmatic, business-case grounds that had motivated Netscape. They brainstormed about tactics and a new label. ... A month later ... the participants voted to promote the use of the term 'open source', and agreed to adopt with it the new rhetoric of pragmatism and market-friendliness that Raymond had been developing.

The entire point of Open Source as a concept was to be more flexible and business-friendly than Stallman's Free Software movement. It's both hilarious and frustrating to see people today treat the OSI licenses as some sort of divinely appointed canon of morally good options when the entire concept exists to get away from moralizing.

(I should clarify that I actually am more and more sympathetic to Stallman's vision, but I think that if we're going to moralize we should go all the way, not put a halo around the intensely and explicitly pragmatic open source principles.)

[0] http://web.archive.org/web/20071115150105/https://opensource...

>re: things being produced as their exclusive property, what is the issue with this? They did the work to make the thing, therefore it's their property

The more the world is built out of "other people's property", the less freedom you have. And the specific property-like law Congress chose to make people pay for software is incredibly strict within the narrow range of its scope - i.e. perfect if you want literally anything to do with technology to become a coercive monopoly.

> why does open source need to "win" (why can't there be multiple options) and re: things being produced as their exclusive property, what is the issue with this?

Part of core infrastructure should be in the commons (consider the roads) to lower friction. There are many road builders but to a great extent anybody can drive on any road with any cargo without asking for permission. And that being the case unlocks more innovation.

> It's worth people revisiting Rich Hickey's "Open Source is Not About You".

Thank you for the reference to Rich Hickey's note. I used to subscribe to his view more or less completely, which is essentially that open source creators owe their users nothing and that if you want something fixed it's your responsibility to do it.

Over time, however, I have come around to think that Richard Stallman put his finger on the core issue, which is user rights regarding the code: to understand the code, to make changes to it, to make copies, and to propagate changed versions. [0, 1] This is different and more fundamental than the question of the obligations between members of the open-source community, which are much more amorphous than licenses for code.

In the original GNU manifesto, Richard Stallman made a prediction that succeeded beyond all expectation.

  Once GNU is written, everyone will be able to obtain good system software free, 
  just like air.

  This means much more than just saving everyone the price of a Unix license. It 
  means that much wasteful duplication of system programming effort will be
  avoided. This effort can go instead into advancing the state of the art.

Not all code needs to be open source. But conferring user rights like the four freedoms on the basic building blocks of software system has unleashed enormous human creativity. I want to work on those building blocks and I'm willing to trade that privilege for other things like making more money. In doing do I acknowledge we do owe users rights to use the code as they please. And I want to live in a society where those rights are upheld, at least for certain types of software.

[0] https://www.gnu.org/gnu/manifesto.html

[1] https://www.gnu.org/philosophy/free-sw.en.html

The author completely neglects the "non-open-source/open collaboration" quadrant as if it didn't exist. Best examples are games with easily accessible modding communities. While development is centralised and you definitely don't have any kind of open source licence, collaboration is generally open in terms of releasing mods to the game.

They also bash the Commons Clause purely using the definition (zero mention on why they think the restrictions are bad, it's just handled with "not OSI" and that's it)

Of course this position can be understood better when you look at who are sponsoring this organisation. (in a short way, opensource.net is an OSI front, and OSI is lobbied heavily by the software industry)

OSI isn't the software industry. It's sponsored by Google, amazon then ms via github in order of sponsor donations. It's financed by the hyperscalars. https://opensource.org/sponsors

This is such a crazy retelling of history that it's hard to take seriously. Open Source was not some brainchild of megacorps but very much so a mission of the wider software industry to band together to have an alternative to hugely expensive proprietary software vendors -- namely Microsoft and Oracle, and to prevent people from imposing their will on users via software in a manner that users had no possible redress from.

OSI was founded by Perens and Raymond and very much represents the interests of the wider software industry. You're free to argue that at some point they have stopped in recent but Open Source has meant the same thing for going on 30 years.

OSI is from 1998 so it’s much younger than the open source movement. BSD was approaching 20 and Linux was what 7 years old at that point?

The free software movement is for sure older than that but I can't find any verifiable source of "open source" being used in its current form earlier than the 90's.

