Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

https://github.com/WhisperSystems/whispersystems.org/issues/...

Moxie saying "Every lawyer we've spoken to has confirmed that [having a warrant canary] would not work." Which isn't surprising. When the government tells you not to communicate something, you aren't allowed to communicate it. Not not not not communicating it isn't some clever loophole.



In that same issue you linked, there's a link to the EFF's Warrant Canary FAQ ( https://www.eff.org/deeplinks/2014/04/warrant-canary-faq ) that disagrees with what Moxie/WhisperSystems is saying.

I'd take the EFF's word over WhisperSystems'.


Note that WhisperSystems is actually concerned with the legality, because it's their own necks on the line, and they don't want to put their own neck on the line for something that their own legal counsel says won't work.

While the EFF is . . . well, they don't promise to bail you out of jail and pay all your court costs and go to jail for you. The EFF is an advocacy organization, begging people to believe in magic beans, by telling everyone that magic beans work. If one of them does and finds out it doesn't work, what is the EFF going to do then? Say "I'm sorry" really nicely?

The hired lawyers' viewpoint is easy to see. If the government orders you not to communicate something, then you aren't supposed to communicate that information. Doing it in a really roundabout way is still communicating it.


The EFF doesn't suggest actually using the warranty canary without court approval:

> What should an ISP do if the warrant canary is triggered?

> If an ISP with a warrant canary receives gagged legal process, it should obtain legal counsel and go to a court for a determination that it cannot be required to publish false information. While some ISPs may be tempted to engage in civil disobedience, EFF believes that it is better to present the issue to a court, to help establish a precedent. If you run an ISP with a warrant canary and receive gagged legal process, contact info@eff.org if you would like help finding counsel.


What's the warrenty canary worth then? You have no way of knowing if it was triggered or silently ignored.


Luckily USG has issued a menu of two permitted disclosure options: http://www.justice.gov/iso/opa/resources/3662014127160184071...

It would be a tough case to make that switching to an explicitly government-permitted method of disclosure was a violation of an order.


If I'm reading that correctly, you're not allowed to say that you haven't received any NSLs. The best you can do is to say that you've received fewer than 250.


Exactly. And if reddit's next transparency report says that they've received 0-249 NSL/FISA notices, they'll be well within the bounds of the Deputy A.G.'s letter.


If you are saying "they can use this as the side-channel to communicate that they've actually gotten an NSL", then you've failed. By establishing that this is something that they are using to communicate information, it's clearly a method to communicate information.

It's true that reddit could change to that new format. However, if they are using the change to that new format to communicate information, then they are sunk.

This is the real conundrum:

1. if It Is Known that this is how reddit will communicate an NSL, then it's clearly communication

2. if it's not known, then this isn't a good way for reddit to tell its users that they got an NSL.


What I'm saying is the following:

In the eyes of a court, either it's reasonable to infer that a disclosure notice compliant with the Dep. A.G. letter is a sign of having received a NSL, or it isn't. If it isn't, then the court can have no objection to such a notice. If it is reasonable to make that inference, then the Dep A.G. letter amounts to a formal authorization from USG to issue a document permitting that inference.

In the eyes of the internet at large, a 0-249 notice never actually means 0 – by the Gricean maxim of quantity, if a range is given then it can be assumed that it was necessary that a range be given.


In the eyes of the internet at large

This is the problem: if you set up a codebook of "perfectly legal things to say," it doesn't mean you escape gag order by using that codebook to talk about things in the gag order.

Here's a hypothetical (which someone has probably already suggested):

Say the reddit admin who puts this report together always includes a number from 0-999 at the very end of the report. To that admin, it's a signal of a gag order having been imposed. Maybe when it's an odd number, or when it's prime, or when the middle digit is the sum of the two outer digits, or the number is greater than 900.

If only that admin knows the system, everything is okay. But it's also useless as to what a warrant canary is used for: telling users you've been subject to a gag order.

So the admin tells people that it's intended as a canary and the method to decode. It's pretty obvious this is communicating exactly the information that you've been forbidden to communicate.

So if the Internet community decides that "we have received no NSL letters" means no NSL letters and "we have received 0-249 NSL letters" means there has been an NSL letter, than changing your message to comply with that understanding is an obvious communication.


All this tells me, really, is that national security letters and the laws surrounding them are a complete farce. I have a hard time believing the government could ever win a case against someone accused of disclosing the receipt of one.


