8

Yesterday, a fairly new user (4 months, 9 reputation) posted a question about rounding in C#. When reading it for the first time, I had a feeling that I knew what they wanted to ask, but unfortunately the text was full of weird grammar, to a point where it asked a different question than I presumed was intended.

It didn't help that the examples of the desired output they gave included some crucial slips, and clarifications added to the output rather looked like alternative outputs contradicting each other - in short: it was a confusing mess.

What I understood is they were (and still are) looking for a pair of rounding functions with built-in precision determination. But the only code they found or wrote (and included in the post) so far was such that you had to specify the precision as parameter. They claimed these functions "do not work" (I could see these didn't do what they presumably wanted; coincidentally they did what they approximately wrote they needed).

All this felt like the result of a language barrier to me. Assuming the different intention and giving them more leeway in phrasing (e.g. "they do not work" meaning "they do not work for me" as in "they are not what I need"), the whole question did make sense to me.

Anyway, multiple users (including me) in the comments tried to get confirmation that our intuition about the presumed intent was correct, and guided them towards edits such that the text would eventually precisely reflect that intent.

Knowing how intimidating Stack Overflow can appear to new users when a question gets closed, I also left a preemptive comment that closing due to lack of clarity was probably imminent, but gave guidance on what they needed to change to fix that. For the time being however, I voted to close the question due to lack of clarity. Eventually the question was closed for that reason.

As I already knew what the edited question would become to be about, I prepared an extensive answer, ready to be posted as soon as the question was re-opened after the edits.

Meanwhile, OP made multiple incremental edits as suggested by the comments, clarifying what they wanted, correcting the mistakes in the examples, improved formatting, explaining how what they have is not what they need, etc.

As soon as the question was fixed, clearly understandable and thereby also distinct from a suggested duplicate, I immediately voted to re-open (and at least by now, a second person voted the same).

But I just came back to the question, to find it failed the re-open review, supposedly because "the original close reason(s) were not resolved". I do not understand that reasoning. As far as I can tell, all clarifications and suggestions from the comments were incorporated into the question body. So - what should I do now?

16
  • 3
    After reading through the question, I agree; it seems clear now. There are some conceptual problems (floating point numbers don't exactly have a specific number of digits), but that's a reasonable misconception and really the sort of thing best addressed in an answer. I cast the third reopen vote.
    – Ryan M Mod
    Commented Apr 19, 2023 at 3:50
  • @RyanM First of all, thank you, I could finally post my answer. But secondly: apart from this specific question (that tag was not added by me) where your vote was sufficient to re-open - what do I do in general? Create a post on Meta every time? Flag over on SO for moderator attention? Just wait until hopefully a random third person comes by? Bump the close voters in the comments?
    – LWChris
    Commented Apr 19, 2023 at 6:14
  • It feels so weird that my close vote is direct, but my re-open vote (which feels like the inverse action) is indirect. I am trusted to judge that I see a reason to close, but not trusted to judge that this reason no longer exists after an edit.
    – LWChris
    Commented Apr 19, 2023 at 6:21
  • 4
    I don't understand the "direct" vs "indirect" distinction. Both close votes and reopen votes are simply votes; there needs to be enough of them before the requested action is taken.
    – tripleee
    Commented Apr 19, 2023 at 7:06
  • 3
    I'm not sure I understand your direct vs indirect. It needed 3 close voters to close it: stackoverflow.com/posts/76033141/timeline and four re-open reviewers stackoverflow.com/review/reopen/34234755 to leave it closed.
    – rene
    Commented Apr 19, 2023 at 7:06
  • I applaude your effort to salvage that question, both on main and meta.
    – rene
    Commented Apr 19, 2023 at 7:24
  • 1
    Does this answer your question? How do I get a question reopened?
    – Abdul Aziz Barkat
    Commented Apr 19, 2023 at 7:58
  • 2
    @rene I think my mental model of voting was just wrong. In my head, it was a scale, and if the scale is +3 on the close side, the question is closed. When voting re-open, I removed my +1 close vote from the imaginary scale, i.e. new balance is +2 = "not enough". I expected the question to re-open (basically as if it had never been closed after never reaching 3 close votes). But it's not how it works and I'm fine with that, was just confused if you imagine a tally going up and down.
    – LWChris
    Commented Apr 19, 2023 at 8:03
  • 2
    @AbdulAzizBarkat I don't think voting to close this as a duplicate is appropriate. One of the purposes of the Meta sites is to allow issues around actions (like closing, reopening and deleting) on specific questions to be addressed.
    – Adrian Mole
    Commented Apr 19, 2023 at 8:04
  • @AbdulAzizBarkat For the general procedure, yes, thank you, that's exactly what I wanted to know for the future.
    – LWChris
    Commented Apr 19, 2023 at 8:08
  • 1
    @AdrianMole Ah, yes I mostly wanted to address the comments asking clarification for the general case. I'll retract that close vote and leave the auto comment there to address LWChris's comment.
    – Abdul Aziz Barkat
    Commented Apr 19, 2023 at 8:36
  • The main effort on that Stack Overflow question should be spend on finding the canonical question (perhaps with an intermediate closing as unclear to surface the real question). Questions about rounding abound on Stack Overflow and there are some questions with very, very good and comprehensive answers. Especially concerning rounding of boundary values, like 237.05 (to 237.1 or 237.0).
    – Peter Mortensen
    Commented Apr 19, 2023 at 8:44
  • @PeterMortensen Beware however that this is not "rounding" but rather ceiling/floor to a precision to be determined automatically based on the input. That last bit is what makes this question unique compared to other rounding questions I found, but also impossible to solve.
    – LWChris
    Commented Apr 19, 2023 at 9:03
  • Related: meta.stackoverflow.com/questions/402422/… Due to less chances of reopening, If you think it's salvageable, I would suggest that you shouldn't vote to close. It's easier than convincing reopen reviewers or the meta. Even if you manage to get it open(as you did), just look at the amount of effort put. Withholding the VTC or salvaging it yourself is much easier
    – TheMaster
    Commented Apr 19, 2023 at 11:25
  • @TheMaster For small edits, I agree. "Improve over remove" as I call it. But for new members, I prefer to guide them towards writing a good question themselves rather than "silently" tidying up behind them. And: in this particular case, the difference between what they wanted to ask and what they actually typed was quite steep. I think non-OP edits that fundamentally alter the question asked are generally frowned upon ("not representing the asker's intention" or something similar).
    – LWChris
    Commented Apr 19, 2023 at 17:39

