Hi all,

Just read an answer in SO and 'seemed' to learn something new about Java references.

But did I?

Point being, in Wikipedia, the more external references an item has the more valid it 'seems'.

I think that a concept as this should be incorporated to the SO sites in order to add more validity to answers - when possible of course.

I would also encourage this with some reputation points.



Added screen shot from Wikipedia to be a bit more clear. alt text http://www.freeimagehosting.net/uploads/e4e051b1a9.jpg

Edit 2:

Reputation had been raised as a possible incentive but it is not the point. The point is that arbitrary users of differing knowledge and experience can choose to up-vote a particular answer over others where in actuality that answer is wrong. Later on someone can stumble upon this while googling up their pain-point of the day and go about to get the wrong impression due to the fact that, say, 10 people preferred one answer over the rest, and that is if he's lucky, since most questions on SO don't receive such a high amount of participation.

You (out there that have replied by voting against my suggestion) can then go on by saying that if the masses think something is right then it must be so. But that is not always the case.

My suggestion is to find out a scheme where people on their daily hunt for reputation points do a more thorough job.

Wikipedia faces a similar obstacle with their quest for the truth, I believe - and I do not always take all that is written there at face value. I have a custom to check the reference count in order to try to assess the credibility of an item. But then that might be my flawed outlook...

  • Love all you down-voters who deem it unnecessary to leave a comment. – Yaneeve Apr 13 '10 at 13:15
  • 14
    All a down-vote means on meta is that someone disagrees with your idea - not that it's necessarily a bad post. – ChrisF Apr 13 '10 at 13:34
  • 4
    @Chris, don't forget that you also get downvotes for bad posts though also. – Earlz Apr 13 '10 at 16:18
  • @Earlz - true. Perhaps I should have removed the "all" and inserted an "also". – ChrisF Apr 13 '10 at 17:07
  • Hey, can you leave me enough reputation points to offer a bounty at least ;-) – Yaneeve Apr 13 '10 at 17:14
  • Enjoy the bounty points :) – Yaneeve Apr 15 '10 at 12:41

It is common for good answers to link to other resources, and this is part of the motivation for people upvoting them.

What you want is already in place: we can create links and lists, we get reputation for including them.

If you know an especially great reference to include in someone else's answer, leave it as a comment. If it gets 10x "great comment" votes, it'll count towards your Pundit badge.

The point is that arbitrary users of differing knowledge and experience can choose to up-vote a particular answer over others where in actuality that answer is wrong.

People suck; always have, always will. They will upvote and downvote for reasons others find inexplicable or just plain wrong.

For example, I got two unexplained downvotes just today, on two different answers. One of those was even late in the day, bringing my upvote rep to 198 after hitting the rep cap several times over (38 upvotes total). So I'd be a bit annoyed if today doesn't count as hitting the rep cap, but it's not the end of the world.

That same attitude has to be applied to votes as a whole: take it as an indication of the community's attitude, but not gospel. Always evaluate yourself, if it matters. You can never delegate your responsibility to someone else.

  • 2
    Specially links to other Stack Overflow questions should be encouraged – Daniel Lopez Apr 13 '10 at 14:28

External references shouldn't especially contribute to reputation.

There is a different goal on both sites. Wikipedia wants to have the most references, to prove its information, to rely on sure things. The SO family relies on giving answers to questions, proven or not.

Stack Overflow and its brotherhood relies on reputation, and votes, to determine whether a question or an answer is good or not. External reference can be a plus, but most of times, it doesn't fit. Only few questions are about general facts, and are good to back with references, official sources.

The rest is mostly about simple questions, personal ones, dedicated to a particular situation. And for these ones, I would say that the least links there are, the better, almost. It is better to have a fully explained solution on SO, dedicated to the question, than a link to some official generic thing.

For most questions on the SO family, there are simply no references. The point of these sites is to create references for particular questions and situations.

