A recent meta question seems to me to be asking the wrong question. However, there is some truth in the underlying issue raised:

The majority of golfers browse the latest and featured questions.

Or in the word's of ais523's answer:

Newly posted challenges also have major advantages in terms of participation and voting, compared to old challenges... So if you assume that people's incentive to post to the site is at least partially based on either a) views on their answers, or b) reputation, the incentives to post on a new rather than old challenge are very large.


How can we encourage people both to answer old questions and to look at and vote on new answers to them?

Previous ideas

  • \$\begingroup\$ Is this about old questions which are answered or unanswered, because what's wrong with bumping an old question every once in a while by making a minor edit? \$\endgroup\$
    – Beta Decay
    Commented May 20, 2017 at 9:58
  • \$\begingroup\$ @BetaDecay, either, but will bumping drive views? \$\endgroup\$ Commented May 20, 2017 at 10:00
  • 1
    \$\begingroup\$ Recently, when I've bumped an old question I've found that you get a couple more upvotes and a couple more answers and therefore more views. \$\endgroup\$
    – Beta Decay
    Commented May 20, 2017 at 10:03
  • \$\begingroup\$ @BetaDecay not when you bump older, more challenging questions :/ I was bored once, browsed unanswered by votes, picked one that had a lot of upvotes and looked hard, answered it, and that's all that happened \$\endgroup\$
    – Stephen
    Commented May 20, 2017 at 16:27

5 Answers 5


Run a monthly contest

We have an annual contest which attracts quite a bit of attention to the winners inherently, and a bit more because it uses bounties for prizes. It's too process-heavy to run more frequently, but a process-light contest would be possible.

The rules I envisage are:

  • Each contest runs during one calendar month, UTC.
  • An answer is eligible if it was posted during that calendar month on a question which was at least three calendar months old at time of posting.
  • Answers can be self-nominated, or nominated by others.
  • The judge may at his discretion review the New answers to old questions list to supplement the nominations.
  • The judge's decision is final, and may reflect his biases.
  • Notwithstanding the judge's personal biases, attempts to corrupt the process via bribery or blackmail will be strongly frowned upon.
  • Prizes will be awarded as bounties. The envisaged prize structure is one prize per month, awarded by the judge; other members are welcome to form their own judgements of the nominated answers and award their own prizes.

In order not to clutter meta, I propose running it in a chatroom.

  • 2
    \$\begingroup\$ NB I had this idea in mind when I posted the question, but I didn't want to post it straight away in case it shut down other answers. I'm posting it for feedback rather than just doing it, Wiki-style, because while everyone has the right to use their rep for bounties with few constraints, I don't want to do this if it isn't going to go down well with the community. \$\endgroup\$ Commented May 23, 2017 at 8:06
  • \$\begingroup\$ I like this idea. Who is the judge(s)? I'm imagining several judges each month that indicate that they are willing to put forth the rep, and they each get to choose one? Otherwise you have to have some sort of nomination process to select a single judge. \$\endgroup\$ Commented May 23, 2017 at 20:38
  • \$\begingroup\$ @Nathan, I'm offering to judge and provide prizes but, as per the last point, anyone else who wants to and has enough rep to offer a bounty could also judge and award their own prize. \$\endgroup\$ Commented May 23, 2017 at 21:32
  • \$\begingroup\$ OK I will not bribe or blackmail then. :) \$\endgroup\$ Commented May 29, 2017 at 7:43
  • \$\begingroup\$ Is meta clutter an issue? Seems meta only has a handful of questions a week. \$\endgroup\$
    – Tas
    Commented May 31, 2017 at 4:08

Clarify how to handle changing standards

Visibility and incentives aside, answering old questions is awkward because of changing site standards. I've given up on submissions I started to write because I didn't know what I was allowed to do.

Can I submit a function as allowed by default now, but not when the challenge was posted? Am I expected to figure out what was standard then? If an existing answer spends bytes to process input like l=map(int,raw_input().split()), do I have to do the same? I worry that if I take advantage of looser modern standards, I look like I'm winning unfairly by playing by different rules.

Moreover, most challenges from 2014 and earlier would be closed as unclear if posted today. So, I don't even know if I've met the spec, and if I ask in a comment, I'll probably never get a response. Should I just make a judgment call and edit that in? Is that rude, or maybe biased? What if some submissions interpret it differently? What if there's some dumb unobservable requirement that half the answers ignore, or a subjective bonus?

I find this uncomfortable enough that I often don't bother answering old questions, even when I've already golfed a solution and come to terms that it won't get much visibility.

Some community guidance on what's OK would make me feel less like I'm treading on thin ice. Or, community projects to clean up old dupe targets so that we can make decisions as a group.

  • 3
    \$\begingroup\$ Earlier when I went to answer our site's fibonacci question I noticed it was terribly unclear by today's standards. I asked in chat if anyone would object to some stronger specifications. Everyone seemed in favor of a stronger spec, so I edited it. I think as long as a edit doesn't invalidate existing answers it should be done, other wise the question should be closed as unclear (or in some cases as too broad). \$\endgroup\$
    – Wheat Wizard Mod
    Commented May 20, 2017 at 22:31

Just bumping an old question can help

Every now and then I like to pick a page deep into the challenge list and answer a question from it. Yesterday (a little before this question was asked), I answered this question, which was on the last page sorted by "active" at the time. Since I did that, the question's had a number of other answers (if you look at the dates of the answers, there's an obvious split, with some being posted in 2011 and some being posted in 2017, with nothing in between).

I don't think randomly bumping questions is all that useful – some questions aren't worth bumping, for instance – but looking through old questions for one you like the look of, and answering it, is a good way to increase the supply of interesting questions that people will find.


Vague idea: improve search

Imagine having to answer an old challenge, which has 3 pages of answers. I assume you should never post a duplicate answer (at least I never would).

Before posting your code, you should look through all the answers, filter those that are in your language, and check whether your answer adds anything new.

I'd be happy if we had a "search for answer in a particular language" possibility, but I cannot define what exactly it should do. Moreover, this feature has been needed for a long time now, seeing that the competition across languages is usually meaningless.


Or just delete the old ones.

So our current approach seems to be: every question can only be asked once. And each unique answer can only be given once (usually just a couple of answers per language).

It seems to me a bit like a crossword book at a library: after a while, all the crosswords have been done, and it isn't much fun for newcomers.

Wouldn't it be better to delete the old questions, so that newcomers can discover them and have the fun that the old-timers had? It's much more appealing to answer a new question with 3 answers than an old question with 200 answers.

  • 1
    \$\begingroup\$ Just because the question isn't new doesn't mean it's no longer fun or interesting. One of my favorites I found while browsing was this challenge in which I was over 2 years late to it. Answering old challenges also helps refresh them by bringing them back into the fold of active challenges where others might find interest in them. \$\endgroup\$
    – miles
    Commented May 24, 2017 at 6:11
  • \$\begingroup\$ Of course. But your example has only 4 answers (including yours). Answering an old question with 50 answers isn't so much fun, because your answer will never be seen, and it probably can't beat the other answers in the same language. \$\endgroup\$ Commented May 24, 2017 at 7:35

You must log in to answer this question.

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