2 Answers 2

Reset to default
6

To me, this question (especially the title) still seems unclear, as it did at the time of my "Leave closed" review.

The title suggests (to me, at least) that the decision to round up or down should depend on the (value of the) last digit (the phrase, "according to the last digit"); however, the examples given show cases of the same two numbers being rounded both up and down.

The answerer (and poster of this Meta question) has assumed that the rounding direction is something that is specified but that is not mentioned anywhere in the question – hence my decision to leave it closed.

Perhaps my review would have been better had I left a comment explaining my decision. I did (briefly) consider making an edit but I wasn't sure enough that my understanding of what the OP actually wants is correct.

There is also the issue that the C# language has a distinct decimal type (the term used in the title) but the OP's shown code uses the double type.

3
  • I see. My language intuition tells me that "decimal value" simply means "a value with decimals", and "according to the last digit" means "such that only the last digit is removed". I agree it was too far fetched without context, but everything else about that post (talking about 2 distinct functions to round up/down, having the same number each rounded up and down once in the examples) points in that direction. I think the "according to last digit or" bit could be safely replaced with "to" in the title to nail what they wanted to ask.
    – LWChris
    Commented Apr 19, 2023 at 7:55
  • 1
    @LWChris Your interpretation is most likely correct. However, questions on Stack Overflow shouldn't rely on potentially ambiguous interpretations and guesswork but should be complete clear in themselves.
    – Adrian Mole
    Commented Apr 19, 2023 at 8:01
  • Yes. To be perfectly honest, I remember I read the edited body, and seeing the "to one less digit than the input" had been added to the title. I may have overlooked that "according" part in the title.
    – LWChris
    Commented Apr 19, 2023 at 8:06
6

From the other answer:

The answerer (and poster of this Meta question) has assumed that the rounding direction is something that is specified but that is not mentioned anywhere in the question – hence my decision to leave it closed.

It was specified from the initial revision of the question, namely in two distinct method signatures RoundUp() and RoundDown().

This is why reviewers must continuously ask themselves:

  • Do I know that this question is unanswerable in its current form, or
  • Is it, at a first glance, not crystal clear what is being asked here and do I not have the experience/determination/English as a second language skills to try and read what's actually being asked here and can I help them by editing?

If the former, you should of course feel free to close-vote and to review as "remain closed". Otherwise, there's no harm in clicking Skip.

1
  • 1
    And I would also say: have some leniency. If it can be understood with a reasonable guess, just let it happen. Or at least post a comment to ask for clarification instead of immediately voting to leave it closed. The effort should be to get questions which are not doomed reopened, the effort should not be in keeping them closed.
    – Gimby
    Commented Apr 21, 2023 at 12:33

You must log in to answer this question.

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