Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Facebook 'unintentionally uploaded' 1.5M people's email contacts without consent

businessinsider.com

Post reply on HN
Print view View original

11–20 of 310 posts

Re: Facebook 'unintentionally uploaded' 1.5M people's email contacts without consent

#11

Post by AnthonyMouse »

ludston wrote: post #8
brianpgordon wrote: post #5

Earlier quoted context omitted.

Apparently Facebook is claiming that the functionality came from a separate "import contacts" feature that used to exist. But I agree; the idea that the import logic could have slipped into the login process accidentally is ludicrous. Or at least it indicates an outrageous lack of care on Facebook's part.

It requires just one developer and a couple of reviewers to make poor choices. Which begs the question, how do you structure your organisation such that a foolish developer that only barely understands the change that they are making can't write code that makes arbitrary queries to particular data sets in unapproved contexts?

Start by keeping the primary copy of the user's data on the user's own device so that the developers never have access to it to begin with. Then, if you ever have to hold a copy of the user's data, make sure it's encrypted by the client and your servers are never in possession of the plaintext.

To access the user's data, your developers should have to intentionally crack the user's password. And if they attempt to do that they should be fired.

Obviously this is not how Facebook works, but ideally it's how the thing that replaces Facebook will work.

Re: Facebook 'unintentionally uploaded' 1.5M people's email contacts without consent

#12

Post by AnthonyMouse »

darkpuma wrote: post #7
jammygit wrote: post #4

First they ask for email passwords. Then the new users assume Facebook won't comprehensively mine their emails. Then Facebook awkwardly gets caught uploading 1.5 million users' email contacts. It doesn't make sense for people to trust the service at all unless you assume one of two things: 1 - Despite all the outrage on hackernews, and the NWT stories, our neighbours down the street and family members still don't kno…

> "I really do think people have trouble assuming the worst about the intentions of others and are inclined to be trusting." I think you hit the nail on the head. Even on HN, it's not uncommon to see a few comments on each negative story about facebook accusing the media of a conspiracy against Facebook; claiming that the media is wrongly maligning Facebook who is merely the unfortunate victim of a series of coincide…

I think you see that more with other tech companies.

There is a certain amount of anti-silicon valley sentiment in the media and as a result there are a lot of stories maligning tech companies in ways that aren't always fair. Especially when the media companies are campaigning for some kind of problematic legislation that the tech companies are on the other side of and so will take any excuse to try to make them look bad.

Then there's Facebook, about which nobody has time to write a story maligning them unfairly because there is never that long between any of the stories maligning them fairly.

Re: Facebook 'unintentionally uploaded' 1.5M people's email contacts without consent

#13

Post by emerongi »

AnthonyMouse wrote: post #11
ludston wrote: post #8

Earlier quoted context omitted.

It requires just one developer and a couple of reviewers to make poor choices. Which begs the question, how do you structure your organisation such that a foolish developer that only barely understands the change that they are making can't write code that makes arbitrary queries to particular data sets in unapproved contexts?

Start by keeping the primary copy of the user's data on the user's own device so that the developers never have access to it to begin with. Then, if you ever have to hold a copy of the user's data, make sure it's encrypted by the client and your servers are never in possession of the plaintext. To access the user's data, your developers should have to intentionally crack the user's password. And if they attempt to do…

There should be a name for this sort of software design. It's not just encrypted/privacy-oriented or whatever. It's a software design with a clear contract on who owns the data: the user.

E.g. Google Drive, which claims to take privacy seriously and also encrypts your data. But the data is not encrypted with a secret unknown to the server. How should my family members differentiate between the encryption Google claims it has and client-side encryption? For them it's all the same.

Maybe we need some commonly understandable name that a regular user can look at and know that this software is data-agnostic.

Re: Facebook 'unintentionally uploaded' 1.5M people's email contacts without consent

#14

Post by sonnyblarney »

AnthonyMouse wrote: post #11
ludston wrote: post #8

Earlier quoted context omitted.

