# Filter out questions with long-accepted answers?

For some reason, people have been dredging up old questions where a answer was accpeted long ago, and adding some tidbit. The ancient question now pops to the top of the list since it was recently modified, so pushes other questions where it's still possible to make a meaningful contribution off the list (off the first page really, which is basically the same thing).

So, is there some way to filter out questions with answers that were accepted more than maybe 2 days ago? The reason for 2 days is that sometimes OPs jump the gun and accept answers much too early and some good discussion follows anyway.

That brings up another point. I don't think it should be possible to accept a answer until some time, like maybe 24 hours, after the question is asked. Until then, you could easily get another better answer than the best so far. Twice now I saw a unanswered question only a few minutes old, so spent some time writing a detailed answer only to find that while I was typing another answer was posted and accepted. That's a big demotivator.

• You may be talking in part about the effects of me wandering around the system with inadequate experience of how to use it properly :-). If so, constructive guidance is welcomed. For example, I wasn't aware of the pertinence of one or other answer being accepted or not. And am still not fully aware. If this affects rankings then I can see it may be considered relevant. If it affects answer usefulness then the relevance is less obvious. To me it would seem that even if an answer has been accepted it may be valuable to add to what has been said if one has something useful to say. Maybe not :-)
– Russell McMahon Mod
Jul 6 '11 at 4:43
• @Olin - I don't agree. It's not only the OP who may be looking for help, and the search function may/will bring up older questions too. The accepted answer is not always the best. Other views should be welcomed. Nevertheless, I also find it unnecessary that questions top the list again after minor changes to existing answers. But new answers deserve attention. Jul 6 '11 at 8:27
• About accepting very rapidly, I do agree. @stevenvh also made a remark on this recently, (and his was the accepted answer!). The OP probably thinks: I ask a question to get an answer, and if I get one I'm happy. I agree with @Olin that this isn't motivating. Can we use the number of views as an acceptance criterion? In the example it was accepted after only 8 views. Jul 6 '11 at 8:35
• I think Federico is right on this: others may also be interested, and more answers means more angles. I often arrive on older questions by clicking one of the "related" links. So maybe it's not necessary that the top the list again for new answers to older questions. But new answers should always be possible. Jul 6 '11 at 11:01
• I'm not against new answers, and sometimes there is legitimately new information added, but lately dredged up old questions seem to be drowning out the new ones. I guess there is no easy answer to this. Jul 6 '11 at 11:55
• @Russell - Olin's view here runs contrary to traditional SE use protocols. Your answers are a great addition to the site, and all the ones I've looked at add new information or an alternate viewpoint, which is great! Posting a more thorough answer... Jul 6 '11 at 15:43

There are a lot of points to be addressed, so this will be a long one.

TL;DR: Answering old questions is a good thing: the front page is important, and should not be deluged with low-value edits, but quality answers are always welcome. New questions are here. There's a 15 minute delay in place, and I often advocate for 24-48 hours as good practice. It's unlikely to increase to 24 hours, but you could try.

For some reason, people have been dredging up old questions where a answer was accepted long ago, and adding some tidbit.

This is a Good ThingTM. They're adding information to our repository of knowledge. Hopefully, questions here will not be localized in time, so information at any time can be useful to any reader. (Alternatively, they might be updating something that's time sensitive, which is laborious and not dependable, but undeniably useful!) We're not just providing answers for the OP, we're providing answers for anyone who has that question at any time.

The ancient question now pops to the top of the list since it was recently modified

Also a Good ThingTM, we want to be able to review and vote on new content.

[and] so pushes other questions where it's still possible to make a meaningful contribution off the list (off the first page really, which is basically the same thing).

Contributions to old questions are still a Good ThingTM. That said, the issue of pushing stuff off the front page is important. Low-value, high-volume edits should be discouraged, but new answers like those that Russel has been posting (containing information not presented in the other answers or an alternative approach to solving the problem) should be welcomed!

There is not a catch-all definition of a low-value edit. Use your best judgement when trying to determine whether your contribution is valuable enough to push one question off the bottom of the front page. As a few examples, I'd classify a spelling fix in the body of a post as low value, but a spelling fix in the title would affect searchability and be valuable. Adding another item to what's already a long list (like these examples..) would be low value, adding a code sample, schematic screenshot, or quote from documentation where there is none would be high value. Fixing LaTeX markdown when the question or equation is still readable is low value. If you're unsure, it's probably better to make the edit (it's still added value), but be careful or start a discussion on Meta before making a lot of them.

Answering individual questions isn't the only goal here. The vast majority of our content isn't consumed by the original author of the question, or even users of this site, but by visitors from Google! We understand that your time is limited, so we want to avoid making this a support center for individuals, and instead create an easily searchable database for the public.

So, is there some way to filter out questions with answers that were accepted more than maybe 2 days ago? The reason for 2 days is that sometimes OPs jump the gun and accept answers much too early and some good discussion follows anyway.

Yes. I'm shocked that you have been posting as quickly as you do without this utility! Click the "Questions" IC in the banner, and then choose the "Newest" tab, to bring you to this page with only brand new questions. Some of these do have accepted answers, if you want only unaccepted answers, you should search for hasaccepted:0 and choose the "Newest" tab.You can also use the RSS feed to save your F5 key.

That brings up another point. I don't think it should be possible to accept a answer until some time, like maybe 24 hours, after the question is asked. Until then, you could easily get another better answer than the best so far. Twice now I saw a unanswered question only a few minutes old, so spent some time writing a detailed answer only to find that while I was typing another answer was posted and accepted. That's a big demotivator.

