In our discussion of whether the Theoretical Computer Science is too narrow for this site, I suggested a compromise idea that would have different forums like this site, all linked together more so than the average SE sites. It has been observed that this idea can even tackle the beginners forum issue (though I am not sure that would be a good idea). One major impediment is that the current software does not support such an idea, so the SE people would have to put in considerable effort to make it happen.

The idea went as follows:

I think it would be great to have a CS stackexchange site that had multiple "forums" (for the different fields of CS) like cstheory. The membership could be shared. The reputation points could transfer across the forums. But each site should have its own homepage which featured only its own questions - tags and filters are not nearly enough. Perhaps the featured bounty questions could be shared across forums, as could questions with very high scores. That way, we could see what exciting things are happening in the other fields and would be tempted to occasionally browse the other forums without their questions clogging the theory ones. This would increase the user base as users from all the forums would occasionally do some browsing and answering each others' questions.

The response seems to have been positive. I am curious:

  1. Is there support within the community for such a proposal (if done right)?
  2. Assuming there is community support, is this something the people running StackExchange are willing to consider implementing.

Does anyone see a downside? Is there opposition to this idea?

I guess cstheory users can only answer question 1 and the people running SE can only answer question 2.

  • A thought: what if we had a voting-based approach to determine on which "forums" a question appears? (I'm not referring to any heavy-weight "vote-to-migrate" approach. It would be more dynamic; for example, a research-level TCS forum would display all questions that have currently sufficiently many "research-level TCS" votes.) – Jukka Suomela Feb 13 '11 at 17:37
  • Sharing accounts, reputation and privileges should be very easy to realize. – Raphael Feb 13 '11 at 20:04
  • 2
    One could argue that a cheap way of realising this is to keep things as they are. Then tag certain posts with "beginner" (this can be a volunteer + community activity). Then ask the StackExchange people to implement 3 views of the questions, namely, those with the beginner tag, those without, and all questions. (Note that much of this can be emulated anyway.) – Dave Clarke Feb 13 '11 at 20:20
  • 4
    I think that having a beginner tag is exactly the wrong way to realize this proposal, and I would be against it. Also, the focus of it my proposal is not regarding the beginner questions issue at all, but rather regarding having many research-level communities linked together. – Lev Reyzin Feb 13 '11 at 22:17
  • 1
    I was playing Devil's Advocate. Which research-level communities would you like to link together? – Dave Clarke Feb 14 '11 at 10:01
up vote 7 down vote accepted

While I like the idea, this has the feel of the US Congress making a resolution about Egypt's political future. Which is to say, why discuss this matter here, and what power do we really have ? meta.SO is the right place for such discussions.

Moreover, I don't know if the time is right yet for this kind of thing. After all, we have only one example of a 'research level' forum in the SE family (since as has been made abundantly clear, MO "doesn't count")

Update: Consider "Is it Possible to Create Multiple Stack Exchange Sites that Share Some Questions and Answers?" on meta.SO.

  • You're right, but I wouldn't have wanted it brought up on meta.SO unless we actually wanted such a thing. Also I figured that since the issue was brought up by Joal on our meta, we can discuss it on our meta. – Lev Reyzin Feb 14 '11 at 13:30

I'm not opposed in principle to this idea, whereas I think it's pretty obvious that I don't suport some other proposals to change the scope of the site. I'm not all gung ho for it either. Some thoughts:

  1. Who would the moderators be? The currently-elected mods I am sure thought they would be moderating a TCS web site in which homework questions are not permitted. Would they be ok with moderating a web site that might include a forum full of HCI questions, or a beginner's forum where differentiating between homework questions and "legitimate" question would require psychic powers?

  2. How would rep be calculated? And, how would mods be elected? Say there are a kajillion users on the beginner's site. What is to stop them from engineering a "coup" and overturning all our meta policies on homework, etc.?

Also, I did a few google searches of users on other sites this morning, and I will say that the following is probably (though not with certainty) true:

At least ten users on the front page of our users list are "notable" in Wikipedia's sense (meaning, practically, that there could be an article on them in Wikipedia, or there already is). By contrast, it appears that none of the top users of StackOverflow, or Cooking, etc., meet that notability criterion.

