One of the stated goals of the site was the have one comprehensive answer to each question. I think we can definitively say that it hasn't worked out that way. I think the reason is reputation. If one person answer the question with 40% of the information that a comprehensive answer would contain, and another with another 40% with 20% scattered among three other answers. Who's answer will get edited with the combined information and get the further upvotes? It just doesn't happen. People don't want to edit questions and mess with someone else's reputation (quite rightly, IMO - how would you feel getting a downvote for something added to your answer that you didn't write?).

One idea to get towards that goal would be to have a community answer on each question, with say a small reputation bump (2 points?) for participating in making it. To prevent reputation inflation, you can only get the 2 points from a given question one time, with a daily limit (say 10 points) for doing it on questions asked by others.

Then when the question is seen for posterity, it will be the first answer.

The issue is what to do with the accepted answer? One possibility is to only open the community answer once there is an accepted answer and pre-fill it with that answer.

Does any of this make sense to anyone else?

  • 3
    You want an answer about the answer or an answer that is set to wiki for all and primed to be the answer?
    – random
    Oct 28, 2009 at 15:09
  • Either this can be easily misused by contributing crap to the wiki answer and gaining rep for it or I misunderstand your suggestion. Oct 28, 2009 at 15:16
  • @random, I mean an answer that is set to wiki for all and primed to be the answer.
    – Yishai
    Oct 28, 2009 at 15:20
  • @John, I agree with the misuse potential, that is why I was suggesting such limits on the rep (2 points an answer, 10 points a day max, only waived if it is your own question).
    – Yishai
    Oct 28, 2009 at 15:21
  • Sounds like Wikipedia
    – MarkJ
    Oct 28, 2009 at 18:00
  • This definitely is an issue, but I am unsure of whether this would help
    – Casebash
    Oct 29, 2009 at 8:32
  • @Ladybug Killer, you leave those cute little critters alone :'( You're breaking my heart here. Jun 15, 2011 at 8:52

3 Answers 3


There is no need to have one, single, definitive answer for each question. Perhaps you could link to the stated goal so I can understand your perspective better.

In fact, it's rather nice having many different perspectives and different ways of explaining the same or similar solutions, as well as the several solutions a single question may give rise to. People visiting the question later on may not have the same exact need, but the varied answers may provide the answer to their specific question.

Further, when you have a problem, you tend to read all the answers in a question, so there's no compelling reason to try and contain everything in one entry. Having several answers is fine for those who are struggling with the problem.

In other words I don't feel there's a compelling motive for your feature request.

Further, as you've discovered, moving the site more towards a wiki style site results in complex questions regarding reputation. It would require a significant change in how the site operates, and the benefit, if any, is minimal.


I think this would be useful in the case of questions such as those discussed in What's the rationale behind closing a useful question with 70K views?, where a question has been locked and/or closed because there are a huge number of answers. In many such cases, there are numerous duplicates or near-duplicates lurking in the undulating moebius strip of answer pages simply because people have eagerly added their contribution without taking the time to scan the 72 pre-existing answers to see if theirs is already there.

So allowing moderators to establish a single 'meta-answer' for such question would provide a way to merge duplicate and similar answers into one comprehensive (and comprehensible) whole. The meta-answer would be "sticky", i.e. be presented first. It might be necessary to only show it to those with a certain privilege level, or to visibly indicate it as a draft (or both) until such time as it's considered representative of the general body of answers to the question.

One problem I can see with this idea is that it would potentially involve a lot of tedious editing work. The amount of work needed could be reduced / made remotely practical by providing a way to mark answers as having been incorporated into the meta-answer; such a flag would presumably be disabled or visibly expired if a compiled answer was edited.

I can't really think of anything that compilers could stand to gain from such work, since it would certainly have to be considered a CW-type undertaking. Maybe a "maintainer" or "compiler" badge for people who do a lot of such edits?


My goodness, what fortuitous timing. I'd been thinking about asking about this very thing.

Besides parsing HTML and working with datetimes, the most frequent question in seems to be about the mail function. For the uninitiated, the built-in mailing function is terribly unreliable and provides no feedback when things go wrong. Combine this with the average quality of shared hosting, and it's no wonder why we get so many questions about sending mail through PHP.

Thankfully there is a known, finite set of solutions to these mailing problems. Many are solved by the various anti-spam techniques documented on Jeff Atwood's well-known blog post about mail deliverability. Many are solved by using one of a handful of modern third-party mailing libraries instead of the built-in function.

Because there are so many duplicates, it's hard to actually find a good enough duplicate to suggest a closure. So I was thinking that writing a definitive guide covering the most frequent problems might serve a worthwhile purpose.

Such things have been tried before, like "Reference - What does this symbol mean in PHP?", but with mixed success.

I think that spreading the word that such a resource exists might be problematic, followed closely by making sure that questions actually get closed as duplicates when they fit the pattern. While knowledge would spread slowly throughout frequent contributors, I'm actually far more worried about dupe closures. Some users don't care for having their questions closed as dupes, and some users think that they are special snowflakes that are totally asking a different question for which the answer is still the same. (The answer being that mail() sucks and you shouldn't use it except as a last resort.)

You must log in to answer this question.

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