12

I recently got access to the question closure stats (yay?), and I've been thinking about the reasons why we close questions.

About 56% of closed questions, over the last 90 days, are closed as either "unclear what you're asking" or "off-topic - Questions on problems in code you've written must describe the specific problem...". I'm wondering whether we can invest work on our off-topic reasons to help make expectations for questioners clearer, while also providing more relevant and structured feedback on question closure. (I'm assuming we have some way to modify these, since they're customized for Salesforce?)

Would it be worthwhile to see about changing or adding to our roster of off-topic reasons? A few suggestions that occur to me:

  • "You've asked a question about your code, but didn't provide the relevant code. Please edit your question to add this information, but consider including only affected lines rather than pasting your entire class." (We don't really insist on "short, self-contained syntactically-valid examples", so perhaps we can be more direct and specific?)
  • "You've asked a question about an error, but you didn't include the error or its location. Please provide the verbatim error message, line number, and affected code."
  • "Questions about testing and code coverage should focus on testing specific lines of code or areas of functionality. For help getting started with a test class, see (Trailhead link)."
  • "Questions about how to achieve specific requirements or tasks should show your own research and work towards that goal. Requests for sample code or complete implementations are off-topic."

These are just some thoughts; feel free to shoot them down independently of the broader question of whether this is a tree up which it's worth barking.

I think it'd do us good to be able to express some of our more common question issues structurally, rather than leaving it up to individual commenters.

6

We are limited to three custom close reasons.


For reference, below is the current state of the close reasons:

Currently Active:

  • Questions on problems in code you've written must describe the specific problem and include valid code to reproduce it. For help writing short, self-contained syntactically-valid examples, see: SSCCE.org.

    This close reason is our catch-all for good reason. It describes in a compact way many of the most common problems with bad questions on this site:

    • Short
      • When a post contains hundreds of lines of code, for example, it is simply too much to scan through for the question to be of any value to the broader community.
    • Self-contained
      • When a post does not contain enough information to reproduce the issue in another org, it becomes difficult or impossible to verify an answer with certainty. Posting answers as a guess is not really what we're after.
    • Syntactically-valid
      • When a post contains numerous syntax errors, it again becomes very difficult to post one concise, clear answer which can address all issues. Incorrect paraphrasing or pseudo-code is worse than excessive length.

    So I disagree with your contention we do not require SSCCE. Another name for this concept that we bandy about more regularly in these parts is Minimum Viable Reproduction (MVR). I haven't been able to find a good online resource to explain this concept but would be happy to include one if it's out there.

  • Programming questions not specific to Salesforce are off-topic here, but can be asked on Stack Overflow.

    This tag is completely redundant and can be safely removed. There is a standard close reason which reads:

    belongs on another site in the Stack Exchange network

    Once you select this reason, you can just type in Stack Overflow.

  • Question that are intended to propose ideas, solicit opinion or provoke discussion are better asked in an alternative forum.

    I think this reason would be improved by linking to a specific forum. Perhaps this one?

Pending approval:
(Before this reason can be approved, one of the currently active reasons must be deactivated.)

  • This question shows zero evidence of any research, and can be answered through a quick internet search.

Deactivated:

  • This question appears to be spam.

  • Questions specific to a unique user scenario are off-topic because they will be of no use to anybody else. A good example of this is asking for people to write the test coverage for a particular class or trigger.

  • Still fleshing out my response, what's above is just a baseline. – Adrian Larson Jun 15 '18 at 20:28
  • It's a shame we are limited to 3 reasons. Thank you for the insight! – David Reed Jun 15 '18 at 20:29
  • 3
    I think I'm more annoyed that many of the off-topic questions arguably could be the "please give us code" flag instead. There doesn't seem to be consistent usage with them. – sfdcfox Jun 21 '18 at 23:25
  • 1
    Still noodling on this one and hoping to come back and add some proposed new reasons. – Adrian Larson Jul 3 '18 at 16:58
4

My observation over the course of my participation in this forum is that someone will vote to close a question without making a comment to ask for the relevant information. If that were to happen consistently and a reasonable amount of time allowed for updates to a question, I believe the number of questions closed for the reasons you've cited would be dramatically reduced. At times it almost appears as though there's a "race" to see how quickly a question can be voted for closure. To me, that's something that can make this place seem unfriendly to new users.

