According to this meta post we can never recreate old challenges. Of course not, its a duplicate. Makes sense to me. No problems.

Except this challenge can no longer be answered with a language newer than the challenge, even though the requirement that "languages be newer than the challenge" is no longer a valid. If I want to be able to solve that specific challenge ("output a-z") with newer languages, I'd have to post a new challenge first. But it's a duplicate.

So, which takes precedence:

  • Challenge spec disallowing new languages
  • No duplicate challenges

Or do we recreate the challenge, remove the date specification, and close the older one as a duplicate of the newer?

Yes, this meta post is in response to my answer (and at least two others) being deleted, however as I was not the first person to answer the challenge recently: I had assumed that the date-invalidity to have been true when I answered and I would not have seen the challenge unless it had been answered by another user first.

  • \$\begingroup\$ I'd say the you can use newer languagese, as the challenge outdates our new rule. \$\endgroup\$
    – flawr
    Commented Oct 2, 2019 at 17:38
  • \$\begingroup\$ @flawr Normally I'd agree, but that challenge explicitly states that new languages are disallowed. (FWIW, I'd like those answers to stick around if possible, but I was the one who deleted them while handling some flags) \$\endgroup\$
    – DJMcMayhem
    Commented Oct 2, 2019 at 17:47
  • \$\begingroup\$ In case you are curious, I have flagged a total of 12 answers to that question, nine for violating the rule on language age (two of those nine were flagged just minutes ago and are still pending), one for not being a serious contender (I don't remember in what way it wasn't a serious contender), and two for straightforward invalidity (producing the wrong output). \$\endgroup\$ Commented Oct 3, 2019 at 3:07
  • 3
    \$\begingroup\$ Its those nine that I feel should not have been deleted, @pppery, for the reasons I mentioned in the comments and in this meta question. If the deletion is the correct course of action regarding those answers, then this meta question is regarding "what should those answerers have done instead?" Currently there is no consensus (meaning, ironically, both that deletion is correct and the answerers weren't wrong to post them). \$\endgroup\$ Commented Oct 3, 2019 at 3:58

3 Answers 3


So, which takes precedence:

  • Challenge spec disallowing new languages
  • No duplicate challenges

"Precedence" is not how it works.

When a particular, unusual, case arises which introduces a tension between two policies, the thing to do is to take the case to meta. This scenario isn't so common that we need to have a general policy about it. Ask a meta question about the , tagged , and set out the options and the arguments. As you can see by looking at those tags, this has happened before, although not often.

(This approach would also give you the opportunity to explain why it's worth reposting such a trivial question as the one you mention).


We have reposted challenges before

Personally, I disagree with that meta ruing. It's 3 years old, multiple challenges have been revived/reused/reposted since then without issue, and we already allow certain types of challenges to be reposted.

I believe that challenges should be ok to be reposted, with a few constraints:

  • Obviously, still active challenges, no matter how old they are, should not be reused. Hello, World! is 4 years old, but still gets multiple answers a month. How long a challenge is "active" for is, IMO, another matter for discussion (although my gut leans towards longer than a year or two).

  • If the challenge specifically prevents you from answering it now rather than when it was posted, it should be allowed to be recreated. The cited challenge (The alphabet in programming languages) has this listed as one of its rules:

    The programming language should have existed prior to the writing of this post, on this eighteenth of April 2011.

    This specifically disallows you from using a new language, rather than being one of the thousands that disallowed newer languages because it was site policy.

    For example, this challenge is from 2013, but there is no rule preventing you from choosing a random language and answering it, so I see no value in reusing it. For challenges saying that the language must already exist (as was site default for many years), I believe this meta ruling overrules that.

  • The challenge author should be fine with the challenge being reposted. Its just respectful to let someone know when their challenge is being redone, as the newer version is likely to take traffic away from the older one. Obviously, this gets a little more difficult if the author is no longer active, but that leads on to my next bullet point.

  • Meta agrees that its worth posting. As you can see from my experience with this, I posted to meta beforehand asking if it was acceptable to the community to do this. I doubt that we'll be wanting to revive multiple challenges a day, or even a week1, so its unlikely that requests like this will clutter the front page, and, given that the community will decide whether the repost is a duplicate, it's good to 'have their permission' (for lack of a better term).

1: And if we are, then that in itself becomes an issue, one which can be addressed if we ever cross that bridge

So, in conclusion, I'm all for reposting certain, inactive challenges, so long as the community sees value in redoing them.

  • 8
    \$\begingroup\$ I would prefer to modify the old challenge before reposting if the only change is to remove the posted-by restriction. \$\endgroup\$ Commented Oct 2, 2019 at 17:45
  • \$\begingroup\$ @NathanMerrill So would I, but rule changes to a challenge should (IMO) be approved by the original author, and that can be an issue with old challenges (which this question is asking about, I believe) \$\endgroup\$ Commented Oct 2, 2019 at 17:48
  • \$\begingroup\$ For challenges saying that the language must already exist (as was site default for many years), I believe this meta ruling overrules that. I cited that meta in one of the comments that I believe I left on another of the deleted answers. I can't verify as I don't have 10k rep, but that did not stop deletion as pppery, Jo King, and DJMcMayhem♦ believe that specifying the rule overrides that meta. \$\endgroup\$ Commented Oct 2, 2019 at 18:04
  • \$\begingroup\$ Comment contents for non-10K: In this case the challenge itself has the rule about languages predating it – Jo King Sep 22 at 23:02 \$\endgroup\$ Commented Oct 2, 2019 at 18:06

EDIT: It seems I am mildly off topic and this particular issue I am addressing was solved two years ago, but I'm leaving this post around for the sake of posterity.

The spirit of the law is important

The primary purpose of the "no new languages" rule is to disallow specialized esolangs that make that particular problem trivial to solve. Perhaps we should revisit this approach in the default ruleset and create a new rule that is less of a low orbit ion cannon and more of a precision laser rifle in terms of what it bans.

Languages don't really compete with each other in general

Java answers, for instance, are going to be much more verbose than APL or Perl. It doesn't really make sense for a solution in C++ to compete with a solution in 05AB1E. Even versions of the same language could make for a big enough difference on golfing to not really be competing; Python 2.7 vs 3.8 are the prime examples of this. It doesn't really make sense for new versions of an old language to be disqualified in old questions.

I think we should discourage questions that make languages compete with each other on byte count.

My proposal

I think the "no new languages" rule in the standard rules and loopholes should be supplanted by something like this:

You may not use a language that was designed to make the exact posted problem trivial. Typically, these are esolangs that were designed after the creation of the challenge and a solution in such a language will often be 0 or 1 bytes long. This does not include multi-purpose golfing languages such as 05AB1E or V, but does include comically overspecialized languages such as H9Q+.


You must log in to answer this question.

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