The term still predates OSI, but more importantly than the name than what was going on.

The specific transition from sharing as public-domain software vs free, open, shared, and permissivly licenced software dates to ~1953 with A-2 being provided for free and requesting updates be sent back to UNIVAC. https://en.wikipedia.org/wiki/A-0_System Others might point to IBM SHARE from 1955 as being closer: https://en.wikipedia.org/wiki/SHARE_(computing)

Followed by many similar arrangements before the term open source was coined with OSI only showing up much later.

No, because it's a name that was made up then to try to rebrand something that had already existed for decades, but which has a name that those guys didn't like.

My memory from the time was that the OSI was mostly about trying to characterise the various licenses that were floating around, and about legitimising the use of freeish software in industry.

Today the majority of free software that I use seems to be Apache/BSD/MIT - and GPL for Linux - but back in the 90s everyone seemed to make up their own “free” license, and there was a lot of confusion around GPL. My recollection is that OSI was trying to create some clarity around which licenses really were “free”, and which weren’t.

I was just a passive observer, I was not involved in any way, I wasn’t a fan of some of the players, and I even thought calling it OSI was pretentious - but nevertheless they made a dent in the universe by popularising the term “open source”. OSI was far more than a rebranding because it enabled us to speak about GPL - which imposes conditions on use - and free licenses like BSD in the same breath, while excluding what we now call “source available” licenses.

So despite my misgivings, I think the OSI deserves credit. They stood up and did something meaningful, and helped move the entire software industry forward.

Because of this, I think calling it a rebranding is totally unjustified and unfair.

GPL does not impose conditions on use.

Sure it does. If you modify GPL code, you have to give your changes back in some circumstances.

If you modify it and don’t give your changes back, you can be sued.

> Open Source was not some brainchild of megacorps but very much so a mission of the wider software industry to band together to have an alternative to hugely expensive proprietary software vendors

No, you are thinking of Free Software and copyleft, which predates Open Source.

Open Source was invented as an alternative to Free Software.

Please reread my post, I didn't mention anything historical. I never said who founded it nor who started it. Your arguing against a straw man argument you made up. I am saying today, OSI does not represent the software industry and that given sponsorship bias it's likely to represent the views of those companies that pay OSI.

Bloomberg, Capital One, Cisco, Intel, Meta and Red Hat all are sponsoring at the same level as GitHub. FerretDB, Perforce, Tidelift and Salesforce join Microsoft at the next level down.

That's a much more diverse crowd of sponsors than I thought I'd see based on your comment.

Note that it's subscription/service games that often have the most contentious relation with their modding scenes, as they tend to see the mods more as competition. On the other hand, people are generally more willing to develop mods for software sold with perpetual licenses, as the chances of no longer having access to the thing you are supporting is much lower than with e.g. $XXXXX/month cloud services.

Very good point, not only in live service games... I've seen a thing or two and it's usually obvious how developers treat the modding community. In some games they are viewed as an asset (people mod my game, they make it different, this attracts more people to buy my game) and in other games, modders are the scum who always want to cheat/commit "fraud" by not wanting to buy all the microtransaction slop/etc.

Even for perpetual license games, if the game has paid DLC the developers often see mods and cheatengine as a threat to their revenues. Capcom is on the record that this is part of why they put DRM and anticheat into their single-player games now, because otherwise you can get items and cosmetics with cheatengine instead of buying them off Steam/PSN/XBL

Modding games seems like a legal grey area. Permission hasn't been granted, but also, game developers already have your money and they know a lot of players would be unhappy if they tried to stop the modding.

I wouldnt be surprised if most game studio didn't enjoy it/try to promote it in maybe indirect ways. Skyrim is an old game, great game, but an old game. However modding keeps it fresh and interesting and I would imagine it probably is a big reason why they still get sales on it. While it isn't a huge money maker any more, they are still selling copies just for people to play with the mods, extending the timeline of how much revenue the game will create over it's total lifetime.

Single-vendor is not proprietary because I can fork VSCode and I can't fork Microsoft Word.

