2
$\begingroup$

Would it be appropriate to ask questions that elicit community opinion and discussion?

For example:

Title: "What's your favorite data structure/algorithm/problem that satisfies requirements x, y, z? Why?"

Body: "Recently I've looked into data structure/algorithm/problem A and noticed similarities to others: B, C. Clearly A has these benefits over B and C. I was wondering if you (the community) knew of any other similar data structure/algorithm/problems that are similar or could be applicable to the aforementioned scenario. etc. etc."

This would be a question where people could share ideas and opinions about particular CS topics. I would like to imagine this is different than reference request. Although references would be expected in any answer to this, it isn't particularly a request.

I've also seen the opinion-based questions appear to be non-constructive, but this category of non-constructive seems more applicable to a scenario like "I have this problem. What's a good way to solve this problem?". Which would elicit opinionated solutions.

I haven't seen too many of these types of questions. I've posted a similar sort of question. In my opinion these questions are very beneficial to a community. Allowing people to freely contribute their opinions and passions (in some cases) about a CS topic. Not only that, but there is a discovery factor to the answers. Community members may learn and discover new ideas they had not previously known about which again (in my opinion) seem very beneficial to community members.

With that being said, I do see potential issues similar to the reference request issues and one of my own:

  • "They have the potential to grow forever, with everyone and their grandmother putting in their favorite resource."

    I think this could be remedied by making sure the topics are narrow. Broad topic (e.g. "What's your favorite tree?") would yield way too many responses or it would lose value as someone could checkout a wikipedia page on the topic.

  • "They tend to accumulate duplicate answers."

    This, again, would be more prevalent for broad questions where community members may not read through all answers and go ahead and write a duplicate without realizing it.

  • "The votes are primarily an indication of who came first, secondarily an indication of popularity, and only very occasionally an indication of suitability."

    In this scenario, popularity might not be such a bad thing. The more interesting and noteworthy the response, the more votes it should receive. In the end I could imagine it being similar to a poll where community members vote for their favorite response. This, again, might actually be a good thing.

  • "Stack Exchange is a questions and answers platform. It isn't good at being a collection of links. Lists of books or articles about a general topic can go in tag wikis."

    This, I think, would be the most reasonable issue with these kinds of questions. It very well could end up as a list of links. Although I think the well-needed difference here is that these should be much less than a wiki of topic-related links. It would be unnecessary and uninteresting for members to post answers just because they can. The content should (I say this with high hopes) be valuable and interesting.

  • Whence answered, it may die in the abyss, never to see the light of day again.

    It would be unrealistic for a community member or bot to poke these types of questions just to get more visibility. If a question like this gets minimal answers then maybe it was not a good question to begin with, maybe no one really cares or knows about the topic enough to begin with. Not all these types of questions are meant to be eye-opening, if only but a few, but hopefully the good ones would not end up like this. Usually if it's a good topic that enough community members care about or are interested in, it will get upvoted which would attract more attention and hopefully more interesting answers and discussion.

Overall, I feel questions of this nature are a good thing, but I would love to hear other opinions about this.

$\endgroup$
  • $\begingroup$ I think it would be tough. The platform is well suited for Q&A. There is some acceptance criterion required, the voting system is prepared to "bubble" the best answer, people are prepared for such outcome. The idea is cool, it might fill some needs but it would leave two types of questions, more work to maintain, constant struggle to draw a linr between "opinion based" and this new type of question. Many people would fight that "true" OB was intended to be this new type. So without some modifications in the platform, rewriting help center, indicating that particular question is this type cn. $\endgroup$ – Evil May 14 '17 at 21:16
  • 1
    $\begingroup$ It would just hit major problems. All platforms share the same type. It is really hard for new users to comply with current format (I had the same problems at the start). So now giving another type of post would be that much harder for newcomers, even harder for people who are already acustomed to it. Until now there were maybe several questions that would fit this. It reminds me of a request to split answers into theoretical/practical ones, but it didn't happened. Finally, if you reword "favorite" into "the best" given what you optimize, it is just like now. $\endgroup$ – Evil May 14 '17 at 21:23
  • 1
    $\begingroup$ thx for your interest in improving the site with a creative twist. however, site structure is fairly firm and has been established over years. think the main/ best strategy to making changes is to build up some interets/ coalition/ group engagement in Computer Science Chat. & it takes work/ commitment/ engagement over time. $\endgroup$ – vzn May 15 '17 at 15:27
  • $\begingroup$ Those are good points. Maybe the CS chat would be more appropriate if someone wanted to get more ideas about a topic they're interested in, rather than a formal question. $\endgroup$ – ryan May 15 '17 at 17:24
  • $\begingroup$ Hey Algorithms from the Book seems to do well on cstheory. Guidelines are OK, but please, can we have some fun once in a while? $\endgroup$ – Hendrik Jan May 28 '17 at 4:10
  • $\begingroup$ Here is another question that is relevant to this: cs.stackexchange.com/q/108659/68251 $\endgroup$ – ryan Apr 29 at 3:19
4
$\begingroup$

As the help center clearly states: No, opinion polls and discussions are not okay. Here.

Discussing opinions based on facts is a good thing, but this platform sucks for it. It's better for both the site and the discussion if it's held elsewhere. Stack Exchange Chat can be a prudent alternative.


As for your example:

What's your favorite data structure/algorithm/problem that satisfies requirements x, y, z? Why?

That is already a long title, and unfolding the requirements will make it untenable. It an open-ended question that invites opinions, facts, and arguments from all over a whole area of research/work.

Luckily, the proposed body is not about "favorite" things at all:

Recently I've looked into data structure/algorithm/problem A and noticed similarities to others: B, C. Clearly A has these benefits over B and C. I was wondering if you (the community) knew of any other similar data structure/algorithm/problems that are similar or could be applicable to the aforementioned scenario. etc. etc.

This suggests you've already done a lot of research yourself and are beyond asking "what's your favorite?" and have reached the realms of being able to articulate a meaningful question. Properly scoped, those are great questions for Stack Exchange!

$\endgroup$

You must log in to answer this question.

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