8
$\begingroup$

As all of you may know or notice that many people don't read the tour or the help section to ask questions.

We have a lot of

Requests for analyzing ciphertext, finding hash preimages, identifying or decoding some code, or even reviewing full cryptographic designs are off-topic, as the results are rarely useful to anyone else and/or would be too long for this site.

type of questions. My ideas to reduce the submissions of these questions.

  1. When the Ask Question is clicked the cursor focused on the title, and there on the right we have a yellow box with the title How to Ask. I suggest that we clearly write our the most common off-topic reason there.

  2. When you write how decrypt help message on the title then you will see the related questions with 3 closed and one hold. This suggests me an idea, too.

    If we write a canonical question that contains the most the common words of this type of questions and a similarly write and accepted the answer, in the same manner, it will be on the top of the Questions that may already have your answer. It may show an indication. We may need to upvotes for both the question and answer to be on the top.

    We can leave the answer to the community and lock the question, AFAIK, we cannot leave the question to the community.

  3. update 04.02.2017: On the Tour of the Math.SE, they put an item with a link;

    • Hints on mathematical problems (but please read our FAQ about homework questions) similarly, we can do the same for homework and help analyze type questions.

update: I've seen an example that a question left to the community.

Will these help to reduce?

$\endgroup$
  • 1
    $\begingroup$ "How to ask" cannot be changed by mods, that's a SE wide feature handled by the SE team. $\endgroup$ – Maarten - reinstate Monica Feb 2 at 23:55
  • $\begingroup$ As for the generic Q/A, that would work if there was anything like a universal answer for analyzing ciphertext. We could say: send it to the NSA, but I don't think that's very helpful. So although I'm all for reducing the number of these questions, I don't think these two solutions will do it :( $\endgroup$ – Maarten - reinstate Monica Feb 2 at 23:56
  • 1
    $\begingroup$ @MaartenBodewes I've checked the math they have a different text with a link. $\endgroup$ – kelalaka Feb 5 at 9:25
4
$\begingroup$

Listing off-topic reasons in the "How to ask" section

This seems to be a reasonable thing to do.

If there are types of questions that users should not ask, then we should be displaying that information in a clearly visible location. Having it somewhere in the help center is not nearly as good as having it front and center when you ask a question.

I never paid much attention to the "How to ask" section before (I don't ask questions that frequently). I was actually kind of surprised at how little information is actually there. There is certainly a little bit of room there.

Canonical Q/A

I'm not sure about this being as good of an idea.

It would provide a precedent of a well-received (upvoted) question about "analyze ciphertext"-type questions. This may contribute to confusion regarding our policy on such questions.

Also, it might not work that well: Frequently these questions simply expect our community to solve whatever the challenge is for them.

Rarely do they show effort and ask specific questions on how to progress; In such cases, the question can usually be phrased so as to be answerable and not voted closed.

$\endgroup$
  • $\begingroup$ We have great answers that could be canonicalized for how to analyze novel cryptographic designs, at least. Perhaps something similar could be made for ciphertext analysis? Or maybe just something explaining that a ciphertext-only attack is really, really hard, especially with negligible amounts of ciphertext? $\endgroup$ – forest Feb 6 at 4:41
  • $\begingroup$ @forest You're free to post an answer in favor of the idea and explain what you think it should consist of; The problem I see is that when the users want help as opposed to want us to solve it for them, they usually do what they're supposed to and state what they've tried and where they're stuck. In such a case it doesn't have to be closed. The other case is when they simply dump the ciphertext and expect an answer from us; In this case, closing it as a duplicate won't help them, because they don't want to try, they simply want us to give them the answer, which the canonical Q won't contain. $\endgroup$ – Ella Rose Feb 6 at 15:22
  • $\begingroup$ (part 2) And in such cases we don't need it to be a duplicate in order to close it. On the other hand, if the intent of the canonical Q is to simply offer a concentrated list of potentially applicable techniques, that might be helpful for the first case, but people still might get stuck even after reading it and ask questions anyways. $\endgroup$ – Ella Rose Feb 6 at 15:22
  • $\begingroup$ @forest I posted a counter response to the points brought up in this answer $\endgroup$ – Ella Rose Feb 6 at 15:59
1
$\begingroup$

This answer brings up counter-points to this other answer.

Canonical Q/A

It would provide a precedent of a well-received (upvoted) question about "analyze ciphertext"-type questions. This may contribute to confusion regarding our policy on such questions.

This can be mitigated by locking the question, which would put a banner under it that mentions that it is not a good example of how to ask a question.

Also, it might not work that well: Frequently these questions simply expect our community to solve whatever the challenge is for them.

Being comfortable with mathematical notation/ideas and basic programming ability are more or less prerequisites to solving these sorts of problems.

Many (if not most) people are neither comfortable with math nor have any programming ability, so these types of questions are always going to appear.

If we concede that no-effort questions will always be a problem, then we can move on and try to optimize for the remaining cases where the situation can be improved.

Rarely do they show effort and ask specific questions on how to progress; In such cases, the question can usually be phrased so as to be answerable and not voted closed.

A canonical cryptanalysis question could actually prevent some of these from being asked in the first place. If the answer is already written down, that saves a user the trouble of asking and waiting for a response, not to mention the nervous stress that a new user asking a question is likely to feel.

One potential drawback might be: When such a question is asked in a way that it is answerable (meaning asking about general techniques), that answer can conceivably become useful to a small amount of other users.

However, if we had a concentrated source of cryptanalysis how-to, then it doesn't matter so much if those answers are helpful to others, because such help would already contained elsewhere in a more popular and accessible location.

$\endgroup$

You must log in to answer this question.

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