Single-vendor open source is the balance some companies have found between sharing their software with the community and capturing the value of their employees' labor. It's less free than openly developed FOSS and more free than proprietary software. It's unrealistic to expect all software to be openly developed FOSS with today's economics; the hundreds of thousands of contributors to single-vendor open source projects all need rent money, and you can't build a business on providing the open-source backend for AWS managed services.

Companies will move up and down the freedom gradient depending on their needs at any given time. Sometimes they do it well, and sometimes they handle it in a kludgy and myopic way (I'm looking at you, HashiCorp). LinkedIn open-sourced Kafka, and Elastic restricted their license for ElasticSearch. Software doesn't always go from "more free" to "less free."

You can't forkvscode...

https://github.com/microsoft/vscode/forks

27,000 people seem to have done so.

That repo is not actually VS Code. What you get when you download compiled VS Code releases includes proprietary, closed-source components. If you build “VS Code” from OSS components, you do not end up with VS Code.

Chromium is still useful but it's not Chrome.

Most of the differences are configuration though, or at least I didn’t notice any significant differences other than that using an unbranded build

So you're saying Microsoft should share their brand with you?

No, and I have absolutely no idea what you’re talking about.

microsoft/vscode is not the same as the VS Code you download from MS. The latter has a bunch of DRM stuff that many MS plugins require.

Cursor is an AI first fork of VS Code

Thank you, downvoters, for freeing me from ever taking votes here seriously. Downvote a man for being right? About an on topic technical matter? In a technical forum?

Thanks ;)

> I can fork VSCode

Well, sort of. You lose a whole bunch of stuff (e.g., Pylance) by doing so.

If it is, then Open Source has won, right? And we can disband the OSI, right?

I suspect that's not the message they intended to send.

(I read the article, btw.)

As the author of source available software, this article merely seems to be screeching from big players that they can't exploit some software anymore.

I'm not following; how does an onslaught of fake FOSS mean FOSS has won? If anything it sounds like it needs more defending than ever.

How is fake FOSS different from FOSS ?

Is fake kindness different from real kindness?

It means that the expectation is that the source is open. The default used to be closed source, so in essence, Open Source has won.

Personally, I don't see source available as a bad thing. [1]

[1]: https://gavinhoward.com/2023/12/is-source-available-really-t...

Nobody is discussing FOSS here, only OSS. The only solution to capitalistic shenanigans is the software commune, OSS isn't a commune, FOSS is. OSS is, essentially, a software cartel disguised as a philanthropic enterprise.

Advocates of OSS want to live in a world where bigcorps and little guys can coexist, this has always been a pipe dream, since day one. FOSS locks the bigcorps out of the discussion, or rather, forces them to compete on the same playing field as the little guy, which amounts to the same thing.

So you see these articles come out every now and again bemoaning capitalistic greed. Communist utopia is right there bro. Just open your eyes and stop letting bigcorps eat your lunch every couple of years or so.

I thought the article was arguing that say Redis was while “open” the codebase is owned by one company and hence that’s no different to proprietary.

To which I say, yeah we know.

I mean I can fork Linux kernel tonight and announce my big new plans, and there will be zero installs. Big FOSS software is there because they built a reputation of constant quality over years (as successful proprietary companies should and sometimes do)

That social capital is immensely hard to compete against (the Hashicorp forking might manage?)

And that is what “the big players” are trying to exploit - that for a decade they have led / organised / helped / whatever the development of this software and now they want to monetise

It’s also why I think the “foundation” model works surprisingly well - we can trust that to keep software going so it’s a better ask for cash (even if the support model is the “best” imo)

There are tons of forks of the Linux kernel. For example, Debian ships their own, and Ubuntu, which is based on Debian, ships a different one. These are kept mostly in sync with upstream but it's completely plausible that if upstream made a change they didn't like they'd just revert it.

The actual problems are with things like systemd, which are enormous and lack standardization but get integrated into numerous parts of the system. Then you can't replace a piece of it or your piece has to integrate with all of the things the original did, but since none of the integrations follow any standards they can change at any time and people can add new ones that only work with the original and -- surprise -- your replacement is suddenly broken.

