15

Proposal: Computer Science (Non-Programming)

I am having trouble with understanding the scope of this proposal. (Partially because of attempts to merge it with other CS proposals.) My first impression was that this is going to be the CS counterpart to Math.SE and Physics.SE. Am I right in my understanding of the scope?

The proposal currently states:

for computer Science practitioners, researchers, and CS students interested in topics like theoretical computer science, artificial intelligence, programming language design, digital logic, or any other area of CS not directly related to programming.

This is confusing for me. Is it an attempt to create a site that will get questions from all topics in CS of all levels? In which case this proposal overlaps with several existing sites (including cstheory) and its creation will damage them.

Can people promoting the proposal clarify the scope of this proposed site?

  • 6
    I second this concern. For the reasons mentioned above, I strongly believe that this proposal should be oriented towards non-researchers (with research questions being allowed). It would also be easier to build a community that way as people are more likely to know an answer to a lower level question than a research level one and it wouldn't conflict with existing sites either. – Opt Dec 5 '11 at 3:07
  • 1
    math.SE does not seem to hurt mathoverflow, or does it? I think people with research-level TCS questions will naturally go for cstheory.SE because they expect better answers there; similar for other (sub)fields. Of course it may happen that another site becomes redundant over time, e.g. because their experts are similarly active on the new site. If this happens naturally, I think it is a good thing. Personally, I would always refer questioneers that do not receive a timely good answer to the specialist community. – Raphael Dec 5 '11 at 8:02
  • 2
    I do not view SE sites as competitors, but as parts of a larger community. The sites are here to help people, not to mirror someone's world view. – Raphael Dec 5 '11 at 8:03
  • @Kaveh: You argue against specialist communities there but use the reverse argument here. How do the situations, in your opinion, differ? Is it just the order the proposals were made in? – Raphael Dec 5 '11 at 8:39
  • @Raphael: Speaking for myself, I think the size of the community which is online is very important. The research community of cstheory is large enough to support having its own site. Indeed, I doubt many of the members would visit a site for researchers that covered all areas of CS as that would tend drown out the theoretical part. – Opt Dec 6 '11 at 17:01
  • @Sid, I agree. I don't want to argue against cstheory in any way; I only think that cstheory can not be an argument against the new site! – Raphael Dec 6 '11 at 17:27
  • 2
    "A and B overlap, and A already exists" does not imply "the purpose of B is already being filled". I don't think "CS would overlap with TCS" is a good argument for not creating CS, given that there are questions within the scope of CS that TCS actively doesn't want and tries to avoid. – Ben Dec 8 '11 at 2:13
  • @Ben: this is about clarifying th scope of the proposal and make sure the scope is defined in such way that it won't overlap and damage existing sites. It was not about not creating CS.SE. – Kaveh Dec 10 '11 at 23:37
  • @Kaveh: You seem to be assuming that if the scope of CS.SE is not changed/clarified so that it does not overlap with other sites, then it should not be created. I think a site "that will get questions from all topics in CS of all levels" would still be valuable, and doesn't necessarily need its scope changed. – Ben Dec 14 '11 at 6:03
  • @Ben: "Yes, this does mean that these proposals — and, for that matter, any other proposals that would tend to drain audience away from existing Stack Exchange 2.0 sites — will be closed as duplicates." from blog.SO, i.e. if overlapping means that the new proposal is going to significantly damage an existing site it should be closed (or redefined in such a way that that it would not damage the existing sites). – Kaveh Dec 14 '11 at 9:32
  • Math.SE also gets questions from different levels. I have no objection to this proposal as long as this is going to be a CS version of Math.SE. – Kaveh Dec 14 '11 at 9:33
  • @Kaveh: That blog entry seems to be talking about sites that cover a subset of some existing site, which is not the case here. None of the the points 1, 2, and 3 apply to CS.SE wrt the existing TCS. And Ken Li's answer to this question (which you apparently accepted) argues well that CS.SE won't actually significantly damage TCS, by attracting all the questions that TCS doesn't want that appear to have no better place (the low level ones). This is precisely because this CS.SE overlaps with TCS but also covers significant areas that are explicitly not covered by TCS. – Ben Dec 14 '11 at 10:32
  • @Ben: I think the statement I quoted is very clear. The reason for closing as duplicate is general, it is not specific to sites with subset scope. Remember "non-programming" in the title of this proposal is there to make it clear this does not overlap with SO. Same applies to other sites. In fact SE asks community moderators to check area51 from time to time to see if any proposal is overlapping with already existing sites. – Kaveh Dec 14 '11 at 11:03
15

The scope should is all area of CS at all levels. So yes your understanding of being a CS counterpart of Math.SE and Physics.SE is correct.

And no I don't think this proposal will damage cstheory, in fact it should even help it because there will be much less off-topic low level questions on cstheory, as they will be posted on the new CS.SE. Also, nobody is on Theoretical Physics is suggesting to close Physics.SE just because research level question may be asked on Physics.SE, or Math.SE.

Also, while research level questions on TCS will be technically on topic on the proposal, the FAQ should have something along the lines of:

There are certain subjects that, while still being on-topic here, you can get better response on our sister sites:

similar to how Math.SE does it.

