15

I believe this is related to Are the wrong questions and answers getting too many upvotes?.

I am wondering what should be done when a popular question is getting answers that are downright incorrect, to the point of being dangerous? When it's just one answer, the solution is simple, but in this case it is nearly all answers, and they all have high upvotes.

How great is the risk in publicly sharing part of a private key? is the question in mind. In it, someone is asking what the danger of revealing a small part of the private key is. While his example size was 256 bits, it seemed fairly clear that he was talking about public key cryptography. The answers involved some of the following (paraphrased):

  • Why would you want to do that? Just share the public key.
  • Leaking only a tiny bit doesn't matter. A 4k key with 2048 bit leaked still has a 22048 keyspace.
  • Brute force is easier than factoring so it's fine to leak a little bit.
  • It doesn't matter what part of the key is leaked. n bits is n bits.

All of these are incorrect or non-answers. For example, it is untrue that leaking a small bit of a private key is safe, because only a tiny tiny fraction can completely destroy the security, whereas if all that is leaked is the exponent, no security is lost at all. I found only one answer that was correct in the comments:

All information in a private key is not equal, so there is no "equivalent factor". Leak the modulus or public exponent and nothing happened (versus sharing the public key). Leak "50%" of the prime numbers and you're 100% boned

However, this patently incorrect comment has double the upvotes:

As factoring semiprimes is assumed to be so secure that brute force is the best possible attack

This should not be happening, especially with something this basic. If this question were moved to Crypto.SE, the answers wouldn't even be close to acceptable. What can be done about this? Should it be moved to Crypto.SE to let the answers get the downvotes they deserve? Should I just ignore it and move on?

8
  • 4
    On a positive note, now the top two answers are both correct and highlights the risks.
    – Anders
    Dec 14, 2017 at 22:15
  • 2
    As a note, this question got into HNQ that is available on the right sidebar on almost all sites, thus this question got an exceptional exposure from anyone on SE. Unfortunately, some of them have various knowledge & understanding but yet were brave enough to "contribute" (comment/post an answer). That's the source of the problem that we don't have the solution for until now.
    – Andrew T.
    Dec 21, 2017 at 10:13
  • 1
    Yeah I realized. That kind of overexposure can be good or bad. I'm glad people decided to answer, but it's security where accuracy is vitally important, not something like English.SE where a bad answer is just slightly irritating. I imagine some sites like Aviation.SE or Law.SE have similar issues.
    – forest
    Dec 21, 2017 at 10:17
  • 2
    The standard SE advice is to downvote and flag, but that never works in a significant fraction of the most egregious errors. Especially if the question gets onto the Hot-Network-Questions, in which case you can say goodbye to correctness. An alternative is to complain about it in the chat-rooms, and hopefully some experts will take notice and help. By the way, I noticed that you say on your profile that you lost access to three previous accounts. So see here.
    – user21820
    Jan 2, 2018 at 10:40
  • @user21820 Unfortunately one of the accounts I do not even have a password to (it is unregistered), another I lost the password to (I may find it again eventually), and the other I do not have the email for, even though I do know the password. As such I can't merge any of them.
    – forest
    Jan 2, 2018 at 11:42
  • The first paragraph in the help-page I linked to tells you to contact the SE team.
    – user21820
    Jan 2, 2018 at 11:45
  • @user21820 Yep, I did. The response was that I needed to know my previous email since it was, despite being a 10minutemail, "private information". So I gave up on merging them. Not a big deal as it's only something like 3k reputation combined.
    – forest
    Jan 2, 2018 at 12:00
  • Hmm that's troublesome; what a pity.
    – user21820
    Jan 2, 2018 at 12:16

3 Answers 3

8

The simple answer is to use your votes and flags.

Yep - that's pretty much it. This has been discussed at length on various site metas (especially on DIY, where a mistake could be life threatening) and the general consensus is to use votes and flags.

Enough flags gets things removed (and individuals with high rep in a particular tag can also expedite removal) and enough downvotes makes posts fall to the bottom.

(as an aside - I think this is mostly a duplicate of the one you linked...)

