tea.mathoverflow.net - Discussion Feed (Recent questions in complexity theory)

Ryan Budney comments on "Recent questions in complexity theory" (16605)

2011-10-18T09:33:44-07:00

They could potentially be closed as "no longer relevant".

By and large the only problem abandoned questions cause is that if there's no response with an upvote, it will be recycled to the top of the list every once and a while.

I had a look at the questions you mentioned. Wow! There's a lot of them, and the majority have no responses, and little if any comments of substance. I voted to close as no longer relevant, at least for the ones that garnered no interest at all. If nobody agrees with me then no harm done and my votes will disappear in a few days.

Kaveh comments on "Recent questions in complexity theory" (16603)

2011-10-18T06:34:30-07:00

I think the user is not using the MO in a nice way and is posting a considerable number of questions that he doesn't seem to care about (they are left unanswered). The posts are not useful for others, e.g. 1 are just left there. Is there something that can done about this?

Kaveh comments on "Recent questions in complexity theory" (16434)

2011-10-06T21:06:58-07:00

@Ryan, sorry, I use EFF's HTTPS Everywhere and sometimes forgot that MO's certificate is not valid and I have to fix the URLs (now fixed).

@Will, the questions are not off-topic by themselves (they are NOT typical undergrad level). I probably would not notice it if there were not so many posted recently. On cstheory we have a little bit stricter policy and expect the OP to explain the motivation for the question and what they have done to answer it by themselves (particularly for questions in complexity since it is easy to generate difficult complexity theory questions, e.g. "What is the relation between complexity classes $C$ and $C'$?", "What are the consequences of $C$ being in $C'$?", ...). Probably with a little bit more effort the posted question can be turned into interesting and good questions.

@Anton, maybe we can politely suggest that they register and use real names? Also we can ask them to improve their questions by adding a motivation section explaining why they are interested in these questions and what they have tried to answer them?

@Henry, I agree that they don't look like undergrad homeworks, but I have seen instructors giving these type of questions as assignments in graduate courses. Still I am not sure if that is the case here, your theory is also reasonable. I just posted a comment to the last question linking to this discussion. (side note: I don't think that is true in general, some TCS researchers consider themselves computer scientists, others consider themselves mathematicians, and most consider themselves both.)

Henry Cohn comments on "Recent questions in complexity theory" (16433)

2011-10-06T21:06:22-07:00

I doubt these are homework, actually, although maybe I am being overly charitable. A couple of the questions might plausibly be asked as homework in a course, and a couple more could be in some way related to homework, but many of them don't seem that way, and I don't see an overall pattern. There are requests for surveys, questions about open problems, reference requests, questions about particular claims in published papers, etc.

I agree that cstheory.SE might be a more productive place to post them, although that makes me sad. I think the fact that theoretical computer scientists tend to self-identify as computer scientists rather than mathematicians is bad for theoretical computer science and bad for the rest of mathematics (while having them sit in CS departments does not offer great enough benefits to the rest of CS to outweigh these concerns), so I wish the online community weren't fragmented between several sites. However, in practice posting on cstheory will expose the questions to more readers who care.

I'm very curious as to the background of the poster (registering for an account and providing more context would definitely help). It looks like someone who is reading some important foundational papers from 20+ years ago without much expert guidance. I hope they manage to find the guidance they need online.

There's also the issue of the quality of the questions. Many of them are poorly phrased or not particularly well thought out, and there are a lot of questions, which makes me think the poster may just be posting every question they think of while reading. The right strategy would probably be to focus on a few particularly important questions and really try to write them well and explain the context. Has anyone referred the poster to this discussion?

Anton Geraschenko comments on "Recent questions in complexity theory" (16432)

2011-10-06T19:27:25-07:00

It does seem very likely that these are all the same user. Some of them come from the same IP (the ones with the same gravatar images), and the others are likely the same person for other reasons (e.g. same writing style, similar names, ...). It isn't our job to keep people from cheating on their homework, so that side of it doesn't bother me. However, it doesn't look like these questions are very good, which is the more annoying thing. What can be done about it?

Will Jagy comments on "Recent questions in complexity theory" (16431)

2011-10-06T19:10:11-07:00

From looking at cstheory.stackexchange, I see you very active but not "circuits." Would these be easily recognized as classwork at cstheory.stackexchange? That may be the real reason they are here, where they are not instantly perceived as coursework.

Ryan Budney comments on "Recent questions in complexity theory" (16430)

2011-10-06T19:09:09-07:00

Kaveh: when I click on your links I get an "untrusted" warning from my browser. My first response was to not open the link but on a closer look it appears to be because you're using https references rather than http ones.

On your actual concern: I've often wondered why people bother posting these types of questions here when the cstheory.stackexchange.com site seems a far better fit.

Kaveh comments on "Recent questions in complexity theory" (16429)

2011-10-06T18:34:45-07:00

I find it strange that recent MO questions in complexity theory are asked by unregistered users using nicknames. The questions don't look like easy undergrad homeworks, but they might be homeworks from a graduate course (they lack any personal motivation, don't seem to be focused, the nickname changes, the questions are stated in a way that feels like a graduate course assignment). It might be nice if moderators can check if they are being posted by the same user.

users: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, ...