  • 2
    I agree with this analysis. I see a potential gain for CSTheory -- much more likely to have on-topic first-time questions -- and no loss. Added traffic to the new site might even cause added traffic to CSTheory, as people with theoretical questions are directed there. – Aaron Sterling Dec 7 '11 at 14:52
7

There is a precedent for a site whose scope is a subset of the scope of another site on Stack Exchange: Ask Ubuntu vs. Unix & Linux. The existence of U&L doesn't harm AU; AU's traffic is even significantly higher than U&L's. The sites tend to cater with different audiences, with AU tending towards just-make-it-work and U&L tending towards more esoteric questions and understanding in depth. The U&L FAQ recommends AU if the asker is uniquely concerned with Ubuntu.

I expect that the CS site will handle questions at any level on any topic that falls within the purview of a CS department and that is more about science than engineering (for engineering questions, there are already Stack Overflow and Programmers). Researchers will be welcome too, just like they are welcome on Math Stack Exchange, but research-level questions will be a small part of the site. I don't see any need to try to exclude researchers from CS.SE: they will naturally be a minority on the site anyway.

Note that research in most fields of computer science has no home on Stack Exchange. CSTheory and Cryptography specialize on a few fields of CS.

3

I think this should have happened during private beta on the site's meta, but let's go for it.

As for fields, I definitely see those:

  • Algorithmic Learning
  • Algorithms and Datastructures (+)
  • Artificial Intelligence/Machine Learning (*)
  • Compiler Theory
  • Computer Engineering
  • Database Theory
  • Distributed Systems
  • Formal Languages and Semantics (+)
  • Logics (+)
  • Network Theory
  • (Principles of) Programming Languages
  • (Formal) Specification and Verification of Systems (+)

Probably also the computer science parts of interdisciplinary fields such as:

  • Bioinformatics (*)
  • Robotics (*)

Please feel free to expand the list and add detailed subitems. Items marked with (*) are (partly) represented on other proposals/sites, too. Items marked with (+) are represented on cstheory.SE if on research level.

As for levels, these are my thoughts:

  • Soft questions as present on many sites, e.g. discussing the field, its practices or asking for general advice.
  • Academic Beginner questions, i.e. coming from undergraduate students and probably relating to course material. A policy for homework questions has to be developed; I like the model from math.SE.
  • Academic Advanced questions, i.e. questions students may come up with during their studies.
  • Academic Research questions, i.e. questions relating to current research and open problems.
  • Interested Practitioner questions, e.g. opens a textbook and is overwhelmed.
  • Practitioner Problem questions, i.e. a practitioner runs into a problem at work/his hobby that needs some CS insight.

I don't think the word "level" is quite right as they might be orthogonal, but for lack of a better word I stick to it.

  • 2
    the last three levels are also admissible in cstheory. In fact 'academic research' is probably the definition of what cstheory offers. – Suresh Dec 5 '11 at 8:34
  • 2
    Also, compilers theory and database theory are also represented on cstheory - there just aren't that many questions though. – Suresh Dec 5 '11 at 8:35
  • 1
    Regarding levels: Yes, those levels are present on cstheory.SE for TCS (though I think "interested practictioners" are typically sent to math.SE). Not for other fields/topics, though, and that is a point of this proposal. Regarding compilers and databases: I write "theory" to highlight the distinction between what CS does and discussing say SQL, gcc or LLVM. I think there is a huge gap between what would be admissable for cstheory.SE (both topic- and level-wise) and what can safely/productively asked on more practical sites. – Raphael Dec 5 '11 at 8:46
-5

I really don't see the need for a separate Stack Exchange site for this. On the contrary: I think its existence will harm the usefulness of Stack Exchange as a medium for these questions.

All of the subject areas mentioned are programming related; those questions are already being asked and answered on Stack Overflow. I haven't seen any indication that there is a problem with this.

Let's not fragment communities simply for the sake of categorization. The boundary between this and Stack Overflow will only cause a lot of extra hassle for everyone concerned.

So, proponents, please:

  1. Provide a clear delineation of the intended scope that proponents agree on, in terms of types of questions asked ('people interested in Computer Science' is useless as a characterization of scope; 'questions related to computer science' is much too vague).
  2. Prove that such questions will be, or have been, rejected as off-topic on Stack Overflow.
  • 4
    We've already been through this: computer science is not the same thing as programming (it's science), and I have given examples of CS questions closed or ignored on SO. Please read my answer here and this MSO thread. – Gilles Dec 20 '11 at 0:16
  • (Computer science isn't science any more than, say, geometry.) As I wrote, I do not experience a problem with general questions being ignored on Stack Overflow, and the lack of such questions may simply mean that they do not arise so often in practice; but if others do experience a problem here, clearly this must be addressed. Thanks for the pointer, very helpful. – reinierpost Jan 16 '12 at 9:04
  • Since I wrote this, I think the Computer Science site has clearly proved itself, and determining its scope between StackOverflow and Programming on one hand and Theoretical Computer Science on the other hand hasn't turned out to be much of a problem: while misplaced questions are common, it's almost always clear where they do belong. So I withdraw my objection. Still not sure what the downvotes are for, though. – reinierpost Aug 21 '15 at 22:28

You must log in to answer this question.

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