At the request of Mark Trapp, I'm turning one of my answers to another topic into a feature request.
Currently the FAQ does not provide very much specific information on what exactly defines a good question. It does a decent job of explaining what not to ask, but it isn't specific enough about what should be asked. (For instance, there are four examples of bad questions, but not a single example of a good question.)

I don't mean any disrespect to the hard work that has already gone into the FAQ, but it's clear that a large percentage of community members are finding it an insufficient resource to determine what kinds of questions they should be asking. Examples of existing questions which are on-topic and strictly adherent to Stack Exchange subjectivity guidelines would be very helpful.

  • Would also be good to go down each question, point by point, and identify how they meet each point of the criteria.
    – user28988
    Commented Aug 17, 2011 at 2:49
  • 1
    This might be a good thing to have up on the blog if/when it goes live.
    – Adam Lear StaffMod
    Commented Aug 17, 2011 at 3:21
  • 1
    Yes please. I've only written one question on this site. The number of closes and the apparent inconsistency of moderation (what is elegant code was closed; what is robust code isn't yet closed) makes it impossible for me to know if a question is acceptable or not. Commented Aug 28, 2011 at 4:32

4 Answers 4


I agree. It seems that we should do absolutely everything possible to make it easier for new users to write great questions. Honestly I am not sure if Jeff's suggestion of using the most upvoted questions, or @MarkTrapp's modified version are necessarily the best, since there will be inevitable bad questions. I think that the reason a lot of bad questions get asked is because a new user comes on and sees something like "What's the best programmer cartoon ever?" (probably deleted now, but I remember that one from a while back) with 127 upvotes, leading them to ask something like "What's the best soda for programmers?" Promoting questions that really meet the quality guidelines, could really do a lot to incentivise asking better questions. To find the best questions, it seems that they really need to be selected by a real person who is fully aware of what makes a great question. Whether the questions get selected by moderators, or through the meta this really does seem worthwhile.

  • 1
    I'm envisioning a future where we adapt Jeff's method, only to realize that the most upvoted questions are all "hey guys, what the hell is on topic in this forum anyway?" and the entire site becomes some kind of Escher painting.
    – user29776
    Commented Aug 30, 2011 at 19:01

In the previous question about this, Jeff suggested that we take our cues from the top-voted questions list, but that list generally isn't indicative of question quality, just popularity.

There's a new question view that might be more helpful in coming up with a "best of Programmers" list: the Greatest Hits:

Questions that got a large amount of views or a large amount of feedback.

The current algorithm divides the number of page views with the total amount of question and answer feedback received (adding a bonus for high view counts), excluding questions with less views than the median :- 238.

There are still some questions in the list that are duds, but overall it does seem to be better than other question views at capturing the site's intent.


Is this question actually answerable?

The specific site scope is conceptual questions in software development.

The general policy of Stack Overflow is to disallow any quesion that is primarily opinion-based. That commonly gets interpreted as disallowing any question that does not have a single, unambiguously correct answer. Sometimes that policy goes wrong and rejects the equivalent of 'what is 2 + 2'.

Looking on the front page, literally every question is either subjective, trivial or a matter of definition.

Obviously, the set of conceptual question that have unambiguous, uncontroversial but nevertheless non-tautological answers is not going to be large. But is it actually empty, or merely very small?

So I would be interested if anyone could find an example of a question that clearly both should and would not be deleted.


We should permit and actually encourage certain kinds of legal questions. If the best you can come up with is "consult a lawyer" then your clients or employers are going to skull fuck you.

I've been a freelancers part time for twenty years and full time for thirteen.I've consulted lawyers lots of times but they have never done me any damn good. Accountants aren't any better. I was in a world of hurt until I learned enough about the law and business finance that I did not need to consult attorneys and accountants anymore.

While I wouldn't represent myself in a lawsuit, by understanding contract an intellectual property law I don't need to get into situations that can only be resolved by a court. The best kind of lawsuit is one that doesn't happen at all because it doesn't need to. Knowing how to read a contract, knowing when not to sign a bad contract and knowing how to get the other party to live up to their contractual o ligations through diplomatic means are all good ways to avoid ever needing an attorney.

It is somewhat better to say that the answer depends either on your location or your clients, but having said that some effort should be made

  • 1
    This question isn't about deciding what's on-topic or not: it's about finding questions that are already on-topic and can be used as examples for the FAQ as it is now. For the discussion about whether software law should be on-topic, check this question and this question.
    – user8
    Commented Aug 30, 2011 at 0:37
  • Yes, I see your point. I should have posted this as a top-level question. I read both those threads. While I agree that we do not want to encourage uninformed legal advice, there are all kinds of really basic things about the law that every code should know. None of these things require an attorney to find out about. Just paying attention to the world around you and knowing how to read and what to read will take care of it. Commented Aug 30, 2011 at 4:25
  • I don't have any manner of formal legal education. While I have had some expert advice from attorneys, I haven't gotten that much. I am confident that what I consider my legal expertise is completely valid, I learned it all in "The School of Hard Knocks", for example by working on consulting contracts for dishonest clients who were determined to rip me off. Most programmers tend to be very naive about other people and are easily taken advantage of. I want to put a stop to that. Commented Aug 30, 2011 at 4:27
  • What I do know about software and business law would be a great series of articles for my website. I'll see what I can cook up, but I cannot write them anytime soon. Commented Aug 30, 2011 at 4:28
  • I agree with what you're saying here but this definitely seems like it merits its own thread vs being an answer in this topic.
    – user29776
    Commented Aug 30, 2011 at 18:58

You must log in to answer this question.

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