up vote 9 down vote favorite

The question came up: Are questions about Software Engineering ontopic?

share

5 Answers

up vote 5 down vote

The CS.SE site was set up, in my understanding, to satisfy a need: the need for a place for CS students, academics, and professionals to ask questions related to CS, which do not meet the criteria for sister sites (especially StackOverflow and CS Theory). Naturally, a mature CS.SE should be expected to overlap (to some degree) with Math.SE, StackOverflow, and cstheory.SE, but the idea of bringing in mathematicians, theoretical computer scientists and software developers to build a distinct community remains; that is, non-research-level, non-programming CS questions are really only on-topic on Math.SE, and that's not really an optimal situation.

Software engineering, imho, is another story. Programmers.SE already has an established community for asking questions related to software engineering, and they do not discriminate on the basis of question level or theory content (as far as I'm aware). From the Programmers.SE FAQ, questions on the following are the on-topic subjects there:

  1. algorithm and data structure concepts
  2. design patterns
  3. developer testing
  4. development methodologies
  5. freelancing and business concerns
  6. quality assurance
  7. software architecture
  8. software engineering
  9. software licensing

Of the 9 explicitly listed on-topic subjects, fully 6 are unambiguously, unquestionably, undeniably the sole subject matter of software engineering. That is, 2/3 of the subjects are software engineering proper, which leads me to the following lines of thought: there is already a non-research-level SE site where questions of all kinds related to software engineering are explicitly allowed; there is no need to have a new SE allow these questions, although it certainly could; "computer science" and "software engineering" are already understood to be distinct, even if the distinction is not yet strong; the distinction should be expected to become more marked in the future.

Taking these factors into account, I'm not sure I see any value in promoting the asking of software engineering questions here, any more than I see value in continuing to promote the asking of CS questions on Math.SE. I would say that the position should be that, while perhaps questions on software engineering aren't forbidden here, they're not really appropriate, and this community isn't the best one to address them. EDIT: To be more explicit, I say that if we must either completely allow or completely prohibit questions on software engineering, I say we prohibit them.

share
3  
Will academic software engineers feel at home on programmers.SE? I suspect they would rather be here, especially when formal methods (as they call it) enter the game. – Raphael♦ Mar 7 '12 at 15:52
1  
There's something unsavory to me about having CS.SE be a site for non-research-level CS questions on the one hand, and "academic" software engineering questions on the other. I'm having a hard time putting into words exactly what I find unsatisfying about that, but I'll get back to you. – Patrick87♦ Mar 7 '12 at 16:00
1  
Research-level CS was always intended to be ontopic here if it is not TCS as (only) this particular niche has its own site. – Raphael♦ Mar 7 '12 at 16:18
1  
@Raphael You say "if it is not TCS as (only) this particular niche has its own site"... but software engineering already has its own site. It's called Programmers. That's the argument I'm making... that this situation is no different from the situation with cstheory (except software engineering is, arguably, even less related to this site's primary focus than is the on-topic material on cstheory!). – Patrick87♦ Mar 7 '12 at 17:09
2  
Therefore my question: is Programmers appealing to academics, i.e. how do research-level SE questions fare there? – Raphael♦ Mar 7 '12 at 17:19
@Raphael What sort of research-level questions are you interested in when it comes to software engineering? The question that spawned this meta discussion would be solidly on-topic and would have a good chance to get good responses on Programmers - there's a group of people with strong interest in and knowledge of theoretical software engineering. – Anna Lear♦ Mar 14 '12 at 1:46
@AnnaLear I have no interest in Software Engineering myself (beyond reading stuff I come across), but that does not mean that nobody does. I think that in particular formal methods questions should do very well here. – Raphael♦ Mar 14 '12 at 6:29
@Raphael I'm sure there are many things (some even programming-related :)) that people here are interested in. I believe there are existing places that serve SEng questions well and there's no need to bundle SEng with computer science here. – Anna Lear♦ Mar 14 '12 at 13:23
@AnnaLear: I guess we will have to see wether users discover that some questions are answered better here. – Raphael♦ Mar 14 '12 at 13:25
@Raphael In the mean time, what do we do with questions like the one which sparked this discussion: that are not related to mathematical or theoretical aspects of software engineering? Are they off-topic, or not? It seems like the consensus (from this question, which may be a poor sample) is that software engineering questions, in general, don't belong here. – Patrick87♦ Mar 14 '12 at 14:14
@Patrick87: I don't know, really, as the grey area is infamiliar territory for me. I guess we have to decide on a case-to-case basis until we figure out our comfort zone. – Raphael♦ Mar 14 '12 at 14:45
up vote 4 down vote

In my opinion this particular question should be in scope.

Computer Science is a unique blend. On one hand it shares with mathematics the formal aspects. On the other hand it shares with natural sciences the experimental aspects. I suppose no-one would argue for banning questions related to experimental algorithmics. So, why would we ban questions related to other kinds of experiments?

What if someone would ask whether there is evidence that static typing improves productivity? Should that be banned? Before you answer, take a moment to read how Wadler, a very well known researcher in the theory of programming languages, pleads for more experimental research that could shed light on this question.

