26

Premise

There are over 40k tumbleweed badge-giving questions now on SO (41,959 on last check), out of 186,432 unanswered questions. SEDE (although a bit dated) lists 974,000 questions asked, so that's almost 20% of questions unanswered, and 4% doing worse than unanswered: they got ignored. The nature of the tumbleweed badge being that it is awarded once, there are potentially more than 4% of questions that go ignored.

Maybe we're growing beyond what the system as designed can handle. Lots of these tumbleweeds are from 1 rep users, and I'm betting that's turning off a lot of potential answers. Still, there's a growing (is it also growing proportion-wise?) pool of unanswered questions that are getting no attention.

Proposal

To up the value of answering tumbleweed (or tumbleweed-worthy) questions. When a question has the criteria for the tumbleweed badge (low views, 0 votes, 0 answers):

  • Upvote on answer give more rep (15 instead of 10), until there's an accepted answer
  • Accepted answer gets more rep (30 instead of 15)

Closing arguments

Some suggestions have already been made in the past (for example auto-rep bounty) which would arguably be gamed for maximum profit, which would have all sorts of perverse effects. Likewise if we had an answer reward proportional to question upvotes. These are out of the question. But to force a question to tumbleweed so it can be worth more to answer it? I don't know how that would get done. Just looking at the question to see if it's content is tumbleweed-worthy is bringing it closer to not getting tumbleweed.

I believe we can up the incentive on the forgotten questions and produce a lot more satisfaction for users everywhere. (And none of my questions ever tumbleweeded, so I'm not doing this for me :P)

Improve this question
8
  • You make a good case, but it would be important to know how many questions are tumbleweed and abandonded asked by hit-and-run users. It is possible, even likely, that a large percentage of questions is tumbleweed because they deserve to be, i.e. are objectively bad questions.
    – Pekka
    Commented Oct 31, 2010 at 17:08
  • 4
    @Pekka: In my mind that's where the "Not a question" reason for closing is for, questions that are unanswerable. Weed those out, and you're left with potentially good questions that just fell under the radar.
    – MPelletier
    Commented Oct 31, 2010 at 17:36
  • 2
    Nitpick: 41,959 users have been awarded the badge -- a single user asking multiple eligible questions would only receive the badge once. The difference is likely negligible, but just putting that out there for accuracy.
    – Jon Seigel
    Commented Oct 31, 2010 at 19:06
  • @JonSeigel: Very astute, sir. I did not know a tumbleweed badge could not be awarded twice or more. In that case my proposal is flawed, it should be about any tumbleweed-worthy question, not just the one that got the user their badge.
    – MPelletier
    Commented Oct 31, 2010 at 19:31
  • Is it possible to raise the barriers to entry to the site?
    – Tim McNamara
    Commented Oct 31, 2010 at 20:01
  • 1
    @TimMcNamara: There's a catch there. We want to remain open, to be un-elitist. There's a whole stigma that comes with 1-rep users that frequent users know, yet newcomers don't. You either wise-up and realize that you need to contribute, or you leave. That means that yes, there might be a need for 1-rep question weeding, but that's a whole 'nother game.
    – MPelletier
    Commented Oct 31, 2010 at 20:28
  • 1
    They say the Queen Mary takes a whole kilometre to come to a full stop. Over the years my experience has taught be that there are a lot of fly-by questions, and nobody should be sorry about them. I'm more in the favour of an Atwoodian "ask correctly or don't get an answer" attitude now.
    – MPelletier
    Commented Sep 29, 2012 at 2:28
  • @Pëkka You make a good case, but it would be important to know how many questions are tumbleweed and abandonded asked by hit-and-run users. As far as I know, SE (and SO) strives to provide a great database of answers not just for the asker, but also for the searchers from Google. Just because a question is abandoned by the OP doesn't mean the question is unworthy of an anwer.
    – Unitato says Reinstate Monica
    Commented Jul 10, 2017 at 8:02

5 Answers 5

Reset to default
5

I actually favor deletion, if sufficient time has elapsed.

How big a problem are unanswered questions?

we could probably ignore the question score and just delete anything older than 30 days that has no answers posted. This is currently about 7,000 questions, vs ~5000 with score <= 0. I haven't figured out yet how to inspect the vote scores of answers.