What I wonder is whether there would even be a trial, or a public one at least. "Terrorist" may be what you get defined as and who knows what happens then. Sad but is truly my concern. And I'm really not a conspiracy theorist typically...


Can they claim to have received 251 letters if they only received one?


That's a little bit playing with words but if next year same day, this post doesn't re-appear on Reddit then automatically:

    * They don't communicate anything to users
    * Users understand a gag-order was issued
    * They didn't break any laws
To an interested party, a hint such as this, should be enough to act as an alarm. On the other hand if they force the party in some way to release a statement then I guess the users are toasted.


* Users understand a gag-order was issued

Right there? Communication happened.

All these attempts to say "well, I'm not really communicating" are just flopping on the beach. You have information, and because of your action, a third-party now has that information. We'd call that communication any day of the week.


If non-communiticating is taken as actually communicating you are correct. However proving that this is the case in court IMHO is not easy.


Non-communicating would mean never actually having include the clause in the first place. Having it and then abruptly removing it, in this case, is pretty clearly communicating something.

Or, in signal processing terms:

    ______________
                  |______________

    information---^


To me the problem is that a warrant canary could theoretically be completely legal, but the government could bankrupt you with legal fees before you even got to court.

That is probably what WhisperSystem's attorneys have advised them.


This makes more sense—being held in contempt of court is not in itself a ruling (i.e. canary warrants are not in themselves illegal, but disobeying the court is), and is extremely difficult to fight.


Apple disagrees. http://www.zdnet.com/article/apple-omits-warrant-canary-from...

And I think they have at least one lawyer.


Here's what you know: Apple used to have a warrant canary, and now doesn't.

I could put a warrant canary on my website, and then remove it. What does that prove?


Instead of not communicating it. What if you block people from viewing something else as a consequence of receiving a request.

Eg there is a public mailbox folder that can be checked by anyone online and there is an email filter that directs all mail with keywords typically found in requests to that folder. When a request is received you would of course be forced to block access to the mailbox as soon as you were aware of the takedown, but losing the ability to check that the mailbox is clean is the warrant canary. There would always be a small window between when the email arrived and was visible and when the company saw the request and blocked access.


Okay, I'm getting this pattern by people who want to make warrant canaries work:

   1. You receive warrant you aren't allowed to tell friend about.
   2. ???
   3. Friend now knows you got a warrant!  How clever!
It doesn't matter what stuff you cram into #2. I'm sure you want to talk to the court at great length about how #2 means you didn't communicate.

Courts have dealt with this nonsense for a long time. It's called "distinction without a difference."


What if you stuff the first amendment to the US constitution in there?

It's only one sentence, and very blunt:

>Congress shall make no law...abridging the freedom of speech, or of the press...


It should be evident by now that that one sentence is effectively meaningless in the US.


Has it been tested in court in relation to NSLs?


Wouldn't this essentially translate to the government forcing the company to mislead the public? This is different from forcing the company to be silent—that is perhaps understandable. But I would love to understand the rationale behind the government forcing a company to post a false notice with no benefit to anyone but the government itself.


Saying "the government is forcing me to lie!" is ignoring your agency is establishing the canary.


>This is different from forcing the company to be silent—that is perhaps understandable.

How is this understandable? The US constitution prohibits it.


If they act in such a way that they must continue to act the same way to prevent knowledge from being transferred then I could see the courts ruling they must continue to act in such a way even if it would involve lying.

The only ways around this is for the information that a warrant was received to be sent before it is received (to give an example that doesn't violate causality, a person stream 100% of everything they see/hear, so that between the time of receiving such a warrant and realizing what the warrant actually is the information that it has been received was already sent out). And even then I see the courts ruling against them because the courts will argue the spirit of the law was violated.


How does the government prove that you're communicating something when you've never explicitly said 'this is a warranty canary, it's ommission next year means that something is up'.

Take Apple. It had something that was assumed to be a warranty canary. It's no longer there. People made assumptions. Has Apple done anything wrong?


prove that you're communicating something when you've never explicitly said

Because courts aren't that dumb. Blackmail is illegal, too, and frequently the person demanding blackmail doesn't explicitly say "give me money or I publish this evidence." (It's legal to publish information, and it's legal to get paid not to publish it, but it's illegal to demand money not to publish it.)

I don't know what happened with Apple.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: