Wikipedia talk:Stable version to revert to

This page may end up being merged into a policy or guideline but until then its a useful way of dealing with stable versions in the mean time. Crouch, Swale (talk) 18:08, 4 April 2021 (UTC)

RFC

There has been some debate about how long it takes for a change to become stable. There are 2 questions, please indicate which you're answer is for. Things like WP:TITLECHANGES, WP:RETAIN, WP:CFDS and WP:RMCI should be updated if this RFC gains consensus.

(1) How long in general should a version have been in place for it to be considered "stable" for the purpose of obtaining consensus? This isn't intended to be a strict rule but merely give some general rule since sometimes (like with the technical request of Talk:Kereta Api Indonesia#Requested move 8 June 2020) a move made years ago is requested as a revert of an undiscussed move even though the move happened over 2 years ago and in some cases like Omos an unrevert was made that the change was made a month ago even though it had had the same target since 2004 prior to the change.

(2) How long should a version take to be stable for the purpose of WP:RETAIN when a page is edited or renamed against this (and one of the exceptions doesn't apply, such as communality, strong ties or ATDAB)? See examples of Humour and Chinese whispers where this has been debated. @SnowFire, FOARP, Necrothesp, and Calidum: who were involved significantly in the chinese whispers debate on this. Crouch, Swale (talk) 18:08, 4 April 2021 (UTC)

• You pinged a lawyer (me) so I shall a lawyer's answer: It depends. For 1) it really depends how many eyes have been on it and what has been done in the mean time. If literally no-one has even looked at the page since the change was made it is hard to give it much permanence. 2) Same as one. The question is not "how much time" but "what has happened". If there are hundreds of edits and much content added I would say this represents the change settling in. FOARP (talk) 19:03, 4 April 2021 (UTC)
• You're trying to pin down the un-pin-downable. EEng 03:54, 5 April 2021 (UTC)
• I'd agree that the answer is "too variable to even attempt to give a guideline". Anaxial (talk) 05:39, 5 April 2021 (UTC)
• I agree with FOARP and others but would add: if we're talking about an article receiving considerable attention/action since the move in question I'd say three months or more establishes stability. If the article has been essentially ignored since the move no amount of time establishes stability. That said, in RETAIN cases no matter how long a given title is stable if it's an ENGVAR title that is controversial then the original variety reigns, per yogurt. I don't think the move of Chinese whispers should have been moved back after the review. As a result, that title remains controversial. —В²C ☎ 08:41, 5 April 2021 (UTC)
• (Brought here by the RfC robot.) I agree with FOARP on both points and would add that the number of PAGEVIEWS should also be considered, because even if an article has not seen many edits, the more pageviews it has gotten increase the chances that more Wikipedia editors would have read it, and if it was already good quality, might not edit it, but if it was low quality, greater pageviews increase the chance it would get fixed/edited. (for determining ImplicitConsensus) I would guess there might be some unobtainable statistic that x% (maybe below 1%) of all pageviews are Wiki editors. ---Avatar317^(talk) 21:28, 5 April 2021 (UTC)
• Looks like my reply got lost (thought I responded originally), but pretty much see above. This is inherently going to change a lot from case to case. You can maybe set out some principles for what tends to make things more stable and what makes it less stable, but you're not going to find a consistent rule. I would hesitantly put out that RETAIN should be a somewhat firm rule (with the exception of total rewritings of an article, i.e. a two sentence stub doesn't count), and other undiscussed moves will vary by pageviews, number of editors, etc. A move that stood undiscussed for 10 years on a lightly trafficked article could well have less consensus behind it than a move three months ago on a heavily trafficked article. SnowFire (talk) 15:17, 6 April 2021 (UTC)
• Concur with FOARP, EEng, et al. In actual practice, this isn't about lengths of time, it's about onset of dispute, and the stability-to-dispute ratio. If the X that someone wants to change at an article was substantively stable until a fight broke out yesterday over a change boldly made at 14:57 UTC, then the status quo ante to revert to is the state it was in at 14:56. If the fight has been protracted since 2017, and there was periodic squabbling about the same X matter even before that, it's not clear there really is a stable version to revert to, but it might be the version before the onset of the "four-year war" in 2017, or if the bickering has periodically subsided for a few months at a time, it might be the most recent inter-conflict version before current round of bickering started up again 2 months ago. Like much else on Wikipedia, this stuff is a judgement call.

