33

Proposal: Computer Science (Non-Programming)

I originally wrote this question because I believed that this was a duplicate of Theoretical Computer Science. However it does appear that the example questions have already been closed on Theoretical Computer Science, for being too low-level. So while this would appear to be a duplicate, I'm not now clear where the questions asked actually do belong.

EDIT: If this isn't a duplicate that worries me, because that would imply that we have two sites for Computer Science: one for "research level" and one for easier questions. "Beginner level" sites don't do well. No other sites reject undergraduate level questions as "too easy".

10
  • 14
    No other sites reject undergraduate level questions as "too easy". <-- this is no longer true. I am in the private beta of Theoretical Physics, and that site has closed graduate-level questions as being "too easy," that is to say, not research-level.
    – Aaron Sterling
    Commented Sep 29, 2011 at 14:50
  • 1
    Did the Computer Science proposal from earlier this year (?) vanish?
    – Raphael
    Commented Sep 29, 2011 at 22:18
  • 3
    See also meta.stackoverflow.com/questions/80023/… and meta.stackoverflow.com/questions/79351/… as well as discuss.area51.stackexchange.com/questions/612/… for older discussions on the topic.
    – Raphael
    Commented Sep 29, 2011 at 22:37
  • 2
    two level issue has been discussed several times, please read the previous discussions. Non-research-level questions in TCS can be asked on Math.SE, we have checked it with them and they are fine with it. I suggest that you focus on your original question, the situation about Theoretical CS is already clear: research-level on cstheory, not research level on Math.SE.
    – Kaveh
    Commented Sep 30, 2011 at 22:47
  • Undergrad walks into a grad level class... #joke
    – Pratik Deoghare
    Commented Dec 6, 2011 at 13:49
  • Undergrad expects help from grad...#joke
    – DJClayworth
    Commented Dec 6, 2011 at 14:12
  • 2
    There is also MathOverflow which frowns on, and rapidly closes, undergraduate-level questions as too easy.
    – Marc van Leeuwen
    Commented Feb 4, 2012 at 11:29
  • Looks like you're confusing two questions: 1) Is all computer science theoretical? 2) Is it OK to have two sites (for computer science) distinguished solely by the level of the questions? (Research level or application.)
    – reinierpost
    Commented Feb 6, 2012 at 17:35
  • 2
    So, we will get two sites to endlessly discuss the P vs NP question.
    – Victor Stafusa
    Commented Feb 14, 2012 at 23:03
  • I also wanted to ask this question. But now there is scicomp.stackexchange.com, Computational Science, in addition to the Computer Science and CSTheory. So, everything seems ok now. Or, may be, a couple more cs sites?
    – user71451
    Commented Dec 30, 2012 at 13:44

8 Answers 8

Reset to default
24

There's a LONG-running argument on area51 about whether it's ok to have two CS sites or not, or if a TCS site is antithetical to the mix of experts and non-experts that some other SE sites have. It's hard to judge from the set of questions what the actual scope of this site would be, but it's clear that CS is a MUCH larger discipline than TCS, and in that respect alone there's no duplication.

6
  • 3
    The point is that if TCS is a subset of CS, then that's duplication.
    – DJClayworth
    Commented Oct 12, 2011 at 15:09
  • 5
    right. and if the CS site were first, this would probably be in argument in favor of NOT starting a TCS site. But the TCS site did come first, so now what :) ?
    – Suresh
    Commented Oct 12, 2011 at 16:26
  • 9
    TCS needs to be rescoped, imo. (Good luck with that, though.)
    – Richard
    Commented Oct 19, 2011 at 22:03
  • 2
    I agree with @Richard.
    – beroal
    Commented Nov 13, 2011 at 16:34
  • 4
    CS, IMO is just too large of a community for it to have a single stack exchange site. Also, many areas of CS are already served well by a collection of sites :- TCS for theory, meta.optimize/stats.stackexchange for ML/Stats, TCS/or-exchange for optimization, TCS/economics.stackexchange for game theory, etc. There are areas for which we don't have a place yet like Graphics, HCI, PL and Systems but I don't think the solution is to create a site encompassing all of CS.
    – Opt
    Commented Dec 1, 2011 at 23:56
  • Also would like to add that theoreticalphysics.se had to be split off from physics.se and programmers.se from stackoverflow and the former especially has a much smaller online presence than CS so in all likelihood if a CS site gets off the ground, it will have to be split up in the future.
    – Opt
    Commented Dec 2, 2011 at 0:01
22

I think it would be hard to build a community for this. One of the reasons AI.SE was closed was because of a lack of experts.

