0
\$\begingroup\$

Unlike on other Stack Exchange sites, the execution of an idea really makes a difference on this site. My question is, if an idea for a challenge has been used before, does that make any other challenge based on the same idea off-topic, or can an idea be re-used if it's implemented in a substantially different way?

I ask because I recently wrote a challenge based on an idea that turned out to have been used before. I didn't realise this until I pasted it into the site, but I decided to post my question anyway because the execution of the idea is very different, and the kind of answers I hope to attract are very different from the answers to the old question. (The new question is really designed to allow quite long answers that catch a lot of special cases, whereas the old one has a scoring system that very strongly favours short answers.)

There is some (very reasonable) debate in the comments as to whether my question is a duplicate or not. I would like to know what the community thinks, both about this case and about the general question.

My new question: Turn nouns into plurals

The old question: Pluralize it!​​

| |
\$\endgroup\$
  • 1
    \$\begingroup\$ The consensus is: if answers from the old question can be copied over with little modification and remain remotely competitive, then it's a duplicate. \$\endgroup\$ – Martin Ender Oct 14 '14 at 15:49
  • 2
    \$\begingroup\$ There are a few posts here on meta about dupes in general (and some specifics): meta.codegolf.stackexchange.com/… \$\endgroup\$ – Geobits Oct 14 '14 at 15:50
  • \$\begingroup\$ @MartinBüttner that definitely isn't that case here, but the question is still attracting close votes. \$\endgroup\$ – Nathaniel Oct 14 '14 at 16:01
  • \$\begingroup\$ If an answer got zero incorrect in yours, it would likely win both contests, not merely compete. In both cases the scoring is a weighting of code length and correctness. They both highly incentivize complete correctness (zero-score on the old, only counting size on yours with a 50% bonus). Saying yours is designed to attract longer answers to catch edge cases is irrelevant if catching them all would be a perfect score on the other. \$\endgroup\$ – Geobits Oct 14 '14 at 16:15
  • \$\begingroup\$ For what it's worth, it would be pretty easy to make a non-duplicate by just not pluralizing. Make gerunds, conjugate verbs, match rhymes... There is a well of challenges related to English language irregularities waiting to be tapped. I don't see the need for two "pluralize this list of words" challenges. \$\endgroup\$ – Geobits Oct 14 '14 at 16:21
  • 8
    \$\begingroup\$ Had a good laugh about this meta-duplicate. \$\endgroup\$ – TwiNight Oct 14 '14 at 22:09
  • 1
    \$\begingroup\$ @geobits according to the linked questions, the consensus up to now has been that a question is off topic if the existing answers to the old question would be competitive on the new one. You're arguing that my question is off topic because answers to the new challenge would be competitive on the old one. This is different. In the old question there's no way of knowing if you've reached a score off zero, but if you have there's no point in going further. In the new one the score is accessible to everyone, and there's always scope for further optimisation. Thus it should attract more sophistic \$\endgroup\$ – Nathaniel Oct 15 '14 at 0:43
  • \$\begingroup\$ You said yourself that top answers there would fare decently on yours. My comment here was in addition to that, not replacing it (since you knew it already). \$\endgroup\$ – Geobits Oct 15 '14 at 0:55
  • 1
    \$\begingroup\$ @Geobits I'm sorry if I was unclear, but the intent of that comment was the opposite of how you took it. The top answer on the old question would be a reasonable first stab at the new one, but it could be beaten with minimal effort and therefore wouldn't be competitive. \$\endgroup\$ – Nathaniel Oct 15 '14 at 2:30
  • \$\begingroup\$ @Geobits just in case the implications of the rules aren't obvious, the new question is explicitly designed as a kolmogorov-complexity task. It's about creating a good partial compression of the provided data (in source code form) and trading off the length of the compressed version against the accuracy of compression. I was proud of myself for thinking of a challenge that's so easy to attempt but so nontrivial to win. I expect optimal answers to be hundreds or thousands of bytes in length, especially if they claim the bonus. The other question won't generate answers anything like that. \$\endgroup\$ – Nathaniel Oct 15 '14 at 8:02
  • \$\begingroup\$ Gerunds and verb conjugations are not a suitable replacement because there are far fewer irregular ones, hundreds at most - the data just isn't varied enough to be interesting. A challenge based on rhymes would be nice, but it won't work for this scheme because rhymes are not a one-to-one mapping. I could go for another language (maybe German verbs?), but it'd be less interesting for those who don't speak it. \$\endgroup\$ – Nathaniel Oct 15 '14 at 8:12
  • \$\begingroup\$ Since a byte of code and an incorrect word count the same, to do anything but simple regex/replace like the old one, you need to compress at a rate of a single byte per word (leaving a bit of overhead for decoding). I'm not sure that's possible with the word lengths you have included. Maybe if you got the bonus you could just squeak out a win by compressing at two bytes per word, but my guess is that the winner on yours is going to use the same method that the winner on the other did, just like the current answers are doing. Time will tell. \$\endgroup\$ – Geobits Oct 16 '14 at 20:18
  • \$\begingroup\$ @geobits I chose the scoring system and bonus on exactly that basis, that it's just about plausible someone could do it. For most words you only have to encode which of several possible rules apply, so I think a byte per word on average is not implausible. Even just using regexes I think much better scores are possible than we've seen so far. \$\endgroup\$ – Nathaniel Oct 16 '14 at 23:59

Browse other questions tagged .