Thoughts on "Accepted Answer"
-
Two things are important: telling people with the same question "go here" and telling people looking for questions to answer "this one is solved, no need to read". It's the one thing I was missing from the CP boards from the rocky starts. However, all Q&A sites I've seen have draconian incentives for askers to come back and "close" questions (e.g. EE has topic moderators force-accept answers, SO gives reputation and shows "acceptance rates" of the poster with every question. I don't rememebr CG that well, but there was something, too). So, corollary: people who come to web sites to ask questions don't come back to tell if it helped. I'd say: - keep it - encourage (e.g. give users a link to their unanswered questions when they log in, or when they post a new question) - "somehow" separate the dead from the active questions.
Personally, I love the idea that Raymond spends his nights posting bad regexs to mailing lists under the pseudonym of Jane Smith. He'd be like a super hero, only more nerdy and less useful. [Trevel]
| FoldWithUs! | sighist | µLaunch - program launcher for server core and hyper-v serverpeterchen wrote:
Two things are important: telling people with the same question "go here" and telling people looking for questions to answer "this one is solved, no need to read". It's the one thing I was missing from the CP boards from the rocky starts.
Spot on! And that's where we want to be. We have suggestions that appear after you've typed in a question. Hopefully that helps to reduce repeat questions. I wonder how people are actually finding it. The other "important" thing is our current concern.
-
What exactly is the problem? Is it that very few answers are marked as accepted (meaning only a small number of questioners are doing it), or that the questioners don't seem to understand the concept?
.45 ACP - because shooting twice is just silly
-----
"Why don't you tie a kerosene-soaked rag around your ankles so the ants won't climb up and eat your candy ass..." - Dale Earnhardt, 1997
-----
"The staggering layers of obscenity in your statement make it a work of art on so many levels." - J. Jystad, 2001Both :(
-
I'm glad you asked. It's probably fine how it is... the problem seems to be the users, and changing them isn't going to happen. However, you can guide them a bit. Here are my ideas. Each answer must also come with a summary/title (basically, a short version of the answer). Next, the questioner would only get to see the first answer "free of charge". Only the title of the other answers are shown (to anybody). In order to expose the second answer, the questioner must select either "Answer Was Useful" or "Answer Wasn't Useful" ("was useful" would "accept" the answer). And in order to see the third answer, they must click one of those buttons on the second answer. And so on. And an answerer can see all of his/her answers regardless of if they've been accepted/rejected (and delete them before the questioner gets a chance to see it). If an answer has not been accepted/rejected for a week, it will automatically be made viewable. Notifications/updates are also something to consider. Posting an answer shouldn't move it to the top of the list of latest questions, unless it is the first answer and it is immediately viewable (or if all previous answers have been accepted/rejected, in which case posting that answer would mean it would become immediately visible). When the questioner accepts/rejects an answer, then the next answer becomes visible and causes the entire question to go to the top of the "latest questions" list. And in case you are wondering why I think answers should have titles, it would be to give others an idea of what answers have already been posted, so they can at least avoid repeating an answer without actually seeing the full answer. P.S. "Accept" is an ambiguous term for answers. What does it really mean? I can see a questioner interpreting "accepting" an answer as meaning that the answer gets posted for others to see (maybe they think they are the only ones who can see answers until they accept them). I've seen evidence of this with clearly useless answers that have been accepted by the questioner.
Interesting but I'd have to argue against the idea of forcing users to vote. Often the user is unqualified to make a judgement (for various reasons) and so shouldn't be forced to arbitrarily choose one. I often find myself in this situation.
aspdotnetdev wrote:
And in case you are wondering why I think answers should have titles, it would be to give others an idea of what answers have already been posted, so they can at least avoid repeating an answer without actually seeing the full answer.
Good point but won't it be superfluous in many cases? And I wonder how many users wouldn't actually fill this in correctly.
aspdotnetdev wrote:
"Accept" is an ambiguous term for answers
Good point! Suggestions for alternative names?
-
Thiru Thirunavukarasu wrote:
But I'd have to admit that this is also often not the case, in practice.
Well... Only the person asking can really judge if a given answer worked to fix the problem as he actually encountered it. But yeah, that doesn't mean he's capable of picking the best answer to the question he actually asked. IMHO, you should keep it available as a hint to future readers, but make it strictly optional. Many, many users will never come back and "accept" an answer at all, so the more weight you give such a mark the more tempted you'll be to implement some sort of auto-accept feature (based on voting or something)... which then makes the mark completely meaningless.
Thiru Thirunavukarasu wrote:
- Can we salvage the feature but improve it somehow?