Currently if you have a computer science question, there are many places to take it:

cstheory.SE if its research level (the community is rather open on the theoretical part, but much more stringent on the research part)

math.SE if it involves theoretical computer science at a non-research level (including homework)

stats.SE (or metaoptimize) if it involves machine learning, and some kinds of AI

crypto.SE if it is about cryptography at any level

stackoverflow if it is computer science that is applicable or interesting to programmers/practitioners

I think this covers most of the scope. Of course, the person asking has to understand their question well enough to decide between the 5 sites. One could argue that since CS is so widely distributed, it could be worthwhile to have one site for everything... however, I think it would be impossible to attract experts to such a site. Further, how would it build a community?

So I would vote as duplicate, but not of cstheory.SE... but of all 5.

11
  • 9
    There are many low-level questions closed on cstheory, and to my knowledge next to none end up on math.SE. If they do, they get answered, but often (and unavoidably) from a mathematics viewpoint, which is not always helpful. On stackoverflow, anything but questions for man page knowledge does not get answered or answered badly (I might overstate slightly). Programmers is certainly as far away from TCS as mechatronics from physics. stats and crypto covers some areas that (heavily) intersect CS, but are very restricted. So no, I don't think CS is covered in a satisfying way.
    – Raphael
    Commented Sep 29, 2011 at 22:16
  • 2
    Non-research level question in TCS can be asked on Math.SE, they are in its scope, whether users ask them or not is a different issue.
    – Kaveh
    Commented Sep 30, 2011 at 22:50
  • Also MetaOptimize.com, Stats.SE (AI); programmers.SE (Software Engineering); GraphicDesign.SE (Graphics), Electronics.SE (Computer Hardware, Architecture); proposal for computational sciences/scientific computing; us.SE (HCI); ITSecurity.SE, Crypto.SE; avp.SE (sound); I think these cover almost all areas of CS.
    – Kaveh
    Commented Sep 30, 2011 at 22:58
  • 10
    I'd rather have one site for CS than a dozen, simply because field classifications are overly rigorous compared to reality. Overlaps in interest create synergy which a split that fine-grained effectively prohibits.
    – Raphael
    Commented Oct 3, 2011 at 6:48
  • 1
    Agree, thus I've voted to close.
    – mbq
    Commented Oct 3, 2011 at 12:19
  • The question weather having all of it on one site is better is not what is being discussed here. The point is a large portion of CS is already covered by Q&A sites, whether you like it or not.
    – Kaveh
    Commented Oct 3, 2011 at 14:01
  • 1
    I disagree; for this line of reasoning, you should only consider sites out of beta, which leaves the field much more uncovered.
    – Raphael
    Commented Oct 5, 2011 at 5:30
  • You are free to disagree, but the point that there are not sites that one can ask CS questions isn't true.
    – Kaveh
    Commented Oct 6, 2011 at 6:05
  • What about the scicomp.stackexchange.com? Why don't you list it?
    – user71451
    Commented Dec 30, 2012 at 13:45
  • @Val probably because it didn't exist when I wrote that answer. There isn't much debate on this anymore, since CS.SE is up and running, so I am not going to edit and bump this to the front page.
    – Artem Kaznatcheev
    Commented Dec 30, 2012 at 22:30
  • Our current AI proposal which started few days ago has enough experts, but not enough people to ask the right questions.
    – kenorb
    Commented Aug 11, 2016 at 1:04
4

I said it back then and I'll say it again: No duplicate.

One can argue, talk and conjecture a lot. My question is: What harm can be done if you put the site to beta (given enough attention) and see where it goes? If it does not turn out to work, close it and discuss again in which ways we can incorporate areas that are exclusive to the new proposal into other sites.

3

I don't think it's a duplicate. That said, it seems like a site for homework questions. Not sure I like it. The askers would be people who can't figure out their homework, and thus probably ill-equipped to answer most other questions. I don't think a site like that can attract committed answerers, and even if it did the site would have problems due to the clash between competent answerers and incompetent askers.

It seems like "Give me the proof / algorithm" style questions, akin to the "Give me the codez" questions so reviled on Stack Overflow.

2
  • 8
    Believe me, there is plenty of room between homework and state of the art research. For example, all the cranky ideas you get while learning.
    – Raphael
    Commented Sep 29, 2011 at 22:05
  • I don't agree. Surely appropriate topics and problems for Computer Science are learned and asked in undergraduate classes and homework. But that doesn't mean that a Computer Science site would permit blatant asking of homework problems without work shown first. Besides, sites like math.SE and SO have numerous homework problems, and this is considered unproblematic.
    – emi
    Commented Oct 20, 2011 at 20:23
2

The current TCS community somewhat exclusive towards questions that are not research-level, but that's only one of the many areas this proposal is for.

What about?

