We've got a fairly old challenge about the Kolakoski sequence. Unfortunately, the challenge is fairly restrictive in its I/O requirements: in particular, input has to be taken via command-line argument. This prevents most esolangs from competing entirely, which I think is a shame for such a popular (and fun) integer sequence.

Should we post a new version of this challenge, which allows all of our standard I/O methods to allow all languages to participate? While we're at it, we could also relax the output requirements to the recent trend of letting the answerers choose between a) given n output a(n), b) given n output the first n terms and c) print the sequence indefinitely.

The old challenge could either be closed as a duplicate of the new one, or it could be merged into the new one as all the old answers would still be valid. However, they'd likely all be suboptimal (except in languages where reading from command-line arguments is shortest anyway). Since there are only 11 answers, I don't think it's a big deal either way.

(Of course, we could just edit the old challenge, but the OP doesn't seem to want to make this change, specifically because it would make the existing answers unnecessarily suboptimal.)

  • \$\begingroup\$ IMO when discussing a specific question on meta it's good etiquette to post a comment on the question which notifies its OP (and anyone else who happens to read the question and its comment chain). This goes extra when the meta discussion suggests that doing violence to the question is on the table (and I think that merging it into a new one qualifies). \$\endgroup\$ Commented Feb 21, 2018 at 17:06
  • \$\begingroup\$ @PeterTaylor Good point, thanks for notifying the OP. I'll try to remember it myself next time. \$\endgroup\$ Commented Feb 21, 2018 at 17:07

3 Answers 3


Yes, let's repost

The restrictive I/O format has been pointed out in the comments and the OP has responded.

That's a cumbersome I/O format that makes many languages unable to participate. – Mego♦ Apr 2 '16 at 5:26

@Mego thanks for the tip. but in fairness to the current submissions, let's not go changing preferential specs on a 4 year old question – ardnew Apr 2 '16 at 20:01

While reposting the challenge affects the OP negatively, if they do not want to edit it, I think it's preferable to respect their wishes and post a new challenge. While relaxing the input method alone might be a justifiable edit, changing everything from allowing functions to allowing infinite output is a rather drastic change that shouldn't happen without the consent of the OP.

  • \$\begingroup\$ That was almost 2 years ago. I'm hoping OP will pop into this discussion and say whether their opinion remains unchanged. \$\endgroup\$ Commented Feb 21, 2018 at 20:51
  • \$\begingroup\$ I've sandbox the new challenge, but I'll give the OP of the original challenge another week to respond to this meta discussion. \$\endgroup\$ Commented Feb 28, 2018 at 12:16

The old challenge could either be closed as a duplicate of the new one, or it could be merged into the new one as all the old answers would still be valid. However, they'd likely all be suboptimal (except in languages where reading from command-line arguments is shortest anyway).

There is at least one answer which also gave a version with more idiomatic input.

More to the point, although the OP is privileged in some regards (e.g. with respect to notifications), it has always been the case that when we post on a Stack Exchange site we grant a licence to edit the content. There is a sense in which the content belongs to its author(s), but the post belongs to the community. So it is legitimate to weigh the OP's desire to use a specific input format to disadvantage golfing languages against what's best for the community as a whole.

Considering the options:

  • Creating a duplicate is to the detriment of the OP (whose question, being closed, loses relevance and is unlikely to attract many more upvotes) and of the community (the answers from the old question can be reposted with trivial modifications, and explanations of answers like the J one get duplicated; but if answers aren't reposted then people have to look at multiple pages to see what they're competing against).
  • Merging the old question into a new one sorts out the duplication / dilution issue, but is to the detriment of the OP (who keeps rep from votes older than a few months, but definitely won't receive any more).
  • Editing the question also sorts out the duplication / dilution issue; its effect on the OP is more nuanced, because their explicitly stated wishes are being overridden but fresh activity on the question is likely to bring upvotes.

Of these three options, the last seems to be the one which best balances advantages and disadvantages to the OP and the community.

  • \$\begingroup\$ You make some good points. Especially if this answer gets some support that might also help convince the OP that it would be the best option to relax the restrictions on the existing challenge, even if that makes the existing answers suboptimal. (And with only 11 answers it wouldn't be too much of a problem to leave a comment on each to inform them that they can change their I/O format if they like.) \$\endgroup\$ Commented Feb 21, 2018 at 17:06
  • \$\begingroup\$ (note that editing the question will neither bring it to the top of the "newest" stack (only "active") nor cause the NewMainPosts bot to have a message in TNB, and some users may miss the challenge) \$\endgroup\$
    Commented Feb 24, 2018 at 13:33

I sincerely don't mind if you guys decide to close as duplicate in favor of the sandboxed version. Perhaps I was being spiteful with the original requirements :)

Otherwise, feel free to modify the existing question to relax the I/O restrictions. When I rejected that proposition originally, it wasn't my question I was protecting -- it was the existing answers. So you certainly have my consent, although mine isn't the one you should want IMHO.


You must log in to answer this question.

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