Stack Exchange, in general, doesn't deal very well with posts that become out of date.
On other sites in the network, this often happens due to changing circumstances in the outside world that have caused some of the answers to no longer apply to the present way of doing things. In this situation, the best you can do is post a competing answer, but because it's often years after the question was asked and it's not getting much more attention, that answer often ends up never rising above the answers that were posted years ago and were correct at the time, but now are misleading.
On PPCG, challenges get outdated too. This happens in two ways:
- We're getting better over time at identifying ways to write competitions so that they can be fair and fun for everyone involved. This is why we rarely have language-specific challenges nowadays without a good reason for example. Older challenges can often be written in a way that "locks" the challenge into a form that isn't ideal, such as requiring an inflexible I/O format, screwing up the victory condition, or the like.
- New languages and language features are written all the time (not to mention bugs in existing languages being fixed), and although challenges have been known to inspire golfing language features (in the hope that they'll be useful on future similar challenges), most of the time new features that would be useful for solving a challenge weren't written with the intent of cheating/loopholing the challenge. As such, the "languages that postdate the challenge are noncompeting" rule is somewhat unfortunate; it works to close a loophole, but can have a somewhat negative effect on old challenges as a side effect.
Note that there's a notable difference from other Stack Exchange sites: on PPCG, it's the question that becomes problematic, rather than the existing answers.
Now, if the old challenge is problematic enough to outright close, you're in luck; you can just close it, then submit a duplicate. (We explicitly allow near-duplicates of offtopic questions, if the duplicate is ontopic.)
Newly posted challenges also have major advantages in terms of participation and voting, compared to old challenges; they get advertised in chat, and are fairly easy to get onto Hot Network Questions if you want to (or often, even if you don't), thus advertising them to everyone who uses any Stack Exchange site (including the massive Stack Overflow). I'd typically expect around 8 upvotes for a good early answer to a freshly asked question (there's no hard statistics behind this, just my own sense of where my posts typically end up). By comparison, if a question gets bountied, that typically attracts 2-3 upvotes to my answers on it, and if I just post an answer to an old question that I found via search or related questions (without anything promoting it), that's normally worth 1 or even 0 upvotes. 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. (At one time, while I was unemployed, I decided that I wanted to hit reputation cap every day for a week. It wasn't particularly hard, but it required really heavily monitoring the "new questions" list and trying to get in early.)
So I don't think it's entirely unreasonable to want to "restart" a good old question in order to give it the advantages that new questions have; people will be more inclined to answer it, and also the question itself will naturally be better due to allowing more languages (and better still if you fix wording problems and the like in the process). I'd say it's only worth doing so, though, if the question is good, and the new "duplicate" will be even better. Additionally, we don't really want a situation in which people just copy their own (or worse, other people's) answers from the previous challenge, because that's not really interesting at all. These considerations mean that the duplicate will probably have to be posted intentionally; you want to significantly improve on the challenge, and without awareness of the old challenge (and its answers, if any), you're unlikely to make the new one as good as can be.
So although I'd be fairly encouraging of potential new rules allowing challenge reposts, I'd suggest that we conform at least to the following restrictions in any process for doing so:
- The answers on the new challenge should be expected to be mostly distinct from the answers on the old challenge (either due to newer languages, an improved victory condition, or just general improvement to people's golfing skills since the old challenge was posted).
- The new challenge should be comprehensively reworked to fix infelicities in the specification, any challenge-specific loopholes that were discovered, and similar issues. Changing details of the challenge should be encouraged whenever it makes it better (and as a bonus, may end up making the new answers more different from the old ones, too).
On a related question, which is basically your question limited to answer-chaining (which has its own specific concerns), there was support for allowing reposts with the permission of the OP and after a six-month waiting period. For questions more generally, six months seems too short; I'd expect a more appropriate timespan to be two years, or even more (long enough that the original OP may be hard to contact; the linked thread suggests waiting for a month for a response, which makes sense). This inspires a third suggestion:
- If the old challenge is not closed, it should be at least two years old, and the reposter should have permission from the challenge's asker (or else have had no response for a month after asking for permission).