4
$\begingroup$

Firstly, I apologise for the short body of the question. This is somewhat related to allowing amateur ciphers on the site but also relates to questions such as this one.

Questions on such schemes usually boil down to the answers given in the linked post above, and I can't see an advantage on having them as they don't add any real value to the site.

At the same time, the question might be a bit more nuanced than that, ie. only a small portion of the scheme is secret (but can then be modeled with a keyed function or something similar).

Should we reject these questions right from the start, or accept them and add a don't rely on this component in each answer?

$\endgroup$
  • 1
    $\begingroup$ Personally I strongly dislike those questions. $\endgroup$ – CodesInChaos Aug 2 '13 at 8:17
  • 3
    $\begingroup$ As an aside, the fellow's name is Auguste Kerckhoffs, so it's either Kerckhoffs's or Kerckhoffs' (depending on your dialect), but not Kerckhoff's. $\endgroup$ – Reid Aug 2 '13 at 13:32
8
$\begingroup$

I think we should not blankly disallow question just because the person is no expert.

  • Is AES really easy to decrypt if you have the key?
  • Why does ROT13 provide no cryptographic security?

Those two question are valid in my point of view. I am sure there is quite a number of people out there, who have these questions. So it is a good, that these questions (and their answers) show up pretty high on common search engines.

We cannot expect everyone to know the term "Kerckhoff's principle" and use it in their searches.

tl'dr: If the answers to crypto-related questions are helpful to a significant number of Internet users, we should keep them. Regardless of whether they are pretty basic or not.

$\endgroup$
7
$\begingroup$

Perhaps we should have one canonical question about what Kerckhoff's principle is and why it's so essential to modern crypto. Then close all those questions as duplicate of it.

$\endgroup$
  • 3
    $\begingroup$ I think we should do this, but not necessarily close other questions as duplicates of it: we can just link to it to prove our point. (Unless the security of the scheme is entirely dependent on obscurity, in which case the question isn't really repairable.) $\endgroup$ – Reid Aug 2 '13 at 13:30
  • $\begingroup$ crypto.stackexchange.com/questions/9603/… looks like a good candidate for a canonical reference? $\endgroup$ – figlesquidge Nov 11 '13 at 11:41

You must log in to answer this question.

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