24

<Some question about design or product>

What are the pros and cons?

Sometimes, these questions get closed as 'too broad' or 'primarily opinion'. Why is that?

17

The "what are the pros and cons" style questions are slightly obscured polling questions... though only very slightly.

Questions of this type sometimes have a dozen answers that look like:

Pros

  • Point 1
  • Point 2

Cons

  • Counter point 1
  • Counter point 2

and

Something no one mentioned yet is that Foo is really good at doing bar.

The answers to these questions are a poll asking for people to collect various bits of facts into what they think (opinion - a pro for one person is a con for another) into a post. As there are often many of these factoids that goes into making a decision, there can be scores of them (just think of the weighted decision matrix that managers so love - I once had to fill out my opinions on some competing products that was over 100 rows - each with a different factoid of some importance (and I feel it still left out some important items)).

The nature of these questions also makes it so that they are never complete. One can never have an authoritative answer for the pros and cons, nor can one have an answer that includes everything - someone will always find another point that they like or dislike about the design decision. Questions of this format from three years ago still get people adding single pro or con items to it (often just a single sentence).

These types of questions rarely produce good answers and instead seem to have shorter and shorter answers as time goes on.


When comparing products, this falls into a further problematic area of shopping questions.

What are the pros and cons of having a Linux dev system rather than a FreeBSD dev system?

Design decisions are slightly better, though they are still problematic. They are still in the range of 'too broad' and 'primarily opinion'. A better option than asking for pros and cons is to look at trying to write the question as a design review.

Realize though, that asking for a design review rather than just 'pros and cons' isn't a simple matter of changing the finishing up part of the question to ask for a design review. The guidance on how to ask a design review question can be found at Are Design Review questions on-topic?


Many of the questions that have been asked in the 'pros and cons' format in the past can fairly easily be rewritten in a way that isn't a poll and instead identifies the problem and asks for solutions to it rather than polling for tidbits of opinions.

That said, many of the historical questions have answers that are in the 'pros and cons' format that would make rewriting the question while maintaining the validity of the answers difficult.

  • 19
    My experience with pros/cons is that in practice they give a good overview about the most important points that are relevant for most applications. The main difficulty is probably to sufficiently specialize the question. If one has narrowed it down then discussing pros/cons is probably still a good idea. – Trilarion Jul 24 '14 at 7:59
  • 1
    @Trilarion you get questions like programmers.stackexchange.com/questions/4325 and programmers.stackexchange.com/questions/73823 and programmers.stackexchange.com/questions/21230 which have some pretty awful answers in them. These questions don't try to encourage good answers. It is the quality of the answers at the site that draws people here - both to ask them and to answer them. By encouraging answers that are "pro: one sentence. con: one sentence" it reduces the expectations and has other people post similarly poor answers. – user40980 Jul 25 '14 at 13:10
  • Okay, I see your point. They are crappy questions. But this doesn't mean there aren't high voted questions with pros and cons, for example programmers.stackexchange.com/questions/72569/…. I wonder how you would rewrite this now or if it would get closed as too broad nowadays? – Trilarion Jul 29 '14 at 14:01
  • 1
    @Trilarion it would probably get closed as too broad or opinion. That question has some interesting contrasts - Trevor and Andrew do a good job of writing a complete and thorough answer. However, the other three answers are far less well written and comprehensive. This points out the key problem with questions like this - they don't do enough to encourage Trevor and Andrews answers while discouraging the other three (most of the text in those answers aren't even complete sentences). We want questions that encourage great answers - and pros and cons questions instead encourage poor answers. – user40980 Jul 30 '14 at 14:59
  • It may depend on the definition of great answers but for me they are great (in the sense of very useful even if opinioated, not in the sense of exhausting and free of opinion). I completely understand the current rationale. But my impression is that the whole stackexchange system also suffers from not having such questions (like What are the pros and cons of Coffeescript?) anymore. I guess they delivered quite some traffic. I read from your lines that there is practically no possibility to make the linked question conform to current standards. That seems a bit unfortunate to me. – Trilarion Jul 30 '14 at 16:25
  • 2
    The thing is, its not a sustainable SE site (some history). I will point to Slant which maintains as strong of a "what is the best..." focus (just as SE maintains a Q&A focus). As to the old question, it can't be something that is fixed because it would make the good answers there not match the question and I can't guess at the actual problem behind the question that can be something that is solved (I have been able to do such before). – user40980 Jul 30 '14 at 16:37
  • @Trilarion The older the question is, the harder it is to get the participation of the person asking question to fix it... and the more answers it has, the harder it is to craft the new question in such a way that it still matches the old answers (or deleting all the old answers... but then its just easier to ask a new question). This goes in part why its important to close questions fast so they can be fixed while the OP is still able to participate in fixing it and before the answers make fixing it even more difficult. – user40980 Jul 30 '14 at 16:39
  • 2
    "Many of the questions that have been asked in the 'pros and cons' format in the past can fairly easily be rewritten in a way that isn't a poll and instead identifies the problem and asks for solutions to it rather than polling for tidbits of opinions." I think this answer would be even more useful if you used a concrete pro/con example and then converted it the way you're recommending here. – Daniel Kaplan Apr 10 '15 at 18:05
  • @DanielKaplan the difficulty with concrete examples is that people often look at them and say "but my situation is different" - I've tried to keep the meta posts on subjects a bit more abstract so that people don't get too hung up on details. That said, if you were to come up with an example as an answer, I would be more than happy to up vote it and mention it in this answer. – user40980 Apr 10 '15 at 18:15
  • Sure, examples are easy to come up with. I don't know how to make them better though. e.g., "What are the pros and cons of pair programming?" – Daniel Kaplan Apr 10 '15 at 18:21
  • @DanielKaplan you could grab any from this search and then work through an improvement on it. If you have trouble, please feel free to join Software Engineering Chat (the white board) and we could help you. – user40980 Apr 10 '15 at 18:24
  • 1
    The three examples user40980 gave as examples of poor pros-and-cons questions are poor questions independent of the format (pros and cons. For example, "What are the pros and cons of working from home?" could be rephrased as "Is working from home more productive?" and it would be an equally poor fit for Stack Exchange. In some cases, people seem to be wrongly attributing it to the "pro and con" format. – Vaddadi Kartick Oct 7 '16 at 2:53

You must log in to answer this question.