7
$\begingroup$

We have just had our 1000th question; and this is the 12th day we launched; not long, but enough to see how we're doing. Lots of discussions have been done in the review post of beta, while the number of questions is a weak link of us, we do not have enough upvotes, and also there are a number of homework problems, we tried our best to increase the quality of the posts and answers.

Although numbers are not important, sometimes they serve as indicators. I have just observed a drop in the answer ratio; at the time we graduated from beta it was about 2.8, and now it is about 2.5. It is also interesting that in the mean time, our traffic rate is constantly increasing, and the number of active users (users with a positive reputation change) in the late month is passed a hundred, a little increase compares to it was in beta.

Maybe the cause of the drop in the answer ratio is the homework problems, but I am not sure. Is this a healthy decrease (since we do not want to push the site too hard), or is it a signal of lacking enough thrust to keep flying our rocket? What do you think?

Another issue is the bounty system. It seems to me that recently, none of the questions with bounties have been answered in the bounty time period. The same problem occurs in MO, which may due to both sites are theoretical oriented, and since it is related to the system of SE, we do not have too much we can do. Is there any solution to the bounty problem, despite the fact that it serves as a mechanism to attract more attention?

Final thoughts, it seems that most of the traffics comes from a particular region of the world (maybe US), since there is a low traffic gap (for about 8 hours) in the posts. Not a problem but an observation, but I do expect more experts and grad students all over the world to join our community.

Keep going, see how far we can reach :)

| |
$\endgroup$
8
$\begingroup$

I think it's a good idea to assess where we are. I think the current diagnoses are spot on, in that

  • the reduction of big-list questions (thank goodness!) is reducing the average number of answers per question, and
  • the number of homework questions at the end of term is articially inflating the number of questions.

Overall though, we are growing, both in traffic and in number of 'active' users, which is a good sign. I'm also seeing more people come in and out of 'high activity' mode, and I think the law of large numbers will eventually kick in to smoothen out the activity levels.

Two things coming down the road that might help solidify things:

  • The SIGACT News article we all wrote will appear in the next issue of SIGACT News, and that will hopefully draw in some new users with new questions.
  • Talks are under way to have SIGACT formally brand the site, and that should give it an official imprimatur.
| |
$\endgroup$
  • $\begingroup$ Any idea when that issue is due? $\endgroup$ – Anthony Labarre Dec 14 '10 at 8:34
  • $\begingroup$ Either Jan or Feb, I believe. $\endgroup$ – Suresh Venkat Dec 14 '10 at 10:27
5
$\begingroup$

I for one plan to never visit this site ever.

Oops. Change that. What a great site.

| |
$\endgroup$
4
$\begingroup$

Two or three answers per questions sounds pretty good. My guess is that some of the early questions were of "big-list" type, so they are outliers that are skewing the average. As the increase in questions flattens the data, we are just approaching a reasonable data point.

| |
$\endgroup$
  • 1
    $\begingroup$ There are currently 44 [big-list] questions with a total 502 answers; without these questions the answer ratio is only 2.1, which is a little bit low in my opinion. But you are right: there are less [big-list] questions recently, so the ratio is converging. $\endgroup$ – Hsien-Chih Chang 張顯之 Dec 13 '10 at 6:30
4
$\begingroup$

I think we are doing well. I agree that the increase in the number of elementary questions is temporal, and the decrease in the ratio of answers per questions is also effected by smaller number of and CW questions we get which is not a bad thing by itself, the ratio is just getting closer to the ratio of questions which are not lists/CWs.

I don't know if a lower ratio is a bad thing, this might be a result of the theoretical/mathematical nature of theoretical computer science. I would like to know the current ratio for MO. If we think this is a problem we should try to encourage people to provide more answers to the questions from different perspectives, e.g. by putting an entry in the FAQ explaining it that it is good and in fact encouraged to post new answers even for questions which have accepted answers (assuming that the new answer is good and adds some new information).

I tried searching for various combinations of words related to the site. We seem to be on the top page in Google for those combinations I have tried (haven't checked Bing). One interesting thing that I noticed is that Suresh's old dated theory page is still ranked 2nd on Google when searching for the following words: cs theory. :)

I added a link to the external links section of Wikipedia page for Theoretical Computer Science. We may want to ask people responsible for other theory pages like the SIGACT page, and Theory Matters if they can add a link to the site.

Finally, in my view, the next major step for the site is seeing it being cited in conference and journal papers, in the same way Mathoverflow (and to a lesser extent the FOM mailing list) posts are cited. It will probably take sometime, but I think we have passed the stage that the number of questions and users were a major concern.

| |
$\endgroup$
1
$\begingroup$

This is a comment. I am also interested in the cause of the drop in the answer ratio. Homework questions sound like a plausible reason. Or it might just be a temporal fluctuation. But I am not sure, and these might be wishful thinking. How can we test these hypotheses?

They will probably provide a data dump of cstheory.stackexchange.com (and its meta) this month. Can we use it to conclude some useful information about the cause of the drop in the answer ratio?

| |
$\endgroup$
  • 2
    $\begingroup$ The Data Explorer provides database to the problems. Maybe we can wait until the dump is out. $\endgroup$ – Hsien-Chih Chang 張顯之 Dec 13 '10 at 6:34
  • 2
    $\begingroup$ In the newest blogpost of SE, Jeff said that "We won’t be producing a data dump for the month of December 2010". So maybe we have to wait until January. $\endgroup$ – Hsien-Chih Chang 張顯之 Dec 14 '10 at 3:41
  • $\begingroup$ @Hsien-Chih: Too bad. Thanks for the information. $\endgroup$ – Tsuyoshi Ito Dec 15 '10 at 0:06
  • $\begingroup$ the data dump seems like it might be quite useful, especially since we can "borrrow" well used queries. $\endgroup$ – Suresh Venkat Dec 15 '10 at 5:41
  • $\begingroup$ @Tsuyoshi, @Suresh: It seems that our data dump is out. I'll play around with it for a while... $\endgroup$ – Hsien-Chih Chang 張顯之 Jan 10 '11 at 10:01
  • $\begingroup$ @Hsien-Chih: Thanks for letting me know! $\endgroup$ – Tsuyoshi Ito Jan 11 '11 at 14:34
1
$\begingroup$

At the moment, about a third of the questions on the homepage have 0 points or less, or are closed. This is worrisome. Perhaps a good way to ameliorate the situation is to contact a few established researchers and try to get them involved. Was this proposal followed up?

| |
$\endgroup$
  • 2
    $\begingroup$ Perhaps the holiday season is responsible for this? Many researchers are probably on vacation or visiting family. $\endgroup$ – Robin Kothari Dec 18 '10 at 0:44
  • 2
    $\begingroup$ not to mention end of term question flooding. this appears to be an issue on MO as well. $\endgroup$ – Suresh Venkat Dec 18 '10 at 11:40

You must log in to answer this question.

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