[donbronson]

Apple could have a similar situation. Via the iOS SDK, developers can ask for access for a phone’s address book. These contact connections could be harvested by bad developers. They could also be sold by those developers. This is merely one example. All companies with data and APIs are vectors of attack.

[panic]

The difference is that the data is on your phone, not Apple's servers (unless you back up with iCloud, but even then I believe it's encrypted). You can decide which apps to run on your device; you have no control over what Facebook does on its servers.

[donbronson]

Definitely a difference but there is a similarity in vectors of attack: The contact address book can be taken off the phone using the Apple SDK. Then, that data can be anywhere the developer wants to store it.

[valuearb]

It can only be taken off the phone if the user explicitly gives you access to it. And it’s just contact info, not an “attack vector”.

Are you really arguing developers shouldn’t be allowed to write address book utilities on iOS?

[jclardy]

This is non-issue. Do you really want a mobile OS where third party apps can literally do nothing? No contacts, photos, camera, accelerometer, gps, microphone.

Each one of those explicitly requests the users permission. Facebook would track you across the internet just using Share buttons on webpages.

[nkozyra]

Facebook's issues stem from historical overly-permissive API access that could effectively harvest your graph via a single node's acceptance. Some small facsimile of that is represented by your example but not to the scope it was with Facebook. It's a one-to-many but not a graph. I think this is important natural constraint and a noteworthy distinction. It makes connecting data more of a crapshoot.

Apple is equipped with far more resources at present to deal with app review than Facebook has. They've traditionally been far more draconian about their own rules than has Facebook. This means something like what you describe is less likely to fall through the cracks or propagate the way abuse did with FB in the pre-2014 days.

You're 100% right that any access is effectively a vector, but that's needless reductivism. In 2018 society seems to accept some level of visibility but not unfettered visibility. This is the balance that social media and data gatekeepers will have to contend with.

That said, it doesn't take much to tip the scales these days. People tend to have a healthier skepticism toward the giants of tech these days, and if you burn a little bad PR it can have some terrible consequences. Which is why Facebook has pulled some plugs in a panic on the developer side of things in response.

As has been mentioned countless times in the wake of this kerfuffle, MySpace was the defacto social media company until one day they weren't. It's eas(ier) to introduce competition to Facebook over, say, Apple or Microsoft.

Finally, Facebook's product is, well, data. One could argue that it's a platform and data, but the former is a loss leader for the latter.

So for a number of reasons I find it unlikely that even were Apple to have a PR misstep like this that they'd face the same existential crisis.

[untog]

They could have, but that's why Apple has an app review process. It's been criticised (by me, for one) for being heavy handed and slow, but recent events have shown that there's a wisdom to it.

[donbronson]

The app review will catch some stuff but won’t be comprehensive since it relies on human judgement, right?

[untog]

No one really knows. But I wouldn't be surprised if there's a mix - I'm sure there's some automated investigation of how much data is sent, when and what that data is.

[donbronson]

Fair enough. A clever avoidance could find developers activating some evil stuff after the review process itself, though.

[valuearb]

That can be caught too. Apple review has testing automation tools, and advancing clocks and changing locations are simple to automate.

[JumpCrisscross]

> won’t be comprehensive

Something beats nothing.