35

Under the SQL Server tag there are many questions that consist of some sample data followed by some expected data and a request for someone to write a query (example).

My understanding of how Stack Overflow is intended to work is that questions should include an attempt by the asker to solve their problem.

Here is a question being treated more as I would expect, with the only difference being that the first example did include a short text description of the problem whereas this one doesn't.

However there seem to be many people under the SQL Server tag who are more than happy to write people's queries for them and upvote them as well. And many of the people providing answers are very high rep members.

I doubt these questions will be of any use to anyone else in the future because they are specific to the sample data provided, and the titles are usually not very informative.

So my question is, should we be encouraging this? Or is the correct approach to downvote/vote-to-close?

That said, I appreciate that when it comes to trying to build an SQL query it's quite possible not to know what to try. For example if one had never heard of a pivot function one wouldn't know to try it. So how could a good (acceptable) question be asked in such a case?

13
  • 5
    Maybe we need a "Write my query" addition to the SE family? :)
    – Dale K
    Commented Apr 24, 2019 at 5:31
  • 16
    This phenomenon is not specific to SQL Server. There are also numerous "How to convert this dataframe into this dataframe" questions of the same taste...
    – honk
    Commented Apr 24, 2019 at 6:30
  • 3
    Related: the faq-proposed Why should I provide an MCVE for what seems to me to be a very simple SQL query?
    – AakashM
    Commented Apr 24, 2019 at 8:46
  • 2
    From what I've seen, many SQL questions take this form. And to add insult to injury, answers tend to be code-only. It's just the way it is, there are specific tags that kind of have an unspoken leniency when it comes to the site rules; SQL questions should be on-topic, but you often can't really write a question that is 100% according to the site rules just for the very nature of the subject matter. Of course if you bring it to meta then it will be condemned because rules is rules, but the general populace that answers SQL questions thinks differently. The same for HTML and CSS.
    – Gimby
    Commented Apr 24, 2019 at 11:23
  • 1
    @Grimby yeah I can get the unspoken leniency that basically answers the second part of my question. And I guess the fuzzy line of whether to downvote or not comes down to whether the OP has just dumped data expecting an answer or whether they have taken the time to clearly explain what they are looking for, even if they don't know where to start. Or something :)
    – Dale K
    Commented Apr 24, 2019 at 12:02
  • 10
    "won't be of any use" is the core problem, I think. Google just does a poor job of indexing these kind of questions, so nobody is expected to have done any research. The [regex] tag has this problem as well. In both cases a steady supply of SO users that don't mind posting a quicky answer keeps the tags alive. Notable is that their answer rate is much better than other mainstream tags.
    – Hans Passant
    Commented Apr 24, 2019 at 12:14
  • 3
    I wouldn't be apposed to a site dedicated to answering questions about HTML, CSS, regex and SQL to be honest. Questions in the format "I need this result, but I don't know exactly how to phrase it". A: "Like this!", A: "Or like this!". That site can have specifically more lenient rules about question and answer quality and duplication. That is probably what grinds the gears of most people regarding these topics; they're really hard to dupe-close.
    – Gimby
    Commented Apr 24, 2019 at 12:19
  • @HansPassant and even if the titles where meaningful, and google could index them in a useful way, the person trying to solve these problems probably won't know what to search for, and probably won't be able to translate the answer from a different data set to their own, even if the technique is the same. Oh well :)
    – Dale K
    Commented Apr 24, 2019 at 12:19
  • 3
    @Gimby I suppose the other thing is, how much are these people really learning? My guess is that some people are being paid to do a job they don't know how to do, and are getting by because they can post their problems on SO. But whether they can then support/modify/tweak their code...
    – Dale K
    Commented Apr 24, 2019 at 12:22
  • 1
    @Gimby your first comment should probably be an answer. What would be even better is if the general populace could have somewhere to discuss this unspoken leniency which would then be a spoken leniency. Because how questions are treated seems to me to depend on who is online around the time of posting.
    – Dale K
    Commented Apr 24, 2019 at 12:35
  • 1
    @Machavity - not sure that "dump a schema" is accurate... the cases I am talking about just have sample data + expected data, not even a schema.
    – Dale K
    Commented Apr 24, 2019 at 12:51
  • 1
    "how questions are treated seems to me to depend on who is online around the time of posting" - yes and no, there are definitely sub-communities tied to tags that share a similar outlook on things. It also works the other way around, tags such as C and C++ tend to see far stricter adherence to quality standards than what we're used to that also sometimes has an almost weekly meta post about it.
    – Gimby
    Commented Apr 24, 2019 at 12:52
  • 1
    @DaleBurrell I'm trying to describe the problem in general. We talk about this overall problem on a semi-frequent basis.
    – Machavity Mod
    Commented Apr 24, 2019 at 12:57

1 Answer 1

Reset to default
40

Not really, but this is the tip of a much larger iceberg anyways. SO has always had problems with this sort of thing

The problem is that the post that should be closed and downvoted was instead answered (three times) and upvoted, until the meta effect kicked in. Clearly these tools aren't sufficient to deal with this problem.

Servy comment

The proper way to deal with this is to close the question. This is why we have dupehammers, so well trusted users can close questions like this as duplicates, hopefully before any answers are given. Unfortunately, you've got a lot of people who want to Always Help

Helpers can become Always Helpers. They believe that every question deserves an answer. That's what SO is for, right? So they made a typo? Give an answer. They just described a general programming idea without any code? Give an answer. Homework dump? Give an answer. The rules don't matter, just help everyone. And they get Internet Points for doing it. What was the downside again?

As long as people just "want to help" (neither of the top answerers seem to have much interest in moderation or review) and have no interest in trying to encourage some level of effort before doing so, this will continue. Just vote to close and move on. I do that every day...

3
  • 13
    Totally agree with this answer; the problem is that a lot of my close votes age away because in the tags I frequent, the "Always Helpers" outnumber everyone else. Not a problem we can solve, just setting expectations that reader's close votes may get exhausted swiftly, with not much to show for.
    – Heretic Monkey
    Commented Apr 24, 2019 at 14:42
  • 2
    Thanks, its good to know that I'm on the right track as far as the site objectives go. I had started wondering whether I was being too hard on these type of questions because so many high rep users are happy to answer them.
    – Dale K
    Commented Apr 24, 2019 at 22:41
  • 3
    i.sstatic.net/ELnZc.jpg
    – hanshenrik
    Commented Apr 25, 2019 at 8:18

You must log in to answer this question.

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