← Back to context

Comment by direwolf20

11 days ago

They're not saying they violated the license, they're saying they're assholes. It may not be illegal to say you'll do something for free and then not do it, but it's assholish, especially if you said it to gain customers.

They gave code for free, under open source, but you call them assholes if they do not release more code for free. So who is the asshole here? You or them?

  • Nestle started a charity to give free baby formula to mothers in African villages. Once they all stopped producing milk, they stopped the charity and made the formula so expensive it was barely affordable.

There's no broken promise though. It's the users who decide+assume, on their own going in, that X project is good for their needs and they'll have access to future versions in a way they're comfortable with. The developers just go along with the decision+assumption, and may choose to break it at any point. They'd only be assholes if they'd explicitly promised the project would unconditionally remain Y for perpetuity, which is a bs promise nobody should listen to, cuz life.

Continued updates is not and never has been a part of FOSS, either implicitly or explicitly, you simply have a misconception. FOSS allows you to change the software. That's what it has always meant.

> say you'll do something for free

I think this is where the problem/misunderstanding is. There's no "I will do/release" in OSS unless promised explicitly. Every single release/version is "I released this version. You are free to use it". There is no implied promise for future versions.

Released software is not clawed back. Everyone is free to modify(per license) and/or use the released versions as long as they please.

Customers are the ones that continue to pay. If they continue to pay they will likely receive maintenance from the devs. If they don't, they are no longer or never have been customers.

  • It would be interesting to see if there could be a sustainable OSS model where customers are required to pay for the product, and that was the only way to get support for it as well.

    Even if the source was always provided (and even if it were GPL), any bug reports/support requests etc. would be limited to paying customers.

    I realize there is already a similar model where the product/source itself is always free and then they have a company behind it that charges for support... but in those cases they are almost always providing support/accepting bug reports for free as well. And maybe having the customer pay to receive the product itself in the first place, might motivate the developers to help more than if they were just paying for a support plan or something.

    • Well, I think this is what SchedMD do with Slurm? GPL code. You can sign up to the bug tracker & open an issue, but if you don't have a support contract they close the issue. And only those customers get advanced notice of CVEs etc. I'd expect nearly everyone who uses it in production has a support contract.

      1 reply →

    • Some software providers sell software including support, no restriction on number of deployments as well as the possibility for private modifications, then publish under a permissive licence two-three years after a release. That seems to me like a good way of doing things.

I'm noticing this argument a lot these days, and I think it stems from something I can't define - "soft" vs. "hard" or maybe "high-trust" vs "low-trust".

I always warned people that if they "buy" digital things (music, movies) it's only a license, and can be taken away. And people intellectually understand that, but don't think it'll really happen. And then years go by, and it does, and then there's outrage when Amazon changes Roald Dahl's books, or they snatch 1984 right off your kindle after you bought it.

So there's a gap between what is "allowed" and what is "expected". I find this everywhere in polite society.

Was just talking to a new engineer on my team, and he had merged some PRs, but ignored comments from reviewers. And I asked him about that, and he said "Well, they didn't block the PR with Request Changes, so I'm free to merge." So I explained that folks won't necessarily block the PR, even though they expect a response to their questions. Yes, you are allowed to merge the PR, but you'll still want to engage with the review comments.

I view open source the same way. When a company offers open source code to the community, releasing updates regularly, they are indeed allowed to just stop doing that. It's not illegal, and no one is entitled to more effort from them. But at the same time, they would be expected to engage responsibly with the community, knowing that other companies and individuals have integrated their offering, and would be left stranded. I think that's the sentiment here: you're stranding your users, and you know it. Good companies provide a nice offramp when this happens.