Evidence-Based Guidelines for Writing Questions on Stack Overflow
arxiv.org
arxiv.org
I love that the questions are not "good enough" for the community, but still good enough to milk for search placement.
It's a useful place, but I have to wonder if its main purpose was to suck up all the available SEO oxygen and concentrate it. I think they even still nofollow the highly relevant links to other sites, because not doing so would dilute their position as the relevant result for virtually any programming query.
EDIT: After a few seconds of googling, it looks like there has been effort on the nofollow front since I last checked. That's cool. I can see that it's nuanced. Still, my current experience with arriving at a "bad" question with great SEO is super frequent
Stack Overflow is aging pretty poorly.
When you're a guest, you follow their rules. And their rules are fairly simple:
- Stay on topic
- Don't duplicate existing questions (look before you ask)
- Ask a specific question ("how do I fix it?" is not a specific question)
- Be polite (no one is getting paid to answer your question)
- Don't be a spammer (and avoid looking like a spammer)
With coding questions, it really helps to:
- Reduce the problem to a specific detail and be concise, if I haven't finished reading the question and have an idea of the answer in 30 seconds, I move on.
- Provide enough detail that someone else could reproduce your issue. So many people assume I have remote access to their laptop, not to mention what they are thinking inside their head.
- Write it in English, with punctuation, and capitalization. If there's a 10 line block of text without a single capital letter or period, I move on. Same goes for anyone with their capslock key stuck in the on position. I try my best to understand non-native English speakers, because I sure can't talk in their language, but sadly when there's a language barrier, I move on fairly quickly.
Yet, a closed SO question is almost always the first Google result.
What is the correct place to find this type of info?
But since they decided to nix 95% of the useful questions I rarely go there anymore.
Why do they need to separate clearly popular and related questions onto a graveyard site? Sure they might go stale quicker, but I have a load of questions and answers from Android 2ish that are totally irrelevant now and they are allowed to stay apparently.
Stack overflow doesn't have any solution for any answers going stale so I don't think they should use that as a reason to pick on questions that pretty much everyone except the pedantic editors think are fine.
I swear to God I was so pissed off. Here I am , genuinely asking for help, and then these tech oligarch's were blocking me sitting in their ivory towers.
Hope this paper helps me out!
SO wants curation and objective answers to objective questions. Vague or poorly-scoped questions? Toss em in the bin. Questions about "best way" "best library" or "best tool for"? Subjective or not-software -- in the bin.
As a question-answerer, a frustrating question would be one that tries to articulate a bug/problem but never cites any code, compiler error or runtime error. When posed by anyone who is not already an expert these questions very often cannot be addressed without a roundtrip involving "ok, please share the code so we can reproduce the problem".
No joke -- that tip about the MCVE or SSCCE was super valuable. Unfortunately I'd bet that just the act of making this example would probably yield an answer to the person's question most of the time, without even needing SO. Novice question-askers lacked the expertise/experience to know to do this or just the patience required. I'm not faulting them, I remember what it was like to be a novice. It's a reasonable expectation to turn to a mentor and show them the code and ask them what you're doing wrong.
> Hope this paper helps me out!
> > Results: We found that regardless of user reputation, successful questions are short, contain code snippets, and do not abuse with uppercase characters. As regards affect, successful questions adopt a neutral emotional style.
This matches my experience. But I'd love to hear how the results might change if they filtered on low reputation question askers.
When HN bashes SO, I feel conflicted. I totally understand why so many folks are disappointed. But I also think that it's good for what it wants to be.
the first viable answer is giving you the solution, except it doesn't work because the API has changed many years ago.
the second viable answer is telling you that you don't want to do that and it's certainly wrong to try.
the third viable answer is a guy advertising to use another library for just that. (Just because you use C++ doesn't mean you have boost or Qt available but fair enough).
The fourth viable answer is a guy advertising for his own library to solve that same problem. It may or may not be shameless advertising. The github link is dead half the time.
The topic is closed. It will never be possible to post the right valid answer to that problem.
Answers cannot be edited. They will rot forever and mislead an entire new generation.
I never ask questions in SO. To be honest, I'm not dissatisfied either, I just gave up. When I saw a question for which I knew the exact answer, I was unable to answer because... I don't know. I wasn't reputable enough to answer and I couldn't get a reputation because I couldn't answer. Or something like that, I don't remember or I don't care. I just google for answers.
So I would reconsider the "invariably". But that's just my two cents.
Hasn’t happened to me, but nothing is more frustrating that having someone else ask my exact question then it gets flagged.
a tightly defined, clearly written question with a single correct anwswer?
Rejected: RTFM
a question with more than one correct answer (but which almost certainly has a "best answer" which a community like SO is in the obvious best position to judge)?
Rejected: subjective; opinion-based
no axe to grind here--i used to participate in SO (50K rep, and one of the highest rep-to-post ratios)--i honestly just find it too difficult to navigate that narrow straight
there are people (such as the comments above) get judgmental just because a post didn't use capitalization. sure, nobody is getting paid to answer them (well SO kinda is), but there is no reason to berate people for asking questions and missing minor details. just move on.
Or you replied and your answer got deleted for being a duplicate. It was not a duplicate, it was the solution in python3 while the existing solution is python2.
edit: Just remember another one. That horrible time when I couldn't post the reply the OP was looking for, because the answer had to be at least 100 or 200 characters and the solution wasn't that long.
But yeah, in the tags I stick to, many Qs are incomplete. A bespoke template or "wizard" triggered by the selection of a tag would go a long way.
But neither this nor the tips in the paper solve the issue of how to draw attention (and answers) to a Q with high-volume tags.
If you don't know the answer to a question then anything that works is great, regardless of how good it actually is (or how relevant it is X years later).
I'm also convinced a huge number of ARBITRARY_PROGRAMMING_LANGUAGE SO contributors are just random students googling the answers for sweet, sweet internet points (if I was being generous, to build themselves a 'profile' for job hunting or something) but that is another thing entirely.
Can't submit because an edit has to be at least n character changes.
It is basically confirming simple online etiquette produces best results.
Possible rationale: They get more attention
We need to study such things objectively. Also results may change over time as internet users get used to things
But a non common sense result would lead to more studies, that would either confirm or disprove it. That larger body of studies would very likely be trustworthy.