It requires just one developer and a couple of reviewers to make poor choices. Which begs the question, how do you structure your organisation such that a foolish developer that only barely understands the change that they are making can't write code that makes arbitrary queries to particular data sets in unapproved contexts?

Start by keeping the primary copy of the user's data on the user's own device so that the developers never have access to it to begin with. Then, if you ever have to hold a copy of the user's data, make sure it's encrypted by the client and your servers are never in possession of the plaintext. To access the user's data, your developers should have to intentionally crack the user's password. And if they attempt to do…

"Start by keeping the primary copy of the user's data on the user's own device"

It's an organizational policy, procedural, ethics and legal question - not a technical one.

They should have feature reviews before the code reviews. The feature review panel puts bounds on what the code can do.

Re: Facebook 'unintentionally uploaded' 1.5M people's email contacts without consent

#15

Post by brianpgordon »

sonnyblarney wrote: post #14
AnthonyMouse wrote: post #11

Earlier quoted context omitted.

Start by keeping the primary copy of the user's data on the user's own device so that the developers never have access to it to begin with. Then, if you ever have to hold a copy of the user's data, make sure it's encrypted by the client and your servers are never in possession of the plaintext. To access the user's data, your developers should have to intentionally crack the user's password. And if they attempt to do…

"Start by keeping the primary copy of the user's data on the user's own device" It's an organizational policy, procedural, ethics and legal question - not a technical one. They should have feature reviews before the code reviews. The feature review panel puts bounds on what the code can do.

I don't think you need a board to stop this specific case. It's pretty obvious that what they were doing is unacceptable. The problem must have been a pervasive culture of lack of respect for privacy at Facebook, not a single engineer who somehow just didn't know any better.

Re: Facebook 'unintentionally uploaded' 1.5M people's email contacts without consent

#16

Post by mnm1 »

jammygit wrote: post #4

First they ask for email passwords. Then the new users assume Facebook won't comprehensively mine their emails. Then Facebook awkwardly gets caught uploading 1.5 million users' email contacts. It doesn't make sense for people to trust the service at all unless you assume one of two things: 1 - Despite all the outrage on hackernews, and the NWT stories, our neighbours down the street and family members still don't kno…

Don't attribute to stupidity what can be attributed to malice. No, I didn't get that backwards.

Re: Facebook 'unintentionally uploaded' 1.5M people's email contacts without consent

#17

Post by 3xblah »

Can someone use a throwaway e-mail address to sign up for Facebook?

Once the e-mail address is validated, is there any further need for a valid e-mail address to continue using FB?

Historical fact: Going back to the days when a university address was required, if the user created her Facebook account while at university and her e-mail address later expired when she graduated, FB did not disable the account.

Unless one wants to get notifications and other FB crud via email, AFAIK there is no need for a working e-mail address to use FB.

Re: Facebook 'unintentionally uploaded' 1.5M people's email contacts without consent

#19

Post by nvssj »

3xblah wrote: post #17

Can someone use a throwaway e-mail address to sign up for Facebook? Once the e-mail address is validated, is there any further need for a valid e-mail address to continue using FB? Historical fact: Going back to the days when a university address was required, if the user created her Facebook account while at university and her e-mail address later expired when she graduated, FB did not disable the account. Unless on…

Just use a throwaway email account AND keep it? At some point they might decide to lock you out if you log in from a different place, I think it's better if you keep the email account safe.

Re: Facebook 'unintentionally uploaded' 1.5M people's email contacts without consent

#20

Post by nvssj »

jammygit wrote: post #4

First they ask for email passwords. Then the new users assume Facebook won't comprehensively mine their emails. Then Facebook awkwardly gets caught uploading 1.5 million users' email contacts. It doesn't make sense for people to trust the service at all unless you assume one of two things: 1 - Despite all the outrage on hackernews, and the NWT stories, our neighbours down the street and family members still don't kno…

>They don't care about their data privacy. I've heard this claim many times, but the people saying it often change their minds when they read more news stories.

"People don't care about a problem initially, then when it becomes graver they start to care"

So normal, expected behaviour?

Post reply on HN