To me this indicates there's something fundamentally wrong with the question.

Improve this answer
15
  • 14
    Or that it is an extraordinarily difficult question to answer.
    – waiwai933
    Commented Oct 31, 2010 at 17:20
  • 2
    @waiwai but those are supposed to get at least some upvotes, at least in theory. +1, deletion sounds like a good idea
    – Pekka
    Commented Oct 31, 2010 at 17:22
  • 8
    Except, in practice, they often aren't upvoted. I don't mind too much if the question has a negative score, but 0 is a different case. @Pekka
    – waiwai933
    Commented Oct 31, 2010 at 17:23
  • 3
    @waiwai all the more reason to upvote well-put but tough questions.
    – Pekka
    Commented Oct 31, 2010 at 17:23
  • 5
    Or that it was posted at a low activity moment (say Friday afternoon) and fell off everybody's radar.
    – MPelletier
    Commented Oct 31, 2010 at 17:34
  • 6
    @Jeff - sorry, disagree. I had a Tumbleweed question that got answered and another that wasn't, mainly because the subject matter (MediaWiki) is less well know than, e.g. C++. But when they get answered, even after a long time, I check it out and am usually pleased with the answer. So please don't close my questions!
    – Reinstate Monica - Goodbye SE
    Commented Oct 31, 2010 at 17:46
  • @mark we'd delete them, not close them. I am thinking after 120 days.
    – Jeff Atwood
    Commented Oct 31, 2010 at 17:49
  • 5
    @Jeff - sorry, I meant delete, indeed. Even after 120 days, why? Let people choose to close, not algorithms which can only guess if a question is "fundamentally wrong".
    – Reinstate Monica - Goodbye SE
    Commented Oct 31, 2010 at 17:52
  • 1
    Why IS there a tumbleweed badge to begin with, if not to bring attention (or provide a "balm") to the poor chap who's questions went ignored, and by it bring attention to the question? If this is the case, then the badge fails, because (through sheer number?) attention is not given.
    – MPelletier
    Commented Oct 31, 2010 at 17:55
  • 5
    So not only stackoverflow "failed" the user by not having their question answered, but now the poor user's question is even deleted? The proposal in the OP makes much more sense since it helps getting those questions actually answered.
    – Andreas Bonini
    Commented Oct 31, 2010 at 19:16
  • 2
    There may be room for both proposals, as there are a lot of unanswered questions. The majority are garbage (duplicates, horribly written, lacking info, etc) and we'd be better off without them. Even if we lost a small number of decent questions it would be worth it to flush out the junk. As I said, the good ones can be re-asked. But once the garbage is cleared out, upping the reward on what's left seems reasonable. Perhaps "good" could be defined as something like score >= 3 and/or asker's rep >= 100?
    – Brad Mace
    Commented Oct 31, 2010 at 19:52
  • 3
    Also, those of you that are appauled by losing these 'valuable' questions should look at some of the odata queries on my proposal. The instances of great questions that get no votes, no answers, and no followup from the asker are vanishingly rare.
    – Brad Mace
    Commented Oct 31, 2010 at 19:54
  • 1
    @bemace: It's not about not losing great questions. It's about getting an answer for those that really went unnoticed. Weed out the 1-rep questions and the duplicates, and you've got a good-sized pool of unanswered yet potentially answerable questions. Those are the ones I want to put attention on. They might not be the question of the century, but dammit someone wants (and deserves) an answer, and odds are it's not give-me-teh-codez types.
    – MPelletier
    Commented Oct 31, 2010 at 20:10
  • @bemace: I read your posts in reverse. I agree with your proposal.
    – MPelletier
    Commented Oct 31, 2010 at 20:14
  • @Pëkka, Regarding "but those are supposed to get at least some upvotes, at least in theory", In theory, seriously hard questions wouldn't get upvoted, because people write them off without even recognizing them.
    – Pacerier
    Commented Apr 30, 2015 at 9:15
8

Another aspect to this which I don't see mentioned is that it's less intimidating for new users to answer a Tumbleweed question. My first answer was to just such a question, it had sat there for 9 months and I had a good answer for it - so I took the plunge and started contributing to SO.

