Is that really much worse than the status quo, where all iMessage plaintexts are shared with Apple by default, unless every participant in the thread has disabled iCloud Backup?
> So Apple just needs to explain security & encryption to Brussels attorneys to keep iMessage in a silo, for example. (Obviously making iMessage interoperable with e.g. Discord or ICQ will compromise the integrity of the software features of iMessage.)
Are you arguing that it's impossible for Discord or ICQ to implement the same feature set as the native iMessage client?
The wording in the Act appears to not place constraints on exactly what services must be able to interop with iMessage, and it says Apple can't preference iMessage. So presumably the blue bubbles have to go. This means that when chatting in Messages now, one will not be able to know whether it's Apple alone protecting the privacy of the conversation, or whether the weakest link is a company of a single person that started this week using an OSS package to spin up a new messaging app on a Linode VM and was able to achieve parity in iOS messaging simply by virtue of this Act.
That's a very different situation than the status quo.
Similar logic applies to the App Store and payment details, but messaging is egregious because other participants in the chat can make (and potentially change) the security posture of the chat without your knowledge.
(Edit, adding reply to this part.)
> Are you arguing that it's impossible for Discord or ICQ to implement the same feature set as the native iMessage client?
No, I am saying that it's going to be trivial to implement the feature set. I fully expect there to be OSS libraries to implement the iMessage client.
The issue is that when people use iMessage and see a blue bubble, they also know something about how the encryption keys are handled & who does the handling, etc. This Act as written appears to allow any person to install an OSS iMessage server on a VM and achieve OS parity with iMessage, and Apple is prevented from indicating that others in the chat may be compromising the security of the chat. I used ICQ and Discord as examples because Apple exchanging my encryption keys with ICQ is definitely compromising a core feature of iMessage.
Realistically, the average person doesn't even know what an encryption key is and the average tech enthusiast doesn't know anything beyond "Apple handles everything".
> So presumably the blue bubbles have to go. This means that when chatting in Messages now, one will not be able to know whether it's Apple alone protecting the privacy of the conversation
Why can't iMessage to iMessage be blue bubbles with iMessage to anything else being green bubbles? Do you think the average iPhone user understands that green bubbles mean insecure rather than thinking they're "Android bubbles"?
Why are so many people convinced that consumer choice and security are mutually exclusive?
The new law removes that person's ability to even discern this much.
> Why can't iMessage to iMessage be blue bubbles with iMessage to anything else being green bubbles?
This would pretty clearly be Apple preferring its own service over other services, and is prohibited under the new law.
I should be clear here: I think this law is clumsy and will have to be fixed or loosely enforced. But even if one is in favor of these changes, it's fair to acknowledge the reality that the law basically mandates a major change in the security posture of a billion deployed devices. One may still be in favor, but it's important not to lose sight of the fact that there are in fact compromises being made.
> Why are so many people convinced that consumer choice and security are mutually exclusive?
I am 100% not convinced this is true; security and choice can coexist. But this EU law does not aim at the goal of preserving security while increasing consumer choice. Which is strange, given how a consequence of a prior EU Act was the pollution of the Web with cookie banners. (I am aware these banners are not specifically required for many use cases, which is why I said this is a consequence of their Act and not something ordered by their Act.)
I don't agree. Green bubbles indicate that messages travelled 100% within Apple's ecosystem. Indicating interoperability with a 3rd party by using blue bubbles doesn't do anything to prevent those 3rd parties from having their own green bubbles within their own ecosystem.
The entire point of the law is Apple/Google don't get to manage their own ecosystems anymore; EU regulators do that. One read of the intent of the Act is that platforms should not be able to preference their own services over those of third parties. My skim of the Act indicates that it would be in bounds for a regulator to decide that this means no blue/green message distinction, and a direct consequence of that is that anyone in a group chat could be using Facebook Messenger as a client and allowing Facebook to log all the messages.
Engineering decisions henceforth need to weigh what a specific regulator believes the law says, since the law is not very specific about many details. This is not good, especially coming from the people who caused the global cookie disclaimer deluge.
Does it really, though? Is it something that that was explicitly brought up in the drafting of this law? Or is this a doomsday scenario that chooses to interpret the law in its most extreme form?
The legislation has been approved. Let us see how regulators actually enforce it. Until the EU bans green vs. blue bubbles, this is nothing more than FUD.
We're on the same page here: lawyers are now deciding matters of technical import, and that is not a good thing.
I agree with you in general, but it is possible for this evaluation to be done by a simple read of e.g. the iOS marketing page. The Act proposes a regime where it is not possible for a chat participant to determine who holds the encryption keys for the chat. You may consider that better or worse or neither, but it is a significant change in the security posture for billions of deployed devices.
Regulator: what are the others using?
Apple: something very similar but slightly different
Regulator: and it's impossible to develop an open standard thet allows p2p encryption to be maintained?
Apple: Yes obviously in this specific field of technology that's not possible at all.
(Not the) Regulator: Sounds legit
What this Act says is that Apple Messages must interop with any other messaging system. If you spin up a VM running an iMessage-compatible server on Hetzner, Apple Messages must interop with you and cannot privilege Messages (e.g. by continuing to use blue bubbles as a differentiator). That VM may be malicious (for ex: it may log who communicated with whom when), but the Act still requires that it be placed on an even footing with Messages.
Similarly, the Act essentially says that once implemented, anyone using a phone can unintentionally be sending all of their messages to Facebook Messenger, which by law must have seamless interop with Messages. Any group chat could be logged by Facebook by virtue of one person in the chat choosing Facebook Messenger as their default messaging app.
This requirement materially changes the security posture of a billion devices currently in use. You may believe the tradeoff is worthwhile, but it's still not a free tradeoff.
If Apple prefers lock-in over innovation, it's their problem, not the users'.
That's what tech companies like to say. But from experiences with less tech-savy friends, the current way always leads people to share as much as possible. Opt-out is always harder than opt-in.
It's entirely possible that to enable compliance with this will require explicit compromises to the security of iMessage, for example by requiring key exchange with startup messaging providers. The other rules seem to prohibit Apple from describing the risks involved in such a compromise.
> * Use existing hardware and software features without competitive prejudice. E.g. NFC
This appears to say that a malicious app can present UX elements that were previously limited to the OS. Read liberally, that would mean that e.g. any app could now get biometric data by presenting a fake privilege escalation screen (e.g. FaceID/TouchID) and then capturing the results from the Secure Enclave. Is this something people really want?
> * Not preference their services. This includes CTAs in settings to encourage users to subscribe to Gatekeeper services, and ranking their own services above others in selection and advertising portals
This will likely make it harder for users to find safe/private services to use. If every offering can find its way into the default browser/App Store/etc. settings page in the OS, scam services will appear to be endorsed and therefore legitimate.
Edit: Limitation of interop is still possible, the EU is just deciding to move decision making from California engineers to Brussels attorneys. From the Act:
> The gatekeeper shall not be prevented from taking strictly necessary and proportionate measures to ensure that interoperability does not compromise the integrity of the operating system, virtual assistant, hardware or software features provided by the gatekeeper, provided that such measures are duly justified by the gatekeeper.
So Apple just needs to explain security & encryption to Brussels attorneys to keep iMessage in a silo, for example. (Obviously making iMessage interoperable with e.g. Discord or ICQ will compromise the integrity of the software features of iMessage.) I don't think this is going to increase the pace of product improvements.