Not that being Wikipedia-notable is the most important thing in the world, but it seems quite possible that we have more "big names in a field" than all the other StackExchange sites combined. (And, they are active users!) If that is true -- or if anything close to that is true -- then we've achieved something no one else has, and, far from telling us that we are elitists who should be more like StackOverflow, the Area51 people should take the hint that StackOverflow could use some cleaning up to be more like us.

I think it is far more likely that Vint Cerf or James Gosling (or anyone else StackOverflow regulars might term a "god") would stop by to hang out at our party than at StackOverflow, especially if we are able to start having more regular and more serious conversations about the theory of programming languages. Would that increase internet traffc in the short term and the long term? Hell yes.

Although I've disagreed with Spolsky, Atwood and Cartaino, I do believe they respect what we've accomplished, and want to build on it. The high-rep users on Area51, on the other hand, are fighting for a position I can only term "militant mediocrity." We should frankly ignore their proclamations that MathOverflow "doesn't count." I suggest we just continue to build a new paradigm on StackExchange 2.0: "Researchers at the edge of knowledge helping one another expand human understanding," and figure out the best way to do that.

I agree with what Suresh wrote.

Note that there are not many SE sites on scientific research which have passed the beta phase (is there any other than cstheory?) and I would prefer to wait to see if any other site would get launched before modifying the only working example to try to solve problems other proposals are facing. It is not just computer science proposals, I don't know any other scientific research level proposal that has got launched. We can discuss these things but I am strongly against making a decision too quickly. We should always have in mind what we were trying to achieve when we started this site and the effects any changes to the model will have on it. We want and need researchers on the site and anything which will have a negative effect on our ability to attract and maintain researchers is a bad thing. My personal opinion is that one of the reasons that cstheory got where it is right now is following a model on a related topic which has worked, i.e. MathOverflow, and we still haven't reached the place in Theory community that MathOverflow has in Math community.

I also don't see what this is going to solve. Each area needs to attract researchers, putting all of those who are already there won't solve any problem. SE needs researchers in AI to answer AI questions and as long as they are not here nothing will change. I think all of the expected benefits of a unified site can be achieved by having multiple sites, one for each area and then having a unified view like the way is unifying questions from many sites in one place (and nothing more like sharing reputation or votes).

I am also concerned about some of specific points of the proposal and the issues they can create, e.g. based on Physics.SE's experience, I am clearly against sharing reputation, especially if it is going to be shared with a non-research level questions/answers.

Finally, I don't think we have reached a state that we can relax and start to think about ways to modify the site, the site is just a few months old, many active research areas in theory are still heavily under-represented on cstheory, we should first get more researchers on those topics on cstheory before starting to worry about AI or Graphics or Systems. That is what we should focus on. The possible benefits of any modifications to cstheory are unclear in the best case and the possible negative consequences are considerable. I think we should wait for a year to let the site stabilize, changing things just a few months after cstheory's launch is not a good thing, we want to establish a brand in theory community, we have written a SIGACT article advertising the site just a few months ago to attract researchers in theory, any dramatic change will send the wrong signal to the community.

Think of cstheory as your company, would you make radical changes to its structure just after your first successful product is starting to get attention from consumers where the possible benefits of the changes are small/unclear and the risks are considerable?

  • 1
    +1 for some good points. I just wanted to raise the issue in a separate post b/c there seemed support for it. I agree it's probably way too early for any major changes to our framework. – Lev Reyzin Feb 14 '11 at 18:04
  • 2
    Kaveh's points are excellent. We are relaxing too early, while we still have nowhere near the stature in the theoryCS community that we might like to have. – Suresh Venkat Feb 14 '11 at 19:46
  • I totally agree with your points Kaveh. I want to stress out the issue of underrepresented research areas. Individual sites most likely will have a smaller percentage of success. – chazisop Feb 16 '11 at 18:34

You must log in to answer this question.

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