There's no way I would have dived into the Top Questions list, that old question just seemed much safer. Lo and behold someone upvoted it some time later.

So maybe the initial reward for answering a Tumbleweed question could be increased even further if it's answered by a new poster who subsequently earns upvotes on it.

And there are plenty of well written old Tumbleweed questions blowing about out there in the scrub that rank well on Google, I find them quite regularly.

Improve this answer
3
  • That's an excellent point - it's also exactly how I got started.
    – SteB
    Commented Sep 29, 2012 at 15:13
  • @SteB - And the irony of us both posting our first Meta answer on a very old question is not lost on me either! ;-)
    – McNab
    Commented Sep 29, 2012 at 15:28
  • We'll see if it's lost on the "higher-ups" :)
    – SteB
    Commented Sep 29, 2012 at 15:31
8

I like this proposal. Poor questions (negative votes) should be flagged for attention and deleted.

However, I've seen questions with no answers (2 - 4 years old) that were perfectly valid questions. The question probably got lost or viewers couldn't / didn't want to answer it.

As new users register, they may have more knowledge, experience or interest in these questions, while the original poster may never see the answers, other users could find the question and answers useful, increasing the overall quality of the site.

Since a Q&A site is about providing answers, a small rep increase over the usual could provide an incentive for people to provide answers to "deserted" questions.

Deleting bad questions (with down-votes) is one thing, but just because a question has no answers (yet) doesn't make it useless, especially since it could always get an answer in the future. Most 0 vote questions seem to be very technical and/or quite niche.

To expand on the proposal:

Perhaps there should be a no-rep "interest" icon on questions. This would work very similar to favourites, but would allow even 1-rep users to flag and not confer any rep on anyone. But, when someone does provide an answer, after X days of the question having no answers, the rep they get from up votes and accepted would be multiplied by a factor in proportion to the number of "interested" flags.

Perhaps it could also be made easier to find such questions as well, pls point me in the right direction if there already is. My 2c, unless this has already been acted on/resolved?

UPDATE:
I've seen quite a few "tumbleweed" questions that actually did have answers, but they were hidden in the comments. A bit of work (sometimes quite a bit) can provide good answers to these questions rather than just deleting them.

Improve this answer
1

I partially agree with you. I would hesitate to offer more rep for an upvote on a tumbleweed question since that really requires very little effort and will appeal to rep whores who care little about the question itself.

On the other hand, I do think that extra rep for answering tumbleweeds is a great idea as that would reduce the backlog while rewarding those who make the effort to make that happen.

Improve this answer
1
  • 1
    If it was rep whores giving good answers I'd say the added bonus was doing it's job as intended. They could get a new 'Tumbleweed Whore' badge too :)
    – McNab
    Commented Sep 29, 2012 at 15:46
-1

I think it would be good to automatically delete questions where there is no benefit to future users. For example questions where

  • The question has no answers. (Without answers it is of negligible use to anyone)
  • The question has no or very low up votes. (There is little or no interest in an answer)
  • The OP has a rep of 1 and has not been seen in 60 days. (OP is not coming back to look for answers)
  • The question has had no views in 30 days. (It us unlikely to get any more attention. This also means the up votes are old)

The numbers above are WAGs and should be adjusted by wiser people than me.

How many dead questions could criteria like this eliminate?

Improve this answer
2
  • 1
    If a question had no answers it is likely quite difficult or quite niche, but still very useful to people if only there was an answer. Up-voting of questions requires rep, there could be considerable interest by people with not enough rep - early on I was favouriting questions so I could up-vote when I gained more rep. Niche questions can go many days (and longer) without getting views, sometimes the tags and/or title can put people off, not a reason to delete in my opinion. OP has 1 rep & hasn't visited the site in X days - good point- not sure what can be done about this.
    – SteB
    Commented Sep 29, 2012 at 6:43
  • 1
    If a question has no or very low upvotes, that could also be du to it being quite niche. Therefore, not a good reason to delete it just because it tumbleweeded.
    – RobH
    Commented Nov 30, 2012 at 17:20

You must log in to answer this question.

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