10
  • In the case where there are half a dozen incorrect answers, should I flag them all? For now I just flagged the OP to have it moved to Crypto.SE
    – forest
    Dec 13, 2017 at 9:17
  • Remember, you may be wrong with some of your assumptions - the top rep users from Crypto also hang out here and they provide very valuable answers, input, flags etc. I'm looking at some of your comments, and "dangerously incorrect" in some cases is just your opinion, so your votes/flags may be acted upon and may be ignored. You can use them as you like.
    – Rory Alsop Mod
    Dec 13, 2017 at 9:21
  • 1
    While dangerously is an opinion, it is true that the recommendation and implication, if followed by someone without understanding why it is incorrect, can result in complete loss of confidentiality for the key in question. For example saying that it's fine to leak a little because all bits in an RSA key are homogenous.
    – forest
    Dec 13, 2017 at 9:23
  • So yes - use your downvote. That's really how Stack Exchange works. If you are worried about running out of rep, it's not hard to earn more, especially if you are knowledgeable in Crypto.
    – Rory Alsop Mod
    Dec 13, 2017 at 9:24
  • 1
    Alright, I'll just flag and downvote. I have to hope that ~40 crypto experts do the same to drown out the less knowledgeable ones who come due to its popularity. :/
    – forest
    Dec 13, 2017 at 9:25
  • 3
    The good thing is that the less knowledgeable ones are almost always drowned out by the Bears, DW and others who really know this stuff :-)
    – Rory Alsop Mod
    Dec 13, 2017 at 9:26
  • 1
    I hope so! Thanks for the answer.
    – forest
    Dec 13, 2017 at 9:27
  • 2
    I know it can be frustrating. I'm an infosec specialist, not a crypto one, but I feel your pain when I'm on other SE sites and I see security questions with patently incorrect answers. Flagging and downvoting really do help - especially once you have higher rep and you can vote for deletion from review queues etc...
    – Rory Alsop Mod
    Dec 13, 2017 at 9:31
  • Yeah I'm trying to build up my rep (wrote well-received answers a lot anonymously in the past). I suppose it's just another case of someone being wrong and I can only hope that the flag is acted upon. Thanks.
    – forest
    Dec 13, 2017 at 9:36
  • Let us continue this discussion in chat.
    – Rory Alsop Mod
    Dec 13, 2017 at 9:42
10

As a complement to Rorys answer I would like to add: Comment! Engage with posters in a constructive manner! In a nice way, point out errors and suggest solutions. A highly upvoted accepted answer is unlikely to go away from the top spot, but that does not mean it can not be fixed.

1
  • 6
    A good comment will also signal to other readers that they 'should' downvote the answer because it is incorrect.
    – user13695
    Dec 18, 2017 at 10:45
1

Upvotes on wrong but nice-looking answers tend to accumulate because they come from people who are not site regulars, and are not experts themselves, but have been attracted to the question because it was linked outside the site (in the Stack Exchange “hot questions”, on Reddit, on Hacker News, …). An answer that is well-written and looks convincing to a non-expert is likely to rise to the top in these circumstances.

Moving the answer to another site wouldn't help. The problem is not that the Security.SE community didn't see the problem with the answer.

You need to do three things, all of which are important.

  • Downvote the answer. Downvoting is the main signal to convey that the answer is bad.
  • If there is a comment that politely and clearly explains the flaw, upvote it. Otherwise write one.
  • If there are one or more correct answers, upvote them. Otherwise, or if you have something significant to add to the existing answers, write one.

Then link the answer in the Sec.SE chat or show it to your competent friends and let them judge the answer for themselves. If the answer is genuinely bad, it's likely to accummulate a few downvotes that way, and your comment is likely to accummulate some upvotes which make it a bit more visible.

In the best case, the answerer will understand and admit the flaw, and will either edit their answer to correct it or delete it.

If the answer is genuinely dangerous, it may be removed by moderators. But the standard for dangerosity is pretty high. “This isn't best practice” is not high enough. “This is mind-bogglingly stupid and may get you hacked” may or may not be.

If it's any consolation, good answers do have a decent chance of surfacing up over time, if they aren't drowned in a sea of bad anwers. During the thread's 15 minutes of fame, most visitors have little competence and only a casual interest. But over time, a far larger proportion of visitors have a genuine interest in the topic, and they're more likely to read multiple answers and form an informed opinion, if they aren't experts. A noticeable example of this phenomenon is the XKCD password question. It had an answer from the founder of Stack Overflow, and for a while that was the top-ranking answer. That answer has laughable logic (do the math, then ignore the math because the conclusion wasn't the one the author wanted), but it's well-written and by a household name to the kind of visitors who flocked in initially, so it got upvotes. But after two years this bad answer was only a distant #2, and a bit later it was relegated to #3, now a distant #3. So while I wish this answer wasn't present, it didn't turn out to be fatal either.

You must log in to answer this question.

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