The thing is most of those changes are minimal and it is completely unrealistic for any non-megacorp to maintain their own implementation of it. The example here would be Google and Android.

Another great example is Redis - where the main competing product is now not a fork of the original but a new thing by none other than MS, as they have the cash and resources like Google to pull something like this off.

The general problem here is monorepo disease. Instead of composable components with standard interfaces, you get a single huge project which is too big for anyone to replace as a whole but also has shifting internal interdependencies that prevent you from replacing only a part of it.

The article itself is rather, hmm... Melodramatic? Not sure that's the right word, but it is close. That said, the overall idea that "open source" projects that are controlled by a single vendor have problems is true.

I think that having a for-profit company controlling an open source project is a major conflict of interest. Open source does not always result in profit. Often the opposite. And I think we've seen the results of that with all the different open source projects that have re-licensed into pretend open source licenses.

There are ways to run open source projects that support both the open source culture, and allow for for-profit companies to make a profit. But most of those ways mean allowing competition. Which is where the single vendor project conflict of interest becomes apparent. Yeah, big tech will leech off any successful project. Yes, that means less money for the "single vendor". Yes, that is not fair. But I'd say re-licensing is worse than leeching, so...

The other side effect of "single vendor" I've run into a lot, is simply that their paid options are always priced for organizations with very deep pockets. So the smaller orgs (and individual developers) that jumped on the bandwagon early because the project was open source (and they actually could jump on the bandwagon), have no chance at supporting the project. And end up have to find something else because the project stops supporting open source.

I don't see how it's a conflict of interest. The primary benefit of open source is that you can run and modify it yourself if you need to. This benefit is still there if there is a single vendor behind the project.

The risk of a single vendor project is that it's less likely to be supported in the long run. This isn't a conflict of interest though.

I somehow still don’t get what single vendor open source code is, even after reading the article.

Is it code that is open source but the license says “no forking.” And maybe the license says they are allowed to fork it and chance the license to proprietary?

Or is it code that is open source, but nobody has ever bothered to fork it. With the potential spin that sometimes code is just so specific that nobody would bother forking it because it is super tied to some platform or hardware.

The former, that just seems like looming proprietary code. The latter, I dunno, can’t blame people for implementing niche projects I guess.

I like to think of Open Governance vs Open Source on topics like, if someone adds SSO to the OSS repo, will the main vendor behind it reject the PR because it breaks their OSS core business model

I think they are basically saying an OS project with a single commercial entity as the maintainer.

If that’s the case, it seems like a very bigcorp-centered perspective. Indie devs and people who aren’t already well-connected don’t have the ability or resources to get consortiums of companies behind their projects or create open source foundations.

So I don’t know—maybe it wasn’t intended this way, but it comes across as very gatekeep-y to me. Like telling people that unless they are already established and famous enough to start a project that has tons of backing from the start, that they can’t “really” be opensource (despite a very real opensource license), meaning you as the little guy should either build your project for free with no hope of ever making a living from it, or else be excluded from the broader opensource community.

Take Hashicorp's Terraform as an example. Until recently, the code was completely open source with a permissive license. You could do exactly as you like with it. The 'single vendor' decided that was no longer in their best interests.

What's the problem? Now you have 1000's of companies that rely on this software, and if they want to continue to receive security updates, they need to comply with the new license. Of course, people are free to fork projects, and that has been done in the case with terraform, but it's something that a foundation-led project most likely wouldn't do (change the license).

There's another flavor of single vendor open source IMO. I call it 'look but don't touch open source.' Modern software projects have become some complex, and the build and dependency systems so customized, you could have the source code available, and be free to modify it, but wholly unable to do so because of the build-time complexity. A foundation-led project would, ostensibly, ensure that the entire lifecycle of the software is approachable by end users (docs, tooling, etc).

And then you have the look and "you can touch if you give away all rights for what you contribute". Contributor License Agreements where basically you contribute to a "GPL" project, but you sign away the rights to your contribution to the company, so they can still close source it any time they please while including your code.

I wonder how one would view late 90s gcc in this light? Single vendor compiler that made everybody so mad they finally forked it and basically restarted with a new team.

