News for hackers. Stuff that matters.
Re:MacOS 10.12.6 Source by chungy on Thursday September 14, 2017 @08:38PM (#15253671)
It's not the position of the FSF who actually wrote the license. GPLv3 requires that users be able to install their own copies of the program onto the device. The device can allow a user-settable key in order to accomodate the requirement. No Apple keys need to be published.
Re:MacOS 10.12.6 Source by SeanLuke on Thursday September 14, 2017 @09:04PM (#15253794)
> It's not the position of the FSF who actually wrote the license.

The author of a license is not the person with the legal authority to interpret its meaning.

Re:MacOS 10.12.6 Source by thomastjeffery on Friday September 15, 2017 @12:59AM (#15254762)
So you are saying that Apple would want to enforce it in a way that is less convenient for Apple than the way the license author would enforce it?

That is just as absurd.

Re:MacOS 10.12.6 Source by tylerhou on Friday September 15, 2017 @01:36AM (#15254888)
No, he’s saying that if Apple chooses to trust the license author when they say that they interpret it in a specific way, they might get screwed if the license author changes their mind. Should that happen, a court of law would have the authority of legal interpretation.
Re:MacOS 10.12.6 Source by mseebach on Friday September 15, 2017 @03:47AM (#15255402)
No, they're saying that the people actually having to defend the position, the people with actual skin in the game, (Apple legal) believes this.

The opinion of the writer of a legal document carries only tangential weight next to the actual written words of this document.

Re:MacOS 10.12.6 Source by gsnedders on Friday September 15, 2017 @06:43AM (#15256025)
And of course Apple is going to take the conservative position until a court rules on how to interpret the license; it's just not worth the risk to them.
Re:MacOS 10.12.6 Source by crististm on Friday September 15, 2017 @09:49AM (#15257097)
How do you know what they actually believe?
Re:MacOS 10.12.6 Source by mseebach on Friday September 15, 2017 @10:00AM (#15257190)
I don't, as such. But a grandparent made a claim of their position: "It's not the position of the FSF who actually wrote the license."
Re:MacOS 10.12.6 Source by cyphar on Friday September 15, 2017 @05:13AM (#15255709)
Well, neither is Apple. And not just that, but Apple has incentive to spread misinformation about the GPLv3 (to discourage projects from using it so they can continue to use existing software without making their platform freedom-respecting).
Re:MacOS 10.12.6 Source by eridius on Thursday September 14, 2017 @08:49PM (#15253728)
User-settable keys also compromise the security of the device.

The most plausible approach is one outlined in another comment in this thread wherein you get a new key when factory-resetting the device and if you don't write it down it's gone forever, but even that approach will likely cause problems with Activation Lock. And in general it's risky to introduce stuff like this because the potential for unforeseen security compromises is very high, and the upside is very low (the number of people who'd actually take advantage of this is extremely low relative to the number of iPhone users).

Re:MacOS 10.12.6 Source by nitrogen on Friday September 15, 2017 @01:55AM (#15254955)
the number of people who'd actually take advantage of this is extremely low relative to the number of iPhone users

That's only because we live in a world where phones aren't treated like the general use computers they are. MS+OEMs and Apple fought hard for decades to create an appliance mindset around what used to be a user-replaceable component, the OS.

We can just as easily imagine and try to build a world where it's just as easy to mix and match phones with phone OSes as it was to choose DR-DOS instead of MS-DOS.

Re:MacOS 10.12.6 Source by bobcallme on Thursday September 14, 2017 @09:16PM (#15253834)
> User-settable keys also compromise the security of the device.

Not if a physical modification to the hardware, a set of button presses, a hardware token or a fuse is needed to set said keys.

> And in general it's risky to introduce stuff like this because the potential for unforeseen security compromises is very high

There is little to no risk if physical interaction with the hardware is required for setting keys. If you have physical access and enough time, all bets are off anyway.

> ...and the upside is very low (the number of people who'd actually take advantage of this is extremely low relative to the number of iPhone users).

There are many people who dream of having the freedom to roll their own firmware on a particular device and it can be done without compromising the security of said device. This was not just about iThing users (who might/not take advantage of this), this was about people spreading FUD about the tivoization part of the GPLv3. Root signing keys don't need to be shared and the only requirement is that the user can take the sources, compile them and then run the result on the target device. As long as the OEM distributing the GPLv3'ed program gives a clear path for running the compiled code on the target device, then it is not a problem.

Re:MacOS 10.12.6 Source by natch on Friday September 15, 2017 @03:10AM (#15255241)
You and others like you who take this line of argument are either ignorant of, or willfully pretend away, the abusive/spying household member problem.
Re:MacOS 10.12.6 Source by cyphar on Friday September 15, 2017 @05:06AM (#15255676)
Which is why most systems that allow you to change the boot chain or otherwise mutate the chain of trust will clear the storage on the device (such as unlocking the bootloader in Android). And on Chromebooks, you cannot easily (read: without getting a flash programmer and doing it manually) subvert the boot chain and not let the user be aware of it (that's what developer mode is used for -- it shows a scary warning if the measurement and secure boot are disabled).

The reason nobody mentions this is because the solution is so trivial there's really no point in repeating it each time.

Re:MacOS 10.12.6 Source by frenchy on Friday September 15, 2017 @03:38AM (#15255354)
A walled garden OS doesn't do much to stop the abusive/spying household member problem. They can just beat you up until your give them your password, or spy on you and catch you entering it.
Re:MacOS 10.12.6 Source by mseebach on Friday September 15, 2017 @03:50AM (#15255413)
But they can't compromise the device without the victim knowing about it. Not all abusive relationships are violent or even directly coercive like that.
Re:MacOS 10.12.6 Source by natch on Friday September 15, 2017 @12:33PM (#15258448)
It does plenty to help stop it.

So much so that in your attempt to find any toehold to criticize it, your comment had to resort to coming up with scenarios that go outside of the system and rely on human vulnerabilities.

Re:MacOS 10.12.6 Source by josteink on Friday September 15, 2017 @03:47AM (#15255401)
> User-settable keys also compromise the security of the device.

Hog-wash. Me not allowed to be in control of my own device is the ultimate compromise of my security.