What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 127 Stack Exchange communities.

I just failed an audit with this honeypot question. I thought it was a succinct and well-formulated question so I upvoted it. This was obviously wrong of me since the question was closed as NARQ.

But I don't think it was obvious. The question was well-formulated and the asker said his current code was working to some extent. I am no C++ programmer so perhaps I should have skipped it, but I really thought it looked like a good question, worthy of at least an upvote from me.

My questions then:

  • Are we supposed to be restrictive in handing out upvotes? Should I be more careful even outside the review system and only upvote questions I know are technically correct?
  • Do anyone else agree that this might have been a good question?
  • How is this even a NARQ?

Another example just popped up: I also failed this audit which obviously have code for starting a transition. The question is how to call an external function when the transition is done. Once again it was closed as NARQ and I failed the audit. I see nothing wrong with the question, there are no comments indicating anything wrong with it and looking back there is even a good answer.

share|improve this question
4  
It was NARQ because the OP didn't bother to read how the compare() works. –  hjpotter92 Apr 19 '13 at 13:52
3  
Isn't it obvious from the comments on the question?? –  Uphill Luge Apr 19 '13 at 14:57
3  
In the comments, the OP is even apologizing for asking the question. :-) –  Bo Persson Apr 19 '13 at 15:35
    
I still don't think it is good to use such edge-cases as audit questions. They are seemingly good questions, why should a review made by a broad audience catch this? There is no way of locking the audit system to only languages you so it is apparently made to show posts to a broad, general population of reviewers, not a domain-specific audience such as those subscribing to the relevant tags. –  Emil Vikström Apr 26 '13 at 12:30
1  
Because the same broad audience has the ability to close or reopen such a question. There is nothing wrong with the occasional audit being non-trivial. Because guess what, so are the reviews. –  Bart Apr 26 '13 at 12:43
add comment

2 Answers

up vote 8 down vote accepted
+100

This was actually a mistake, the question should not have been used as an audit. The question was flagged as "very low quality", and the flag was validated (note, mods are generally extra careful when it comes to quality related flags because they influence audits). Validating the flag alone wasn't enough to have the question appear in an audit - a perfect storm set up when a moderator deleted the question after receiving advice from its author hours after the low quality flag was validated.

Since the post had at least one quality related flag that was validated, then subsequently deleted, it showed up as an audit. No system is perfect, and we're definitely not pretending that audits are any exception to that rule. It was a case of human error, but it doesn't happen very often given the volume of posts that are legitimate audits being used to help folks hone their moderation skills while catching robot reviewers.

The mistake was validating the flag in the first place. While the question does have issues surrounding being an artifact of a brain fart, being poorly written or horribly formatted isn't one of them. But, it's only one audit and flukes like this are definitely not representative of the system as a whole - they're just flukes and unfortunately rather unavoidable. Like any other system fed by humans, mistakes are going to occasionally happen.

share|improve this answer
    
Wow, I had given up on getting any more answers to this. This seems reasonable and I do feel better knowing the audit questions are, at least implicitly, approved by moderators. Thanks for digging into this! –  Emil Vikström May 3 '13 at 14:23
    
@EmilVikström You're welcome. Since these flukes are unique and case-by-case I try to answer them individually when they come up, but somehow I missed this one. Audits are something that we keep a very close eye on since there are a few moving parts to them. Anyway, better late then never :) –  Tim Post May 3 '13 at 14:33
add comment

That question is indeed well formulated. And, ignoring correctness of statements made, there is not much wrong with it. But well, we can't really ignore that, can we? Reviewing, to me at least, does not stop at a post being cosmetically fine. You can write the most clear, concise and well formulated question, but if it's wrong, it's wrong.

I am no C++ programmer so perhaps I should have skipped it

Yes, exactly. That to me explains why you failed the audit, and how you can prevent it. Don't take any action if you're not certain. We have more than enough people reviewing for you to be able to skip a few.

Now of course, there is always the possibility that you are fully convinced something is just fine. And then you find that you were wrong. It's only a single audit. Don't worry, just learn from it. As long as you don't do this time and time again, there really is no big problem.

Are we supposed to be restrictive in handing out upvotes? Should I be more careful even outside the review system and only upvote questions I know are technically correct?

Ideally, yes. Don't just upvote because it sounds good. Upvote because you know it to be good. Now, admittedly I have upvoted things in the past that I thought were perfectly fine, only for the answer to demonstrate that I understood as little at the OP. It happens. But just try to evaluate correctness to the best of your ability, and vote accordingly.

Do anyone else agree that this might have been a good question? How is this even a NARQ?

Well, no. But not in the sense of "what a drama of a question". At first glance it looks fine. But, as the comments indicate, the whole premise of the question is flawed. Which in turn perfectly explains why this is not a real question. If your question is based on the assumption that "this is of course wrong" and the answer subsequently is "Well, it isn't", you arguably don't have much of a question.

share|improve this answer
2  
I don't agree with your conclusion. The one asking have no way of knowing they are wrong. In this case the user had seemingly working code. Other people may do the same or similar errors. –  Emil Vikström Apr 26 '13 at 12:27
    
Sure, others may make similar mistakes. And I don't blame the OP as such for bumping into this problem (though it would have been resolved by reading the documentation). That does not make the question better though. –  Bart Apr 26 '13 at 12:36
    
I may be unsatisfied with your conclusion but your answer is nevertheless insightful. Thanks for taking your time to answer my rant! –  Emil Vikström May 3 '13 at 14:24
    
@EmilVikström Haha, more than welcome. –  Bart May 3 '13 at 18:31
add comment

You must log in to answer this question.

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