For what it's worth, speaking in terms of free software rather than open source (since the author of the featured article might be receptive to the moral ideals of free software rather than merely the business-pragmatic ideals of open source): if the vendor lets you download the source code of free software already in your possession without an excessive burden (one valid option being letting you download a zip/tar.gz of the entire necessary codebase; one invalid option being forcing you to copy files one at a time) then the software is free software regardless of whether the vendor accepts any third-party contributions.

FOSDEM 2024 talk by the author, https://fosdem.org/2024/schedule/event/fosdem-2024-2190-sing...

> explain the origin and value of the permissionless innovation that we currently all enjoy, and reassert the virtue of software developed in open collaboration, compared to single-vendor software

This article is dramatically simplifying the state of software markets.

>Single vendor isn’t a reasonable way to do Open Source and resist evil proprietary software. It’s just another way to do proprietary software.

>proprietary software is not evil. It’s just inferior.

Based on these statements the author would have you believe there is no value in commercial/proprietary software and we should just never develop it. All software should be open and collaborative. That is obviously silly. While open source software is great, many incredible software innovations and truly valuable software comes from proprietary companies. In fact, these companies are typically the ones that make the large open source ecosystem possible by making massive donations in developer hours as well as cash to orgs like linux foundation.

The interesting discussion is in whether commercial software should be closed source or source available with restrictions. The days of building propriety, VC backed infrastructure software with a traditional permissible license are over and likely never coming back.

yes it's puzzling how simplistic this framing is. I would expect more from the Vice-Chair of the OSI.

>> proprietary software is not evil. It’s just inferior.

> Based on these statements the author would have you believe there is no value in commercial/proprietary software and we should just never develop it.

Do you believe that there is no value in inferior products and that they should never be developed?

I certainly don't, and it's clear to me that the Carrez does not, either.

If I have the option, and can afford to do so, I will select the superior product. But if there is no option, or I cannot afford the superior product, I will accept the inferior one and be better off than if I had no product at all.

> The interesting discussion is in whether commercial software should be closed source or source available with restrictions.

Thank you!

The author carefully uses the term "proprietary software", drawing no distinction between whether it is closed source or source available, as if that distinction is totally beside the point. But for me, as someone who makes software, there is a huge distinction between those two things!

I really hate using tools that I can't read the source of. Just recently I traced some documentation on how python garbage collection works into the implementation for that particular thing in the particular version of the language that I'm using. If python were a single-vendor source-available tool, that would be a bummer and I'd be less likely to use it, but it wouldn't actually affect my work much. But if it were closed source, that would absolutely be a deal breaker for me. I need to be able to go look and see how my tools work, otherwise I'm blind.

I do agree with the author that community-driven open source is better, and I consider projects like the Linux Foundation, BSD, GNU, Apache, CNCF, etc. to be wonderful miraculous gifts. But I also worry that a distressing proportion of the most important software I use has been built on the backs of a series of absurdly under-compensated and eventually burnt-out passionate nerds, and I can't stand that. So I'm sympathetic to a model that has a more obvious (to me) path to creating software tooling that I can use without flying blind, while compensating people adequately for their work.

My quaint dinosaur view is that all of the single-vendor stuff I can think of is aimed at data farming at a scale that is itself presumptively evil. The need to do what these pieces of software do at the scale they do them is a filter for the kind of organization you get becoming involved with them.

I read this article and feel no further informed on this topic. Avoid.

> On one axis, the license used is either Open Source (as defined by the Open Source Initiative (OSI), which I would summarize as coming with all freedoms necessary to enable the permissionless innovation I mentioned earlier) or it’s not. On the other axis is the development model: it’s either developed as a commons, by a community working in open collaboration, or it’s developed (and ultimately owned) by a single entity.

Is SQLite "the new proprietary"? It seems to fit the description perfectly: very permissive license, very closed development process.

That’s why the fork that is more open for contributions is gaining momentum: https://github.com/tursodatabase/libsql

You mean the "single vendor" fork from Turso, the VC funded startup [1]? Thanks but no thanks, I'll stick with D. Richard Hipp.

[1] https://turso.tech/investors

And this is why we need copyleft.

Reply on news.ycombinator.com