I agree that many other venues that allow "software engineering" questions are full of subjective and endless arguments. All the more reason to allow such questions here, I say. But make sure we only answer with facts.

share
"All the more reason to allow such questions here, I say. But make sure we only answer with facts." -- Well said. – Raphael♦ Mar 7 '12 at 16:23
up vote 3 down vote

Since SE is part of applied CS, it is on-topic because I think applied CS is part of CS just as TCS is and we are a CS site, just to add though:

This isn't programmers.SE, so if any SE questions do come up, it should be answered in an objective way preferably with evidence, and if the question is too subjective and cannot be answered with evidence or objectively, then it should be migrated or closed as off-topic. We are a computer science site after all, and anecdotes should have no place here.

share
up vote 2 down vote

The [programmers.se] also allows algorithms and data structure questions. Should we also forbid them on this site?

I think this is not the right attitude to look at the picture. With this attitude, at the end we are going to be left with only non-research level theoretical questions (which themselves were already on-topic on [math.se]).

So if we are going to go this way we should stop building the site right now and not waste our time building a site which is going to be obsolete. There are lots of Software Engineering questions which are more computer engineering questions than computer science question (though the line separating them is not black and white).

According to the area51 proposal these topics should be in the scope of this site. It was known that these other sites did exist and overlap.

I suggest that we wait on the issue of topics overlapping with other sites and see how things develop before making a decision how to deal with this issue. So for the time being, any topic in Computer Science should be allowed.

share
1  
-1. Strawman: programmers.se allows questions on that, but I make it clear that the focus of that site is software engineering; 2/3 of the example topics are software engineering. Slippery slope: we will have plenty of opportunities for questions not suitable for math.se, such as ones related to operating systems, networks and graphs, algorithm design and analysis, etc. Can't agree with this. – Patrick87♦ Mar 7 '12 at 19:39
@Patrick87, and the focus of this site is computer science, not software engineering. You are saying that we should not allow these topics because there are other sites covering them, and the same applies to Algorithms and Data Structures. Systems (Networks, OS, ...) might eventually have their own site, and algorithm design and analysis is on-topic on Mathematics, see these questions. – Kaveh Mar 7 '12 at 19:58
@Patrick87, the point you are making was expressed by me and others in area51 discussions (as I have linked above) and at the end the decision was to continue with the proposal and now it has reached the private beta without a change in restricting the scope. I strongly suggest that you check these and other area51 discussions related to the overlap with other SE sites. – Kaveh Mar 7 '12 at 20:02
up vote -2 down vote

In my opinion, SE and CS will have to separate at some point in the future; methods and mindset are too different.

This has yet to happen, though, and till then SE is ontopic on cs.SE, imho.

share
3  
If you foresee question classes being off-topic, BETA is probably the best time to say that they are off-topic. Otherwise, the broken window theory applies. i.e. it should be off-topic from the start. – Aryabhata Mar 7 '12 at 15:41
They will be offtopic when the academic fields separate, what may or may not happen over a time frame of decades. This is not relevant for us now, imho. – Raphael♦ Mar 7 '12 at 15:43
I never though SE was part of CS :-) but that is just my opinion. – Aryabhata Mar 7 '12 at 15:44
1  
Anyway, now that we have two answers, people can choose. – Aryabhata Mar 7 '12 at 15:46
2  
@Aryabhata It's generally part of CS curricula though (OTOH so are basic math classes). – sepp2k Mar 7 '12 at 15:47
@sepp2k: Aren't those departments usually called Computer Science and Engineering ? Anyway, this is just talk without data :-) – Aryabhata Mar 7 '12 at 15:51
2  
@Aryabhata, here in University of Toronto we have departments of CS and ECE (electronic and computer engineering), and SE is one of the large groups in CS. – Kaveh Mar 7 '12 at 19:27
@Kaveh: What parts of SE do they teach there? Anyway, not enough data points (either way). – Aryabhata Mar 7 '12 at 19:44
Check DCS SE research, DCS courses, CE research, and courses. In many north american universities, Electronic Engineering and Computer Science are in one department (e.g. MIT, Berkeley, ...). – Kaveh Mar 7 '12 at 19:52
@Kaveh: Are those Software Engineering courses mandatory for getting a CS degree? That would be interesting to know. Anyway, as I said one university is not enough, but I suppose getting a reasonable data set would be too much work (either for or against). – Aryabhata Mar 7 '12 at 21:58
@Aryabhata I had to take a course on SE for my CS bachelor and took another non-mandatory one. – Raphael♦ Mar 7 '12 at 22:01
@Kaveh: That is quite interesting. That wasn't the case when I got my CS degree. I guess it just goes to show that I am getting old :) :( – Aryabhata Mar 7 '12 at 22:02
1  
@Aryabhata, you are experienced not old. :) – Kaveh Mar 8 '12 at 1:18
@Kaveh: ................ :-) – Aryabhata Mar 8 '12 at 1:27

You must log in to answer this question.

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