A lot of Q&A sites implement something like this, and... Often, the "accepted" answer just hints at what the final solution was for the user. Sometimes, the user comes back and posts his own answer, describing what he finally had to do to solve the problem, but even when that happens it ends up separated (on the page) from the "accepted" solution. It'd be cool if, rather than just allowing users to "accept" an answer, you would prompt them for a bit of extra information, and display that below the accepted answer. Sort of an epilogue. Granted, most users probably wouldn't bother entering anything, but then most users probably won't bother accepting answers anyway - it'd at least give the feature some additional utility in cases where the person asking the question was conscientious about it.
Shog9 wrote:
Sometimes, the user comes back and posts his own answer, describing what he finally had to do to solve the problem, but even when that happens it ends up separated (on the page) from the "accepted" solution.
Excellent point - a completely valid scenario that no site (I've seen) has successfully addressed! I'll discuss this further in my next post.
Shog9 wrote:
rather than just allowing users to "accept" an answer, you would prompt them for a bit of extra information, and display that below the accepted answer
Not a bad idea but, like you said, it simply wouldn't be used.
-
For Quick Answers we've implemented an 'Accept Answer' feature. Perhaps some of you aren't aware of this and if so that's part of the problem. Let me briefly explain what it's all about. After a member posts a question and receives at least one answer he/she has the ability to click 'Accept Answer' on any of the answers. Our thought is that this would be useful for answer seekers who will be able to more easily identify the best answer(s). It may be the case that the author of the question is best able to judge what is the best answer for his/her question. But I'd have to admit that this is also often not the case, in practice. Accepted Answers as its been implemented and how it's being used just isn't working out well at the moment. Here are my main concerns: 1) Does provide value to the site? 2) If not, should we simply remove this feature and just leverage the best voted/rated answers by the community? Maybe this is sufficient or even more accurate? 3) Can we salvage the feature but improve it somehow? 4) Other thoughts?
Thanks for your thoughts guys! It looks like the concensus is to keep the idea of 'Accepted Answer' around but that it needs to be tweaked somewhat. I'd like to bring up a few notable ideas. Maybe we can discuss these in a bit more detail. 1) The phrase 'Accept Answer' may be unclear. What can we possibly rename this to? 'Useful Answer'? But we also don't want to encourage users to "accept" too many answers (or do we). The original idea is to have one or very few answers as directly addressing the question. I.e. a question with a useful/best/accepted answer with perhaps some equally useful alternatives. 2) Should we allow the question author to post an answer to his/her own question and accept it? Shog brought up an excellent point (here[^]) where this is a totally valid scenario. Maybe we could colour the background of such an accepted answer slightly different than regular accepted answers? Also, there would be no reputation awarded for accepting your own answer to prevent users from gaming. 3) Other thoughts?
-
Oh, and in addition to "Answer Was Useful" and "Answer Wasn't Useful", perhaps offer an "I Already Have My Answer" button (would really do the same thing as "Answer Wasn't Useful", which is nothing but display the next answer for all to see).
How about a "I was just trolling" or "My answer was urgentz. Was. Is no longer urgentz".
cheers, Chris Maunder The Code Project Co-founder Microsoft C++ MVP
-
How about a "I was just trolling" or "My answer was urgentz. Was. Is no longer urgentz".
cheers, Chris Maunder The Code Project Co-founder Microsoft C++ MVP
Those are so common they should probably just be the defaults. The user would have to indicate otherwise.
-
Interesting but I'd have to argue against the idea of forcing users to vote. Often the user is unqualified to make a judgement (for various reasons) and so shouldn't be forced to arbitrarily choose one. I often find myself in this situation.
aspdotnetdev wrote:
And in case you are wondering why I think answers should have titles, it would be to give others an idea of what answers have already been posted, so they can at least avoid repeating an answer without actually seeing the full answer.
Good point but won't it be superfluous in many cases? And I wonder how many users wouldn't actually fill this in correctly.
aspdotnetdev wrote:
"Accept" is an ambiguous term for answers
Good point! Suggestions for alternative names?
Thiru Thirunavukarasu wrote:
Often the user is unqualified to make a judgement
Then maybe add another button: "This Answer Is Unclear". Or a more general: "I Can't Decide Right Now". They can always select a different button later. And forcing them to choose one option will get more people to choose the correct option (although may have the unintended effect of having more people choose the wrong option). I think part of the problem is that people have no reason to indicate what they thought of the answer... they get their answer and ignore Code Project until they need something again. My suggestions forces them to be a little more interactive (i.e., actually open Code Project in their browser, rather than just reading answers in their email). And it makes the accept/reject feature immediately obvious. Otherwise, they might not even notice it or take the time to think of what it means.
