1
$\begingroup$

I have a current bounty on this question of Noam Nisan. Unfortuantely, this question appears to be the Bermuda Triangle for Bounties, as mine is the third bounty being offered which seems likely to expire with no answers produced.

So -- I would be glad if someone would leave a placeholder answer so I could award the bounty, and then the answerer would turn around and offer a bounty of the same size (100 rep) to a question of their choice. Please consider doing that.

$\endgroup$
3
$\begingroup$

The bounty system in the Stack Exchange network works in a little strange way, but I consider your proposal as a slight abuse of the system.

I do not think that a question should stay on top for free without making progress even if it is very interesting. If someone puts a bounty on a question and the question does not get answered, the strategy of putting bounties should be reconsidered.

$\endgroup$
1
  • $\begingroup$ sigh. ok, point taken, everyone. $\endgroup$
    – Aaron Sterling
    Nov 14 '11 at 2:10
5
$\begingroup$

I agree with Tsuyoshi that Aaron's suggestion sounds like an abuse of the system.

If the bounty system fails to encourage people to answer the question after several tries then it could be that we just don't have the right expertise on the site to answer the question. In that case we should consider asking simpler stepping-stone questions, or trying to attract someone that might be able to answer the question as is via twitter/G+/blogs.

$\endgroup$
4
$\begingroup$

I concur with the above answers. The reason the bounty is failing is that either no one can answer the question, or no one cares enough :). In my opinion, the question needs some editing to make the arguments clear - right now, it's a little stream-of-consciousness, and that makes it hard to process.

$\endgroup$
3
$\begingroup$

I think the "bounty system" as such is O.K. (wakes up people). Paul Erdos also offered several hundreds of USD for some innocently looking problems. The difference: Erdos was not afraid to pay in several days ... He hasn't set any deadlines. And answers came only months or years latter. So, why not just to implement a bounty system without deadlines?

$\endgroup$
8
  • $\begingroup$ I am not sure what problem your suggestion is supposed to resolve. $\endgroup$
    – Tsuyoshi Ito
    Nov 14 '11 at 20:47
  • $\begingroup$ @Tsuyoshi: The problem I see it resolving is that currently the reputation points are gone forever after seven days, instead of building up in a bank for the person who eventually solves the "open" problem. So I think Stasys's suggestion makes sense for a research-level site. It certainly annoys me that the only mistake I made was to choose a problem that was "too hard." I should only put bounties on easy-ish problems? $\endgroup$
    – Aaron Sterling
    Nov 15 '11 at 2:20
  • $\begingroup$ @Aaron: Whether you are happy about your bounty or not is up to you, but I do not necessarily see your bounty as a failure. I guess that some users took a fresh look at the problem because of the bounty, and the net effect is that you spent your rep points to attract some users to the problem even if they did not produce the answer. If that was not what you wanted, then I would say that it is your mistake to have put the bounty on the too difficult question. But honestly speaking, I do not like the view which considers the bounty which did not produce an answer as a failure. $\endgroup$
    – Tsuyoshi Ito
    Nov 15 '11 at 2:50
  • 2
    $\begingroup$ I second Aaron: on a research-level site we should allow to state research-level questions. Not only "somebody definitely knows an answer" ones. Of course, this is also not a forum for asking questions which one has hardly thought about for years. This would be not productive. (Unless we create some long-term "research problems" section at stackexchange, with no hope for "quick" answers.) But we should leave a bit more time to find answers even for "easier" questions. $\endgroup$
    – Stasys
    Nov 15 '11 at 18:33
  • $\begingroup$ Now I see the point, but then the proposal should not be about removing the deadline. Instead the proposal should be extending the deadline from one week to, say, one month. I do not think that you understand the role of the deadline of bounties. Without deadlines and with enough number of high-rep users, what would happen? $\endgroup$
    – Tsuyoshi Ito
    Nov 16 '11 at 14:26
  • $\begingroup$ @Tsuyoshi: I understand your concerns. "High reputation" users could then use the "waking attention" system without almost no "risk" by asking "dead questions". (All quoted words are intentionally quoted: Erdos also has offered these dollars not as a deal.) For "hard but not unrealistic" questions it could be perhaps indeed enough just to extend the deadline to, say, at least 3 months? But I am still in favor of opening (perhaps, in parallel) something like "hard but not hopeless questions in TCS" section. Bounted or not. Just for students to look at where the frontiers are. $\endgroup$
    – Stasys
    Nov 16 '11 at 18:14
  • $\begingroup$ Stasys: (1) I think that without deadlines and with enough high-rep users, almost all old questions will be attached some bounties and as a result, bounties will have less impact (inflation). (2) I do not object to a place for “hard but not hopeless questions in TCS,” but that is a little separate issue from the current post about bounties, as you know. $\endgroup$
    – Tsuyoshi Ito
    Nov 16 '11 at 18:25
  • $\begingroup$ Tsuyoshi, yes I understand that StackExchange is actually a "stock exchange": quick flow of ideas (instead of money). But this site could also turn to a very valuable research source (especially for students), if we could find some "stable founds". With hard but not "dead" problems. I.e. an "active" part and an "archived" part. Of course, all this has little to do with Aaron's actual question. But the question--what is (or could be) StackExchange good for--is crucial. $\endgroup$
    – Stasys
    Nov 16 '11 at 18:58

You must log in to answer this question.

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