5
$\begingroup$

I am specifically referring to the Snowden question follow-up here. People have voiced concern over this type of question in comments, and there have been suggestions of taking this to Meta.

This discussion thread is for the "quantum challenge" follow-up, not about the "Snowden" questions in general. While there are undoubtedly controversial points about both, the original one has been well-received by the community (with a net score of +10) whereas the other one has not.


The question makes a lot of assumptions, in particular that we will have commercially available, general purpose quantum hardware in 2050, clearly against the FAQ:

You should only ask practical, answerable questions based on actual problems that you face. Chatty, open-ended questions diminish the usefulness of our site and push other questions off the front page.

Which would indeed seem to directly target this question. The question is difficult to answer objectively and as currently formulated, is likely to result in lots of speculation about quantum technologies in the next few decades, and doesn't ask a single, well-defined question that can be reasonably answered.

However, this is also somewhat true of the original question as pointed out by rath, yet that one was not at all discussed (even though it still has no accepted answer), and therefore it can be argued that both should be leave as is for consistency.

But this raises another point: shouldn't we try to make questions self-contained and independent from one another? Is it a good idea to have "part I, part II" questions around on the website?

Finally, should they be edited to have better names? A hypothetical scenario involving Snowden is fun and all but I am not sure it is helpful to have questions titled "How to say something confidentially to Snowden". How many people interested in the cryptographic aspect of the problem are going to find it in an internet search, versus people simply searching for "Snowden"? Perhaps the title should be edited to something more explicit and less sensationalistic.


I'm not trying to be a killjoy, but I think that while the original question is a good one, the second one does not belong on the site as is, and should be merged into the original one or deleted altogether.

What are your opinions on this?

$\endgroup$
  • 4
    $\begingroup$ I suggest to remove the reference to the recent media hype from the first question because it creates the impression that the main motivation for the question is, to get votes. $\endgroup$ – Hendrik Brummermann Jul 8 '13 at 6:29
  • 2
    $\begingroup$ @HendrikBrummermann I agree with you - using Alice/Bob terminology would be more appropriate $\endgroup$ – Thomas Jul 10 '13 at 6:53
  • $\begingroup$ Joining both questions sounds good to me as that could translate into a better question by providing the complete picture. But that's just me of course... $\endgroup$ – e-sushi Jul 11 '13 at 22:30
  • $\begingroup$ @HendrikBrummermann Provided all answers are edited too since they also contain the name Snowden. Also, I'm not sure if "Alice" and "Bob" examples would make sense as the situation described in the question might become unclear. $\endgroup$ – e-sushi Jul 11 '13 at 22:31
  • $\begingroup$ Do we want to create memes or allow them to go grow? I think not because this website is not meant for that kind of jocularity. Our questions would become flippant, losing their focus on crypto. The first Snowden question was meme-able, and it started to multiply. $\endgroup$ – Patriot Jul 20 at 10:09
5
$\begingroup$

Assumptions are fine as long as they make sense. We might not have quantum computers by 2050 but it stands to reason that we will at some point, and that our current understanding of cryptography will be affected by it. There are many questions about quantum cryptography on this site, all of them perfectly valid. An invalid question as I interpret the FAQ is

What would happen if we suddenly realized OTP is not secure after all?

but of course I might have taken it too far.

I don't find the question hard to answer objectively because all the schemes I've read about so far assume Alice and Bob who want to communicate and hide from some Mallory. In the case of Snowden (who obviously doesn't want to be reached, as I pointed out in my answer) we only have Alice so it would stand to reason that all such schemes would not work.

However, the asker condemned himself in my eyes with the following comment:

Sure, there are two excellent answers and all comments are highly appreciated. But all of them are negative. They have not been accepted because I am waiting for possible positive answers. I think nothing is impossible. Even if current knowing cryptographic methods cannot solve this problem, we have no reason to exclude future possibility, say after 2050

First of all, assuming a better answer might come along is OK, and you should wait a reasonable amount of time before accepting one. In my case, a few months ago I got all good answers on my question, and even accepted one untill this gem came along, which could be put directly in a book chapter of the same topic.

In the case of Snowden I might be wrong, there could be a way to contact him that I'm not aware about. Refusing to believe there is no such scheme even if all answers point to that direction is what I take issue with. After all, optimism is good untill you get peer-reviewed. It's not about getting your answer accepted, it's about being stubborn. Sorry.

Which brings me to the second issue I have with the comment to which I've already replied so I'll quote directly:

I didn't agree with @e-sushi when he said the question is hypothetical but now I do. We can always assume there is some scheme no one knows about that lets us do whatever we want. And if cryptography doesn't answer the question, the question does not belong here. Maybe on Security.SE or a proposal about spooks and spies I saw on Area51 (if it still exists).

I was willing to accept the follow-up as a valid question because quantum computation is about computation - it does not magically conjure a communication channel. That is obvious to me but may not be obvious to someone else, just like the difference between a hash and a permutation wasn't obvious untill I asked about it.


In conclusion, I believe the question initially had some merit but in the way it's evolved I see it has none. A possible rephrasing of the title would immediately yield the problem:

Does using quantum cryptography make it possible to contact someone if I don't know how to contact them?

As it stands I believe the question should be put on hold. The same of course applies to the original question.


Edit: I now see that in my answer I proposed some ludicrous things such as posting agents all over Venezuela. Someone else suggested using Lady Gaga as a cryptographic primitive, both clear indications that something is wrong as we were talking in hypotheticals.

I also saw that both questions were asked by the same person. This alone makes the follow-up bad, as the original could have been edited to include the quantum component.

$\endgroup$
2
$\begingroup$

I would like to see the questions being edited so that they actually add a bit more "crypto" value to crypto.SE because otherwise, they are too broad and might as well fit into the security.SE community. Yet, even there they would have a big potential to collide with user votes and feedback because "hypothetical questions about hypothetical situations can and will result in hypothetical answers".

Let's be honest, the current state of the questions could result in weird answers like

Let's suppose that — in a perfect world — cryptography will be safe tomorrow morning at 09:00 am UTC. If so, there's no problem to talk to Snowden securely because from tomorrow on, even your prayers to [enter-your-god-here] will be encrypted and safe from man-in-the-middle attacks.

If that's not the kind of answers we want, we should stick to the rules... and at least ask the questions to be edited so they add a bit more value to crypto.SE. That way, they might be a bit more acceptable too.

$\endgroup$
  • 1
    $\begingroup$ I suppose prayers are considered save independently of cryptography (if they are not shouted loudly). $\endgroup$ – Paŭlo Ebermann Jul 14 '13 at 11:46

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .