44

Occasionally, a great answer is added to an old question. Too often, simply because of the age of the question and the lack of public visibility, that answer will not get any votes. Thus as it stands, the highest-scoring answer is often not the best, but simply the best of the first.

How does and how should Stack Exchange sites deal with this problem?

7
  • 5
    Or if SO is still around in 5 years, technology changes will mean that former best answers should be changed or updated to reflect the current state-of-the-art.
    – Steve Moyer
    Oct 12, 2008 at 20:09
  • 1
    Hey, that's why they invented the Necromancer badge. ;-).
    – Gamecat
    Oct 12, 2008 at 20:21
  • and that's why so many people are able to edit others' answers.
    – nickf
    Oct 12, 2008 at 21:00
  • Maybe all questions and answers should turn into wiki's after a set time period. Say 2 months
    – tdyen
    Oct 13, 2008 at 6:22
  • That's a good idea, automatically turning into a wiki after a period.
    – tpower
    Oct 13, 2008 at 7:45
  • 1
    What? Turning "old" (for some arbitrary time period) questions into wiki sounds like a terrible idea. What would the motivation be for answering old questions when no reputation will be given?
    – hlovdal
    Mar 13, 2011 at 0:56
  • Related: Solving the FGITW problem by sorting questions based on best estimate of "true" score rather than upvotes - downvotes.
    – blahdiblah
    Oct 3, 2012 at 8:35

11 Answers 11

17

Notify users when their questions are answered. This should allow the user to revisit their older questions and choose the better result as the correct one.

2
  • 2
    The issue I see with this is that it only counts for 1 vote. I acknowledge that the person asking the question could make the new answer the accepted answer, but that does not really have the community opinion that I think people use SO for.
    – Stephen Bailey
    Oct 12, 2008 at 20:34
  • Is this a request to have SO notify users when there's an extra answer to their questions? I think that when an answer you've given gets new comments, it shows in a different colour on your personal page; I'm not clear that it does the same for new answers or comments on other people's answers. Oct 12, 2008 at 21:07
10

I would give some +1 bonus reputation (only few times per day) when one votes a recent answer to an old question.

Moreover, I would make the Necromancer badge a gold one. Well, a platinum one. This would be an incentive for people to keep answering old questions. And if you don't want to give a redundant answer, you must read other people's answers first.

Maybe some more necromancy:

  • Awakener (Reviver?): bronze, +5 after 20 days
  • Necromancer: silver, +5 after 6 months
  • Raiser of the dead: gold, +10 after 1 year
5

The SO/SE engine and the scope of the SE network sites has changed and expanded quite a bit since this question was first written and answered, so I wonder if others have new opinions on this topic.

My take: Just as commitment decays over time in Area 51, up-votes for answers should decay over time. That way, if a new answers comes in after a year and receives 2 up-votes, that could count the same as, say, 4 up-votes for an answer posted in the first few days after the original question was posted.

Note: I am not proposing that the reputation gained from an answer decays. Just that the answer shown as the top (or second-to-top if any answer has a check-mark) be modified, as well as perhaps showing the decay-modified vote count somewhere. Perhaps this last bit could be a privilege, similar to the privilege to see up-votes and down-votes separately.

1
  • 2
    I like the ideal of using "upVotes per 100 views" rather then number of upvotes to sort answers. Dec 6, 2011 at 11:00
4

I like Dimitry Z's answers, but I might extend it a bit further so it caters for more than just one vote : to notify everyone who voted on any answer that a new answer has been added.

Would that create too much spam ? Perhaps not given the fact that the person cared enough to vote.

2

Votes aren't that important.

If you contribute for some days you will quickly have a stock of answers. These gets voted even days or weeks after you wrote them, and you'll very quickly face the situation that the maximum of 200 points you could get per day are used up overnight.

I get surprising comments on answers I wrote for old questions. These questions may look dead and you don't get the fast kind of feedback that you get from the newest ones, but if the question is good or your answer is good the feedback will come. Either you get points for a week old answer or an interesting comment makes you rethink your question.. Sometimes you get nothing. Who cares.

For me that works. The point-hunting game was fun in the first days but it quickly lost it's fascination. Instead I now answer because you want to learn, have nice chit-chats in the comments and give back to the community.

1
  • 1
    It is not the points that I am interested in, but the value I find in the relative vote that answers are given. I typically read them to be the community opinion on a question. So typically I look at the highest scoring and the negative scores to see what people think is good and terrible.
    – Stephen Bailey
    Oct 12, 2008 at 20:36
2

Perhaps sorting w/in a vote level is by newest first, rather than oldest first. Say, for instance, that you create a new answer. It will appear before the other answers at the same score (0).

1

To throw in a possible idea, perhaps there should be some sort of frequency metric votes/view ?

But this may be flawed in that the views are currently counted on questions not answers ( Well actually threads, but you can be sure the question was not read not all the answers ), but still perhaps something like that.

1

On this topic, I find that some old questions have incorrect answers. I try to leave a comment on those questions. Most of the time this seems ignored.

To be honest with you, who made me the authority for determining what is the "correct" way to write a singleton pattern for c#? I did my best and chose the best answer at the time. Stuff changes and a new "more correct" answer may pop up. Just because I asked an question first, does not make me the authority for "the answer", in SO it does.

I think that after a certain amount of time, questions should become community owned. And "the answer" should become community owned (by rolling up all of the data into a single answer)

Perhaps we should allow for duplicate questions, if the new one is more concise with rolled up answers (from possibly multiple questions) - referencing the original.

Perhaps the answer is just to comment-and-wait.

Perhaps we need a meta-moderation scheme.

Perhaps we should allow people to vote that a question is not really answered.

2
  • 1
    This is why I upvoted you: "Perhaps we should allow people to vote that a question is not really answered." Oct 13, 2008 at 6:36
  • That is why I would not typically mark an answer as the correct one, because who am I to say that is correct ( I might do it however if I though that the votes were wrong... but really that is anti community )
    – Stephen Bailey
    Oct 13, 2008 at 7:01
1

Amend answer sorting by votes so that answers with a bounty awarded for being exemplary gain upvotes (for sorting purposes only) equal to the amount of the bounty divided by the reputation gain for an upvote.

e.g. 250 bounty = 25 upvotes for sorting purposes.

(A different "conversion rate" could be used if analysis shows this would be too 'heavy').

enter image description here

Anyone with sufficient reputation (and altruism, I guess) who encounters a great answer to an old question can thereby reward the answerer, and boost the chances of future visitors benefiting from it as well.

0
0

For me, anyway, I have found the way books are rated on amazon to be very helpful. I've also become more critical of reviewers as time as gone by (I can think of one computer book reviewer who must spend his entire life reading basic texts and never gives less than 4 stars). On amazon I take notice of 1: the number of reviews 2: the star ratings of the reviews 3: how likely the reviewer is to be a sock puppet.

0

If the question is popular and the new answer is a really good one, it could be voted up by people who stumble upon the question and do not think that the top-scored/accepted answer is the best one.

You must log in to answer this question.

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