7
$\begingroup$

A recent example is 6 questions posted by turkistany.

Right now when I look at https://cstheory.stackexchange.com/questions?sort=newest I see the following 11 questions posted by him among the newest questions:

  1. Best bounds for the longest path optimization problem in cubic Hamiltonian graph?

  2. Is there a complexity theory analogue of Rice's theorem in computability theory?

  3. Major conjectures used to prove complexity lower bounds?

  4. What is the complexity of computing a compatible 3-coloring of a complete graph?

  5. Sparsity of Horn satisfiability?

  6. What is the most efficient algorithm to sample graphs with trivial automorphism groups ?

  7. What are the best known upper bounds and lower bounds for computing O(log n)-Clique?

  8. https://cstheory.stackexchange.com/questions/596/which-algorithms-have-the-greatest-impact-on-our-civilization

  9. What are the different notions of one-way functions?

  10. PCP characterization of NP

  11. Are there alternatives to using polynomials in defining the different notions of efficient computation?

IMHO, this is not a nice practice, and I consider this as a possible sign that the author is not really interested in the questions (and maybe haven't spent much time trying to find out the answer by himself/herself).

Should we discourage this kind of behavior?

ps: I know that we are still in beta and this might be OK for this stage.

| |
$\endgroup$
  • $\begingroup$ I think this experience demonstrates that having only +5 for upvoted questions is a good idea – cf. meta.cstheory.stackexchange.com/questions/37/… $\endgroup$ – Jukka Suomela Aug 27 '10 at 20:30
  • 1
    $\begingroup$ I think it actually demonstrates the kind of behaviour that is encouraged (and perfectly rational) if questions have half of the "value" of answers. $\endgroup$ – András Salamon Aug 27 '10 at 22:19
  • 2
    $\begingroup$ For reference, I wrote about this in a blog post: Asking the First Questions. Feel free to elaborate in an answer. $\endgroup$ – Robert Cartaino Aug 28 '10 at 0:22
  • 3
    $\begingroup$ I agree with Robert Cartaino's blog post in the comment above: the questions set the tone for what the site will become, so the fact that we're in beta should not be used as an excuse for poor questions. $\endgroup$ – Ian Aug 28 '10 at 2:05
8
$\begingroup$

So long as the questions are good, I don't see any particular harm. But yes, it is best for people to put time and effort into asking questions.

| |
$\endgroup$
  • 5
    $\begingroup$ I think it sets the wrong tone. Ideas are cheap, execution is hard: I am sure people on the site could generate dozens of one-liner questions on demand. For a fully automated version, just pick two random classes from the Complexity Zoo and ask how they relate. This isn't useful, interesting, or even cute. $\endgroup$ – András Salamon Aug 27 '10 at 22:27
  • 6
    $\begingroup$ I don't think that good ideas are cheap. I think asking a good question is actually very difficult: It shouldn't be too easy to solve. It shouldn't be a well-known open problem in disguise. It should be natural and well-motivated. It should attract the attention of experts and somehow trick them to spend their valuable time to seriously think about the problem. The question itself should be an eye-opening experience, regardless of whether someone is able to answer it or not. $\endgroup$ – Jukka Suomela Aug 27 '10 at 23:15
  • $\begingroup$ Agreed that great questions are valuable. But I haven't seen many great one-liner questions, since they require putting together well-known pieces. As such, they have typically been asked before, repeatedly. Most great one-liners first have a few definitions: let X be (some new thing), let Y be (some other new thing), let R be some interesting property, now does R(X,Y) hold? $\endgroup$ – András Salamon Aug 29 '10 at 11:28
7
$\begingroup$

I believe this behaviour is essentially destructive of the ethos of a high-level site for research-level questions. I think it should be discouraged, perhaps by posting the same questions in more fully fleshed out form (with a link back to the original question for attribution). This might be a better use of moderator attention than copy-editing poorly written one-liners.

Some of the value of a question is in sketching (or at least suggesting) the scenery within which the question lives. Even Randall Munroe will usually provide a few props to support the stick figures in xkcd.