  Stability is a state of equilibrium, and it can come about quite suddenly, e.g. as the result of an RfC, RM, or other fairly formalized discussion, or simply because two parties who had been incessantly bickering finally ceased for a while and wandering off to go do something else, until they picked it up again later and punctuated that equilibrium with some more chaos. Stability can also come about gradually and organically (and mostly should). If we were to create a rule of thumb about it, maybe: the status quo before the dispute, to revert to pending a clearer consensus emerging, is the last version of the X under dispute that was not substantively changed for longer than the current round of dispute has been open. So, if people have been gnashing their teeth about it for 1.5 months, and the X content was only stable-ish for a week before that, and was preceded by another a month of striving, then that 1-week version of X is not the "stable" version to revert to. But if the version before that was left alone for 6 months, then obviously that is the one. In another case, if there's been 2 months of see-sawing, preceded by 3 months of stability, before which was 2 years of slow editwarring and other conflict, then that 3-month version is the one to revert to (it likely was the hard-won compromise of that longer dispute period).

  I would add the caveat that making a change no one notices for a year, because it's in a page people don't pay much attention to, doesn't magically make it the stable version; WP:BRD and WP:RMTR#Requests to revert undiscussed moves are both often enough invoked about questionable changes after that long or even longer. But they would not be at an article like Jacinda Ardern which has boatloads of watchers and many active editors. The stable version is a version before the currently disputed one, unless the currently disputed one has good evidence of enjoying some actual consensus (e.g. because it's in a guideline or policy (in a section people actually cite), or an article with many watchers, and has been around long enough that plenty of people have seen it and not raised objections or done reverts. It's also important to remember that we only revert to last stable version when current attempts to establish a consensus have failed, after considerable effort. You can't invoke status quo ante without doing the D in WP:BRD in good faith with the other editors who care to participate. This is a lot like the MOS:ENGVAR, MOS:DATEVAR, and WP:CITEVAR stuff: if consensus fails to emerge from a debate, including with regard to the last stable version, then the final fall-back position is to revert to the choice made in the first non-stub or "major" version of the article. This doesn't mean that the person who made that edit has more say in the discussion moving forward, nor does it endorse their version as better, it's simply something to do to bring that round of "discussion" to an end. But the discussion has to happen first.
   — SMcCandlish ☏ ¢ 😼  16:25, 6 April 2021 (UTC)

• I mostly agree with what others have said that the amount content etc has changed is a significant factor not just time passed however I would say that if an undiscussed change has been made over 2 years ago it should be considered stable even if it could reasonable have been deemed controversial at the time it was done. An exception would still be for moves against express consensus or RETAIN. Crouch, Swale (talk) 17:00, 6 April 2021 (UTC)
• A few additional points: where the new title has been endorsed in an RM discussion (and then confirmed by a move-review) as in Humour, this surely shows the name has become "established" for WP:RETAIN purposes. Further argument based on the original title are not supported by WP:RETAIN. Similarly, eliding from the actual original title to a new desired title (e.g., Telephone game -> Telephone (game), as in the Chinese whispers case) is not a strong argument - if the original title was not actually any good then you cannot use WP:RETAIN to argue in favour of a different title. Yoghurt -> Yogurt is not a precedent-setting case when Humour (which was a reviewed discussion) exists. Finally, people repeatedly raising the issue on the talk-page is not evidence that something has to change - these issues will always be raised on the talk page. For example the remerging of Where's Waldo? back into Where's Wally?, which happened in 2011, is still being raised as a problem on the talk-page there despite there being no actual grounds in policy at all for doing so (they are obviously just different names for the same British-origin media franchise). FOARP (talk) 09:58, 7 April 2021 (UTC)

  The humour RM was closed as "not moved" which indeed probably establishes consensus and in any case like we pointed out about the ATDAB exception in Chinese whispers, editors pointed out in the humour RM that the article may still have been a stub when it was changed so I don't think the automatically reverting back per RETAIN would have applied (and would probably have gone to move review even if more editors supported) but its quite clear that both a significant amount of time and the article has changed significantly would suggest its (relatively) stable at its current title. Crouch, Swale (talk) 10:13, 7 April 2021 (UTC)

• I am going to backup FOARP here and second the prevailing thought that what has subsided for hundreds of edits and in some cases years should be considered to be consensus. ♥ Solarra ♥ • _{T ♀ C} 18:34, 8 April 2021 (UTC)