Seeing a question voted for closure before a user has been given a chance to clarify their question, add relevant code, etc seems "unfriendly" to me, esp when it's a new user who's asking. I'd like to see some kind of limit on how quickly a question can be voted for closure (e.g. an hour minimum) to allow for those kinds of comments to surface but don't know if the system could support that.

In the past, we've had similar discussions about users downvoting without making comments. It's a similar kind of thing. It's not helpful to close a question without commenting on what a new user could do to improve their question so it could be answered. I don't know of a way to require a comment before the first vote to close a question can be cast. Yet, it seems that it's exactly that kind of thing that would be helpful to reduce this problem.

I appreciate you bringing up this topic and agree that it shouldn't be happening.

  • 2
    I think I'm actually approaching the same issue from a very different direction. I tend to see fast closure (and fast reopening, where merited) as an ideal paradigm. (I'm not saying the way this currently works is ideal). Interaction with individual commenters is, in my view, a relatively poorer way to help questioners improve, not least because it prioritizes person-to-person communication without reference to clear community standards. I would love to see improvements allowing questions to be closed fast with clearer, standardized messaging, easily reopened when fixed, and not downvoted. – David Reed Jun 25 '18 at 2:33
  • 1
    Yes, very different approaches. I see that as impersonal and unfriendly with the reasons we have now. Too many member use close reasons like "unclear what you're asking" because they don't have the knowledge or experience to understand the context of the question. I can understand that, but they'd be better off allowing someone else to respond to it. Questions that are too broad are much more obvious. – crmprogdev Jun 26 '18 at 13:11
  • 2
    I really like that closing a question forces a user to edit their post. I see a number of low rep users who have a question closed, when it could be edited, with comments asking for edits, who just don't come back. Honestly, we don't even close that many questions. Our closure rate is about 5%. Closer to 6% if you just look at the last year. SE is at 5%, Software Engineering is at 9%. Workplace at a staggering 46%, Interpersonal skills at 20%, 11% for UX. Honestly, I think we do alright here. – battery.cord Jun 26 '18 at 15:03
  • 2
    I think the real problem is that once a question is closed, its not obvious or intuitive how to re-open that question, or that you can even re-open it at all. This seems to be an inherit limitation to the platform, not just our stack. Hopefully the new UI coming in the next few months gives them a chance to overhaul these systems, and provide "just in time" messages on actions a user can take, rather than directing them to a wall of text, without any information on what items are relevant to them. – battery.cord Jun 26 '18 at 15:05
  • 1
    @battery.cord I've little doubt that it can feel overwhelming for a new user when their question is closed. After editing, they also have to wait for someone to nominate it for reopening, then for it to be voted for reopening to receive an answer. As I've said before, if that can be avoided, I believe new users would find this place more welcoming. I don't think its a matter of merely improving the process of reopening a question. To a newbie, its as though they've had the door slammed shut while standing in the doorway asking for help. – crmprogdev Jun 26 '18 at 16:01
  • I think once a vote is cast to close, the user should be notified & provided some information about whats happening & what actions they can take. Usually they is done indirectly via comments, but it scales poorly, and I think were seeing the result of that scaling failure. I wish more users lurked on their questions, more comment activity, more edits, anything, would help keep more questions from being closed. I see a lot of users ask a question, and then walk away, and come back 12 hours later - too often to a closed question. This is all from my personal experience, however. – battery.cord Jun 27 '18 at 20:49
  • 3k+ users know the details for each close reason, but a 1 rep user won't, unless someone copies it into a comment, or directs them to the help page. AFAIK, they don't see it until the question is actually closed. – battery.cord Jun 27 '18 at 20:51
  • 1
    @battery.cord If you re-read what you've written, I hope you'll conclude the problem truly isn't with how to reopen questions, but instead with how questions are quickly closed without providing useful feedback to the clueless users who wrote them. – crmprogdev Jun 28 '18 at 11:45
  • 1
    I completely agree with you. I think we just disagree about methods - which I'm accepting of. I'd like to see more re-opened questions. I'd also like to see less questions closed! – battery.cord Jun 28 '18 at 13:48
  • Just a note, no one has to wait for nomination to reopen after editing. Any edit automatically enters a post into the Reopen review queue. – Adrian Larson Jul 3 '18 at 16:58
  • @AdrianLarson The last time I edited a closed question, I recall being asked if I wanted to nominate the question for reopening. – crmprogdev Jul 3 '18 at 18:13

You must log in to answer this question.

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