This is a legitimate problem, but a little off-topic for this post. On Stack Overflow, traffic is much higher, and answers are faster, so they implemented a 15 minute delay. I'd personally be in favor of a delay of 2 hours or so, and maintain that 24-48 hours is a much more appropriate time period (and advocate for it in comments), but I think you might have a good chance at getting this bumped up if you made a specific feature request for it ("Please increase the accept delay based on traffic patterns, we're not the same as SO").

• What's the definition of a low-value edit? Is there an official statement about this? I've tried to find it but I couldn't. Jul 6 '11 at 16:45
• I agree with Kevin's answer. Still I also understand @Olin's frustration when he feels that his answers to newer questions get drowned in a flood of older ones. About accepting answers I support Federico's idea of a minimum number of views. If the question has at least 1 answer, and, say, 25 views, then at least other people had the chance to answer as well. Jul 6 '11 at 17:02
• I guess there is no simple answer to what appears to be a more complex problem than I originally thought. I did originally view newest versus recent (I think those are the names) questions. That seemed to leave out some things, so I switched to recent and forgot all about that choice. The recent flood of re-opened old questions made me notice the issue. With the occasional old question I probably would never have noticed. Oh well, I'm still getting used to this site. Jul 6 '11 at 22:25
• This issue brings up once again the fundamental difference between optimizing for the respository versus for the exchange between those interested in EE. I'm more interested in the latter and not sure how much value there really is in the former, but this site seems to have different priorities. Jul 6 '11 at 22:26
• @OlinLathrop, we are following SO's example. When I have had to do programming and google problems I often get an SO question from 4-6 years ago that perfectly answers my question. Would it not be great if after 5 or 10 years of this site we started making one of the better repositories of knowledge for EE. Slowly as your question base builds up so do our google results. Jul 6 '11 at 23:08
• @Daniel - No, as far as I know there's not an official definition of a low-value edit. You'll have to use your best judgement. I added some examples of what I'd consider low-value edits to the question. Jul 7 '11 at 16:20
• @Kevin, I think I found it. Take a look: blog.stackoverflow.com/2009/04/in-defense-of-editing Jul 13 '11 at 11:02
• @Daniel - Great, thanks! The points I take away from that blog post are (1) fix the whole post while you’re at it, (2) be diplomatic, and (3) when in doubt, move on. That's a little different from my post! Jul 13 '11 at 12:41

I am getting back to one part of the question only (How can we avoid that questions clutter the front page just because of some small edits like fixed spelling). My feature request in that regard would be a checkbox called "This is a minor edit". Many wikis have those right under the editing box and I think they are very useful because checking this box keeps the edit off the standard list of recent changes, which is somewhat equal to the front pages on SE.

Such a feature would not solve the issue completely, and I am sure sooner or later someone will start telling others in a more or less sarcastic way ("Ooh! Ooh! Looky world, I done found me something I can pretend to be thmart about!") to use the checkbox, but the front-page-noise caused by small edits would decrease big time.

I am, like many other users, not a native English speaker and, like many other users, I still want to write as well as I can, which sometimes seems to require a number of small edits.

Also, even if it sounds strange at first, I am convinced that people learn while they answer questions, which makes it necessary to give them the chance to edit and improve their answers, and which is possible and just great here on SE.

That being said, I would like to encourage everyone here to continue improving even old answers, and even if it's only because someone caught a spelling error or a formatting issue. Most information here is timeless, and I regularly get turned off using other Q&A sites that work in a strictly chronological order because

• there is often no way of editing, so the threads become noisy with excuses for spelling errors, posted as yet another answer.

• people are frequently told to stop posting because the OP is two years old, even though the two-year-old information is still perfectly good, and even though I just came here using a search engine and because I am looking for information I need today.

• not being able to edit old stuff often causes people to get in lengthy fights from "answer" to "answer" about who's very right and who's very wrong.

The biggest reason why I enjoy spending time on SE is its wiki nature, i.e. all the details built into it to make an individual person's question useful for an entire community and no matter if it was posted today or ten years ago.

Engineering is all about improvement, and SE is cool because it's all about improving information that first started out as just a little question.

Edit:

In a comment to this, Jeff Atwood states that putting edits to the top of the list again is even meant to be an incentive to improve your own or others' questions or answers. I believe this is good, but it seems to be yet another reason for the checkbox I've mentioned: Those who do the editing can decide if they want the extra top-of-the-list-fame or not. If the box is left unchecked, the default (question or answer is bumped up) will happen as intended.

• I was thinking the same yesterday. I completely agree. Jul 7 '11 at 15:50
• +1 to learning while answering questions. Teaching something to someone else requires you to really understand it, and you'll often realize that you misunderstood something until the point you tried to teach it to someone else. Nov 15 '12 at 15:06

So, is there some way to filter out questions with answers that were accepted more than maybe 2 days ago?

You can filter for unanswered question. You have 2 options:

• You can also sort by newest questions Jul 6 '11 at 16:54
• Thanks @stevenvh and Daniel. I have experimented with unanswered and newest versus active. Unanswered is far too few because most questions don't remain unaswered very long. Newest still seemed somewhat restricted, but that was before the recent flood of old questions. I guess the answer is that I just need to get more used to this happening and learn to use the options better. Jul 6 '11 at 22:29