  • Interdisciplinary area between TCS and some other area such as computer vision.
  • Artificial intelligence (There was a proposal on this, but it was closed)
  • Compiler design, programming language design? These are not clear cut programming questions that may or may not fit on SO. Some of the CS questions tend to get downvotes for not being on-topic.
  • Some other area of CS that there is no SE for that doesn't exactly fit in SO, and not research-level enough to be on TCS, or doesn't even belong there in the first place.

There also seems to be a misunderstanding of what TCS is. TCS only focuses on mathematics of computing and does NOT include areas like artificial intelligence which this proposal covers as well.

5
  • 1
    the first 3 points, are already covered by cstheory at the research level: computer-vision, image-processing, AI, compiler, and programming-language. Thus, the CS forum would only differ in that it concentrates on non-research level questions.
    – Artem Kaznatcheev
    Commented Sep 29, 2011 at 6:32
  • 5
    @Artem, no, currently only theoretical questions in those areas are on-topic on cstheory.
    – Kaveh
    Commented Sep 29, 2011 at 10:19
  • 1
    But AI researchers already have a site (MetaOptimize.com, Stats.SE), Software Engineering has programmers.SE. There is a proposal for computational sciences/scientific computing. HCI has a proposal already. Graphics have a site as do Electronics. So in my opinion, the only uncovered area is Systems (Operating Systems, Networks, ...).
    – Kaveh
    Commented Sep 29, 2011 at 10:19
  • @Kaveh There's a lot of programming language theory (typing, compilation, static analysis, …) that falls through the cracks between SO and CSTheory. Programmers.SE isn't a reasonable place for scientific questions. Even algorithm questions, which are officially on-topic on SO, tend to get closed if they aren't tied to a particular programming language.
    – Gilles 'SO- stop being evil'
    Commented Oct 4, 2011 at 23:01
  • @Gilles: I am not an expert on PL, but surely research-level questions in PL theory are on-topic on cstheory. Non-theoretical questions about compilers, ... are not PL theory and to best of knowledge are usually considered part of systems area.
    – Kaveh
    Commented Oct 5, 2011 at 6:20
1

It is not a duplicate, but based on its description it overlaps heavily with several existing sites including cstheory:

Proposed Q&A site 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.

2
  • 3
    I believe you could say the exact same thing about physics vs. theoretical physics, yet both sites exist. Research and theory questions I suspect are always welcome at the non-research level site, but the reverse is not true.
    – Michael McGowan
    Commented Oct 7, 2011 at 21:29
  • 1
    @Michael, you don't seem to be a user of either, neither a user of cstheory. No, the situation is not similar. The general-level site existed before the research-level site, and it was created specifically for research-level questions, see the discussions on physics.SE's meta.
    – Kaveh
    Commented Oct 9, 2011 at 6:29
1

Going beyond first glance, it's clear that the proposed site is pitched to fill a need that currently exists.

Stackoverflow.com isn't great for theoretical computer science. Theoretical computer science shuts out anyone who isn't working at or beyond a doctoral level.

I would suggest that a broad computer science site would be an excellent idea.

-10

I've voted to close because the description is an obvious duplicate.

Edit

I have a problem with some of the questions closed at cstheory - a few examples certainly don't belong on SO, since they are algorithmic and/or theoretical in nature ... but the FAQ on cstheory says it's just for research-level questions.

I'm no professional CSTheorist, but what defines "research-level", and who cares if you get "lower" level questions, too?

5
  • I've actually changed the question. It does seem like there are questions that are off-topic at TCS and I don't know where they would belong.
    – DJClayworth
    Commented Sep 27, 2011 at 14:22
  • @DJ Clayworth - see my update
    – warren
    Commented Sep 27, 2011 at 14:53
  • 11
    @warren the community at cstheory cares if they get lower level questions. Asking a low level question is very easy compared to the effort one must put into asking an interesting research level question. If cstheory did not actively close low level questions, it would get swamped with them, and all the researchers would leave (this has happened with physics.SE).
    – Artem Kaznatcheev
    Commented Sep 29, 2011 at 6:01
  • 3
    @warren, the issue about the scope of cstheory has been discussed several times. If you are not part of cstheory community and haven't read the previous discussion it is better not to get into that discussion here. The scope of cstheory is defined in our FAQ.
    – Kaveh
    Commented Sep 30, 2011 at 22:40
  • 10
    ps: as a side note, I don't understand users who are not part of the community of cstheory and are not theoretical computer scientists expressing opinions and voting on issues that they are not familiar with. But generally non-research questions in TCS can be asked on Math.SE (we have checked this with them and they are fine with that).
    – Kaveh
    Commented Sep 30, 2011 at 22:44

You must log in to answer this question.

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