| |
$\endgroup$
  • $\begingroup$ so you're saying that the moderator should take over the question, and repost under their ID ? That would make sense if the question could be closed. $\endgroup$ – Suresh Venkat Aug 27 '10 at 22:38
  • $\begingroup$ I don't think the original question needs to be closed. I'm saying that if people see that throwaway questions may get "taken over" (with attribution) by people who spend more time on making them real questions, then they may be discouraged from posting them in the first place, perhaps putting in a bit more effort themselves in the first place. However, I'm not sure what the long term consequences of this would be. $\endgroup$ – András Salamon Aug 27 '10 at 23:10
  • $\begingroup$ What if someone has access to the internet one day a week, and carefully keeps track of questions they wish to ask during the other six days? What if Richard Lipton has been storing a bag of 20 questions, none of which are fleshed out enough for a blog post, and upon finding this site, enthusiastically posts every single one of them? I think that, in general, it is best to judge questions purely on their content, and ignore who posted them. Spamming has occasionally occurred on MO, for instance someone copies/pastes many questions from elsewhere. That does not appear to be the case here. $\endgroup$ – matus Aug 27 '10 at 23:20
  • 1
    $\begingroup$ If someone has limited access, it is certainly in their interest to work on expressing their questions more fully. If Richard Lipton posts 20 one-liners, I would (with this proposal) expect most of them to become seeds for new questions by people who put the time in to make them real questions. I quite agree that the person posting should not have huge relevance here, but if someone posts a detailed question that is just an expanded version of someone else's one-liner, it would still be decent behaviour to note the original source. $\endgroup$ – András Salamon Aug 28 '10 at 1:13
  • $\begingroup$ Dear Andras, I'm sorry if I sounded contentious; I'll admit I sounded a little preachy. It is certainly extremely difficult to produce one good question, and having many is even more difficult. I merely wished to express that, in general, these situations can be hard to judge. $\endgroup$ – matus Aug 28 '10 at 1:48
  • 2
    $\begingroup$ this discussion would be relevant to this situation, except for the fact that turkistany puts almost no effort into the questions, provides no context and no reason for asking the questions. It's not the job of moderators to take raw material for questions and "massage" them into real questions. There's yet another example: cstheory.stackexchange.com/questions/700/… $\endgroup$ – Suresh Venkat Aug 28 '10 at 5:56
  • $\begingroup$ Dear Suresh, sorry, I made my comments to abstract and general, whereas the question was very specific. Thank you for putting so much effort into this site. I noticed that for many of your edits, you left a comment, which I think/hope will help people construct their questions properly. $\endgroup$ – matus Aug 28 '10 at 8:27
6
$\begingroup$

I'm having a hard time distinguishing the general case from this specific instance. As Shane says, I don't see a problem, especially given that we are in beta. However, I've had to spend a fair amount of time editing questions by this user, and it doesn't seem to have changed the way they phrase the next question.

Many of the questions are actually not that bad, but it takes a good few rounds of interrogation in the comments before the real question emerges.

| |
$\endgroup$
3
$\begingroup$

I think we need to have some policy on this type of thing. Or some way to strongly discourage badly phrased questions. For example, see this thread:

Complexity of a variant of the Mandelbrot set decision problem?

If you count all the comments, it took almost 30 comments to get a meaningful question out, and even then it seems like the OP keeps changing the question from iteration to iteration, much to the frustration of people posting answers. I'm not a subject area expert, but the frustration with this user's phrasing in this particular question is evident in that thread.

I've also noticed that this user's questions only start making sense after:

  1. Several rounds of interrogation.
  2. Suresh edits and fixes the question making it comprehensible (thanks for that, Suresh!)

Can we remedy this situation somehow?

| |
$\endgroup$
  • 3
    $\begingroup$ in all fairnesss, I should point out that this user's questions are improving over time, so there's some learning happening. The mandelbrot set problem was asked a while back, and the last 3-4 questions have been quite reasonable (with corresponding upvotes) $\endgroup$ – Suresh Venkat Sep 9 '10 at 16:38
  • 3
    $\begingroup$ Honestly, I think the best "remedy" is for others to questions of the sort they'd like to see on the site, so the signal-to-noise ratio is low. It seems to me that the real reason for this "problem" is the low # of research questions being asked per day. I haven't asked a question since we came out of private beta, so I'm as guilty as anyone. $\endgroup$ – Aaron Sterling Sep 9 '10 at 17:22
  • $\begingroup$ good point. In fact in comparison to other SE sites, we're doing quite well EXCEPT on number of questions, in which column we are abysmally low. $\endgroup$ – Suresh Venkat Sep 9 '10 at 22:04
  • $\begingroup$ @Suresh: Do you have any idea why that is? I just asked a question that is a genuine research thing I'd like to know, though I think the question was a bit lame. Maybe there's more of a "polymath culture" in math than TCS? $\endgroup$ – Aaron Sterling Sep 10 '10 at 1:48
  • $\begingroup$ How are we doing on number of users? I would guess that number of users is proportional to number of questions. $\endgroup$ – Robin Kothari Sep 10 '10 at 3:49
  • $\begingroup$ based on other SE sites, that's not true. we have a much smaller questions/user ratio. However, we also have many inactive users. I'd estimate that only about half of our users have actually done something on the site. $\endgroup$ – Suresh Venkat Sep 10 '10 at 4:11
  • 2
    $\begingroup$ @Aaron, I don't know. There might be some shyness in asking questions that have "obvious" answers, and it might be a matter of gaining some comfort with the site. There's a fine balance between "textbook" questions and "open research that no one knows the answer to", and I don't think we've hit it yet. $\endgroup$ – Suresh Venkat Sep 10 '10 at 4:12
2
$\begingroup$

My answer to the question β€œIs asking a bunch of questions in a short period of time appropriate?” is almost no, because it is almost impossible to post many well-thought-out questions in a short period of time unless the questioner has kept a stock of well-thought-out questions which are ready for posting. See my answer to another question.

If the questions are well thought-out, I do not mind ten questions from one person in an hour.

| |
$\endgroup$

You must log in to answer this question.

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