Thiru Thirunavukarasu wrote:
I wonder how many users wouldn't actually fill this in correctly.
Well, I'd say you need to think who you are trying to please. There are basically three groups of users. Those who ask questions (questioners), those who post answers (answerers), and those who search for their problems to find the corresponding answers (searchers). It seems you will please questioners by giving them their answers. You will please the answerers by forcing the questioners to provide feedback. And you will please the searchers with the same feedback provided by the questioners. You might also want to consider that there are mostly a few answerers that do most of the answering. Most of them, I'm guessing, would be happy to improve the QA system, so I'm sure they'd fill in titles correctly. If they don't, then their answer is probably rubbish too, in which case other answerers can assume that and provide their own answer with a better title. I think by shifting the responsibility from the questioners to the answerers, it will be in more capable hands, and the QA system will benefit from it.
Thiru Thirunavukarasu wrote:
Suggestions for alternative names
"I Like This Answer", "This is a Good Answer", "I Need A Better Answer", "I Want Another Answer", "This Answer is Useful", "This Answer Isn't Useful".
-
For Quick Answers we've implemented an 'Accept Answer' feature. Perhaps some of you aren't aware of this and if so that's part of the problem. Let me briefly explain what it's all about. After a member posts a question and receives at least one answer he/she has the ability to click 'Accept Answer' on any of the answers. Our thought is that this would be useful for answer seekers who will be able to more easily identify the best answer(s). It may be the case that the author of the question is best able to judge what is the best answer for his/her question. But I'd have to admit that this is also often not the case, in practice. Accepted Answers as its been implemented and how it's being used just isn't working out well at the moment. Here are my main concerns: 1) Does provide value to the site? 2) If not, should we simply remove this feature and just leverage the best voted/rated answers by the community? Maybe this is sufficient or even more accurate? 3) Can we salvage the feature but improve it somehow? 4) Other thoughts?
You might implement a points system (which would be included in the Reputation calculation, like the old TechRepublic site used. A querent asks a question, views answers, and selects the (possibly more than one) as acceptable. The answer providers gain points for good answers, and question suppliers receive points for accepting answers. Those who would answer can view the querent's points and see whether he/she is the sort who accepts answers, then decide whether to answer or not. Those who ask questions but never accept answers will quickly discover that no one is answering. :)
"A Journey of a Thousand Rest Stops Begins with a Single Movement"
-
Thiru Thirunavukarasu wrote:
Often the user is unqualified to make a judgement
Then maybe add another button: "This Answer Is Unclear". Or a more general: "I Can't Decide Right Now". They can always select a different button later. And forcing them to choose one option will get more people to choose the correct option (although may have the unintended effect of having more people choose the wrong option). I think part of the problem is that people have no reason to indicate what they thought of the answer... they get their answer and ignore Code Project until they need something again. My suggestions forces them to be a little more interactive (i.e., actually open Code Project in their browser, rather than just reading answers in their email). And it makes the accept/reject feature immediately obvious. Otherwise, they might not even notice it or take the time to think of what it means.
Thiru Thirunavukarasu wrote:
I wonder how many users wouldn't actually fill this in correctly.
Well, I'd say you need to think who you are trying to please. There are basically three groups of users. Those who ask questions (questioners), those who post answers (answerers), and those who search for their problems to find the corresponding answers (searchers). It seems you will please questioners by giving them their answers. You will please the answerers by forcing the questioners to provide feedback. And you will please the searchers with the same feedback provided by the questioners. You might also want to consider that there are mostly a few answerers that do most of the answering. Most of them, I'm guessing, would be happy to improve the QA system, so I'm sure they'd fill in titles correctly. If they don't, then their answer is probably rubbish too, in which case other answerers can assume that and provide their own answer with a better title. I think by shifting the responsibility from the questioners to the answerers, it will be in more capable hands, and the QA system will benefit from it.
