But if you click on that link you will not find the word "warranty" used in any way, and in fact the bases of the case seem not to involve any warranty-like responsibility at all. So I'm confused.
The situation described in nullc's comment seem pretty horrible!
So, I'm also confused by your response. I wrote that it's not correct that the no-warranty clause of the MIT license has been declared void. And you seem to be saying that this is in conflict with the appeals judgement not talking about the warranty clause. But that's what I was saying! Nothing has been decided about the MIT license.
The lawsuit was originally dismissed without trial since the lower courts thought that the lawsuit did not have a chance of succeeding due to resting on the claim that the Bitcoin developers had a fiduciary duty to Bitcoin users (i.e. to act in the users' best interest). The appeals court thought that there's a legit case to be made for such a fiduciary duty existing. They did not say it exists, or under what circumstances it would exist. Just that there should be a trial to determine that.
You are right, but there is also a level of reality that needs to be injected: If open source can't be confident of a dismissal these kinds of claims on a summary basis it probably can't exist.
The cost of winning at trial is great enough to be ruinous especially relative to a volunteer effort.
In the case of Bitcoin we're probably better off in the sense that there are wealthy supporters willing to step up and take on costs. But there isn't anything about this that couldn't be applied more broadly to impose far ranging duties on other open source developers.
(FWIW, a fiduciary duty is far broader than just a duty to act in their best interest, it is a duty of single minded loyalty to put their interests over all others, including the fiduciary's own).
> But if you click on that link you will not find the word "warranty" used in any way, and in fact the bases of the case seem not to involve any warranty-like responsibility at all. So I'm confused.
I think the issue is that you cannot use a warranty to absolve yourself of a fiduciary responsibility. Otherwise your Bank would use a warranty to absolve themselves of responsibility over your bank account.
The legal case is not that the Bitcoin developers are liable for their software alone. It's that they might have a duty owing to the control they can exert over the Bitcoin _network_ (through their control of the "official" software repository). There are a couple of careful points to temper this judgement:
1. It's has not been decided that Bitcoin developers _do_ have some control over the Bitcoin network. The question of whether Bitcoin is centrally controlled to some extent or truly distributed cannot be summarily decided by a judge or the court of appeals. It _should_ go to trial.
2. It's not been decided that the Bitcoin developers have a fiduciary responsibility even if they do have control over the network.
3. It's not been decided whether or not Tulip could successfully argue that the Bitcoin addresses do in fact belong to them
3. It's not been decided what action, if any, the Bitcoin developers should be compelled to make if all the previous points are decided in Tulips favor.
No, this is false, because you have a direct relationship with your bank in the form of a contract, payment for services, and a direct monetary/depositor relationship.
If this analogy were closer to reality, it would be like claiming that a bank has a fiduciary duty to someone who not only doesn't have an account with them, but also put their cash money in a barrel and deliberately burned it.
Except it's not even that, because there's no bank-like structure going on here. The code itself didn't create the forms of money that followed—it merely recognizes it as valid on an ongoing basis from people who used it to perform that work themselves.
Thank you. Reading your comment, and then the link in nullc's comment, where the warranty language is included at para 110, it is clear that the warranty is not actually the issue but the disclaimer of liability.
I wonder if the license were modified to add, "if you do not agree to these terms you do not have a grant of license for this software."
Separately, when I look at the case, it seems to be a provision of service issue, so the software itself should be a red herring. Of course anything can happen in a lawsuit, even one that appears to be as illegitimate as this one.