  • the "not proven" is what bothers me... – Yaneeve Apr 13 '10 at 12:42
  • No doubt you have a point, yet in those cases where people can add references I think they should and I think that reputation might be a good incentive. – Yaneeve Apr 13 '10 at 12:48
  • @Yaneeve - in this case, they will be quality answers, And people will reward them by voting for them. – Gnoupi Apr 13 '10 at 12:53
  • @Yaneeve, if everyone needs to link a reliable source to say something then no one will say anything, because all the reliable sources will need to link to another source, and so on. – Thomas Bonini Apr 13 '10 at 14:10
  • @Kop: First, I didn't specify references as a prerequisite to reputation. Second, it is my opinion that the more an issue is based upon the more reliable it becomes. For example when speaking about the nature of Java references in the JVM, a link to Sun's JVM manual, for example, might give an answer more credibility. – Yaneeve Apr 13 '10 at 14:23

A stackexchange website that to some degree enforces this philosophy is Moms4Mom where they have a policy called Back it up:

Stated simply, Back It Up! means your answers must be based on:

  • Something that happened to you personally, or
  • Something you can back up with a reference.

It does add value to the site, but I don't think it really applies to SO for a variety of reasons:

  • SO consists primarily of professionals talking to other professionals. Others sites consist of laypeople (or people of unknown skill/experience/training) providing information.
  • Even in the cases that a layperson person is providing info, other professionals quickly check the answer and give their opinion on its correctness.
  • Even when that fails, in most cases the OP can test out the solution and it either works or it doesn't - we work in a profession where results can be tested and evaluated immediately. This can't be done in Wikipedia or Moms4moms (where the results sometimes take decades to appear!)
  • We encourage everyone to answer. If we required (or even rewarded) external references then fewer people would answer (even though they may know the answer) and we'd have more reputation gaming with people who list all sorts of references, none of which are actually useful or salient.
  • It would require people to go off site to check if references are valid, rather than just checking out if the answer itself is valid (ie, another piece of information to check).
  • It may result in more answers consisting primarily of links with little, no, or inadequate information in the answer itself to resolve the question without checking out every single link.
  • It raises the question as to what is a valid reference. A blog post? A tweet? Another forum?
  • What about those things that don't exist in easy reference online, but are really only documented in books?
  • It encourages users and visitors to go offsite to resolve their problem, or future problems. We don't want to encourage a walled garden, but it doesn't serve the community to send good programmers to other resources without good reason. Let them document the answers here.

But I'm primarily concerned about how such a system would be set up to award rep while limiting gaming without increasing the moderator's roles. I don't see an easy way to set it up that wouldn't enable me to get scads of reputation simply by linking resources willy-nilly all over the place. Are we implementing an additional voting mechanism so people can vote in individual references? Are we just giving rep away for any old reference? What's the mechanism we use to vet all the references?

  • 1) Nice link 2) As for your concern, that is why I posted my idea/concern here, to start a discussion... – Yaneeve Apr 13 '10 at 17:39
  • 1
    @Yaneeve - It's a worthwhile discussion to have, and quite thought provoking. Thanks for starting it! – Pollyanna Apr 13 '10 at 17:47
  • happy to be of service (even if most do disagree with my opinions :) ) – Yaneeve Apr 13 '10 at 18:11

Just because a post or article references outside sources does not mean those sources are valid or reliable.

Giving rep for adding external references is also a difficult thing to implement, because each link's value would have to be evaluated somehow. And if there is no evaluation system in place, it would be easy to game the system by posting many unnecessary or irrelevant links.

An answer on a SO site should be evaluated based on things such as how well it solves the question-asker's problem, how well it's written, etc. Adding external references may be part of a complete breakfast answer, but the answer itself should contain the critical analysis needed to solve the problem at hand.

  • We detected an invalid domain linked in your post. This is likely caused due to sample code not being in a code block. (this message will be automatically removed when the link is fixed) – Community May 14 '12 at 5:24

Another point: Just because someone references an external resource doesn't mean that they are correct. I can easily post something that directly contradicts a specification, but to a cursory examination or a simple reference count, my answer will seem legitimate because I have linked to an official, reliable, valid document. One that just so happens to say the opposite of what my answer does.

Any and all mechanisms that are meant as incentives can and will be gamed. Whatever incentive you have in mind for providing references, be it reputation or something different, you can be assured that it will be abused.

Therefore, I am against providing incentives for something like this.

I am, of course, all for proper referencing and linking to documentation. But we can't rely on anything other than people's sense of correctness to encourage this behavior, just as it is with other beneficial editorial contributions. If you want it to happen more often, lead by example, establish a culture of reference inclusion, and let that culture naturally become part of the accepted way of doing things.

You want a social change, accomplish it via social, not technological, means.

  • +1 I liked the social angle of your answer – Yaneeve Apr 13 '10 at 17:34

Including external references is fine, but it's also really nice to at least summarize what that external reference says in your answer. Links change and break, and many of us work at places where many pages are blocked. What kind of reference link will it be in 3 years? How does it answer my question if I can't see the content of your answer?


There's another issue with this - if someone's short for time, they tend to just post a link to a similar answer rather than answering someone's specific issue. My opinion is that such answers aren't as good as someone who takes the time to tailor the answer to the OP's specific issue. However, in a system where external references are prefered, the quick answer with hardly more than a link might be judged 'better' than another answer where more effort was put into the answer.


I think you raise an important point and I appreciate the discussion that it generated here. In addition to the reducing the instances of "bad information", I believe references have the potential to improve the site and the web in general by DRYing it up. When there are definitive references available, as there are for many language and framework issues, having informal explanations lying around in SO or elsewhere, particularly without references, just makes it harder for folks to find and improve the definitive source(s).

To give an example ;-), I've been trying for a while now to get my head thoroughly around Javascript object instantiation in light of the Chrome browser and the "naming" of object classes. There are countless blog posts about this aspect of Javascript, but precious few of them reference the ECMAscript specification or any other definitive source. See http://pivotallabs.com/javascript-constructors-prototypes-and-the-new-keyword/ for an example of bad information, see Clarifying javascript prototype nomenclature and mechanism for my related SO question and see https://gist.github.com/getify/5793213 for a clarification effort focused on the spec.

I accept that any incentive system can be gamed and I don't think I'm experienced enough yet to defend or promote any particular approach to this, but I'd like to see something that not only encourages references, but discourages new explanations when a suitable reference is readily available.

  • If the information is already out there and readily accessible then it means the question was a poor SO question. It means that the person asking the question didn't do their research. The purpose of SO is for it to be filled with information that isn't readily accessible elsewhere on the web, not for it to be a link farm to other sources of content. – Servy Jun 17 '13 at 16:17
  • That's a good point, but I have a couple of followups. The first is that in many cases there is some non-trivial thought required to associate the question with the reference material, so SO is still important in that case. The second is that I haven't seen an SO policy or practice whereby SO questions are removed in the even they can be answered by simple references. – Peter Alfvin Jun 17 '13 at 16:41
  • Such questions are, as I said, of low quality, and if they have not performed sufficient research it means the question should be downvoted. There is some debate about whether or not such questions should be closed or deleted, and as such some are and some are not. – Servy Jun 17 '13 at 16:46
  • That addresses the second follow-up (thank you), but not the first. In the first follow-up point, the applicability of the reference is either not obvious or not sufficient to answer the question, so there's nothing wrong the question. The only issue is that after providing the unique information, the answer provides their own explanation of the "standard" information instead of just referencing it. That's what I think we should be trying to discourage. – Peter Alfvin Jun 17 '13 at 19:47
  • If it's clear that even with a significant amount of time and effort invested by the OP they won't be able to find the information needed, either because it's hard to find, or not clear how it's applicable to their situation, then it's not a poor quality question. It should simply be answered. In such cases just a link wouldn't suffice for an answer, either. If there is sufficient complexity that the information is hard to find, it means that the link would require detailed explanations as to how it's related, a summary, and an introduction. – Servy Jun 17 '13 at 19:50

someone can stumble upon this while googling up their pain-point of the day and go about to get the wrong impression due to the fact that, say, 10 people preferred one answer over the rest...You say that if the masses think something is right then it must be so. But that is not always the case.

I have not seen this happen frequently enough that I believe something needs to be done about it.

Please consider posting references to specific questions where this is the case so we can verify your claims.


Further, when it does happen, people can easily verify whether the solution works or not - this is programming! Fire up the compiler and try it out. If it doesn't work, go to then next answer.

  • Somethings are simple to verify as you say, some are not, like the workings of the JVM for example... – Yaneeve Apr 15 '10 at 14:16

You must log in to answer this question.

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