Thiru Thirunavukarasu wrote:
Suggestions for alternative names
"I Like This Answer", "This is a Good Answer", "I Need A Better Answer", "I Want Another Answer", "This Answer is Useful", "This Answer Isn't Useful".
aspdotnetdev wrote:
And forcing them to choose one option will get more people to choose the correct option (although may have the unintended effect of having more people choose the wrong option)
True. I'm not sure if this would improve things or not. It could go either way. I think for something like this we really need to put it out there and see the results. The question is whether this is a worthwhile risk. But I will talk this over with the team and get back to you on it.
aspdotnetdev wrote:
Most of them, I'm guessing, would be happy to improve the QA system, so I'm sure they'd fill in titles correctly. If they don't, then their answer is probably rubbish too, in which case other answerers can assume that and provide their own answer with a better title
I'm not so sure about this. I think in many cases there isn't a good title for an answer. Maybe the answer is quite involved and highly technical (with code). Or maybe the answer addresses a number of points and provides alternatives. I think this feature would add more friction that the advantages it may provide. But I will bring it up with the team - they may think otherwise.
aspdotnetdev wrote:
I think by shifting the responsibility from the questioners to the answerers, it will be in more capable hands, and the QA system will benefit from it.
Good point! Thanks for the alternate names. I like "Useful Answer" because it's concise and clear though not accurate enough. Maybe we can also bring up a little bubble or tooltip with more details on hoverover.
-
You might implement a points system (which would be included in the Reputation calculation, like the old TechRepublic site used. A querent asks a question, views answers, and selects the (possibly more than one) as acceptable. The answer providers gain points for good answers, and question suppliers receive points for accepting answers. Those who would answer can view the querent's points and see whether he/she is the sort who accepts answers, then decide whether to answer or not. Those who ask questions but never accept answers will quickly discover that no one is answering. :)
"A Journey of a Thousand Rest Stops Begins with a Single Movement"
Roger Wright wrote:
A querent asks a question, views answers, and selects the (possibly more than one) as acceptable
This is the problem. People simply aren't choosing to accept answers.
Roger Wright wrote:
question suppliers receive points for accepting answers
We considered this but thought users would game the system and just accept answers to get points. But maybe it wouldn't play out that way..
Roger Wright wrote:
Those who would answer can view the querent's points and see whether he/she is the sort who accepts answers, then decide whether to answer or not. Those who ask questions but never accept answers will quickly discover that no one is answering.
Good point! If we do implement the answer acceptance rate and make it very visible it could help things considerably. Thanks for your thoughts :)
-
For Quick Answers we've implemented an 'Accept Answer' feature. Perhaps some of you aren't aware of this and if so that's part of the problem. Let me briefly explain what it's all about. After a member posts a question and receives at least one answer he/she has the ability to click 'Accept Answer' on any of the answers. Our thought is that this would be useful for answer seekers who will be able to more easily identify the best answer(s). It may be the case that the author of the question is best able to judge what is the best answer for his/her question. But I'd have to admit that this is also often not the case, in practice. Accepted Answers as its been implemented and how it's being used just isn't working out well at the moment. Here are my main concerns: 1) Does provide value to the site? 2) If not, should we simply remove this feature and just leverage the best voted/rated answers by the community? Maybe this is sufficient or even more accurate? 3) Can we salvage the feature but improve it somehow? 4) Other thoughts?
When using NNTP and posting my questions to the various MS topics on msnews.microsoft.com, I would have to place a [RESOLVED] into the subject after someone gave me an answer that I was looking for that resolved my problem. This way no one needs to read any further if they don't want to (to save their time). Could this single push button system of yours do the same thing? If an answer submitted to my question resolves my issues, could pressing a button automatically alert the rest of the community browsing the questions be told that the question has been resolved so there's no further need (other than curiosity perhaps) to view the thread? I know this isn't the same thing as you are proposing but perhaps it's something to think about? Maybe a check box to go along with your best answer button to show that the answer not only was the best answer, but also completely resolved the problem (assuming the question is about a problem that needs resolved).
-
When using NNTP and posting my questions to the various MS topics on msnews.microsoft.com, I would have to place a [RESOLVED] into the subject after someone gave me an answer that I was looking for that resolved my problem. This way no one needs to read any further if they don't want to (to save their time). Could this single push button system of yours do the same thing? If an answer submitted to my question resolves my issues, could pressing a button automatically alert the rest of the community browsing the questions be told that the question has been resolved so there's no further need (other than curiosity perhaps) to view the thread? I know this isn't the same thing as you are proposing but perhaps it's something to think about? Maybe a check box to go along with your best answer button to show that the answer not only was the best answer, but also completely resolved the problem (assuming the question is about a problem that needs resolved).
ooh resolved - I like that!
jim norcal wrote:
alert the rest of the community
We're working on a subscription mechanism that should address this.
jim norcal wrote:
so there's no further need (other than curiosity perhaps) to view the thread
It may still be worthwhile to look through the thread even if there is an answer that resolves the question. Sometimes there will be equally valuable alternatives. This depends largely on the type of question, of course.
