Wikipedia talk:Redirects for discussion

RfD of the future

J947's comment in a recent discussion prompted me to think a bit about the overall pattern of what we do at RfD. What we have now is a boutique process: we discuss individual redirects, which we consider on their own individual merits, and this works great when we need to decide if X is unambiguous, if Y is a plausible synonym for its target, or if Z has no history worth preserving. Sometimes, however, the discussion of one redirect will centre on a characteristic that this redirect shares with thousands of other (potentially) existing redirects, and it may feel desultory to spend time deciding this one case, when countless others can be nominated in the future, each redirect to be debated on the same grounds every single time.
I feel like we're a new neighbourhood with good systems in place for deciding small matters, like what colour to paint this fence here (all that matters is that it fits with the choice made next door). But we don't have systems in place for deciding larger-scale questions, like the choice of a street numbering system (it's not enough that the street numbers of two adjacent houses make sense; street numbers are any good only of they're consistent from one end of the street to the other).
What we lack is a set of mechanisms for dealing with classes of redirects. We need tools for querying ("List all redirects where %20 stands for a space, leaving out any that have had their targets manually changed in the past"). We need tools for probing user experience – we always presume what readers want but we don't have a way to get any sort of data ("Let 5% of readers who follow any of those 2500 redirects in the next two weeks be presented with a survey question asking them if they've found what they were looking for"). We need tools for implementing class-level decisions: besides "keep" or "delete", we could have "replicate" ("Create this type of redirect for all eligible titles"). We need tools that will allow us to go beyond redirects ("Let the search engine treat the following strings as equivalent.."). We need....
Anyway, these are just my late-night daydreamings. What are your craziest RfD fantasies? – Uanfala (talk) 23:27, 18 September 2020 (UTC)

• Interesting thoughts. The "boutique" approach may be more feature than bug, since it allows for nuance. There are cases where RfD consensus amounts to "this specific redirect is inappropriate" and those that are more like "this type of redirect" is inappropriate. In plenty of cases, I've pointed to WP:NOTBURO and deleted "Bad redirect" after there's consensus to delete "Bad Redirect". I do wish we could make better use of redirect tagging, especially {{R avoided double redirect}}, to help bring in related redirects, to get some of the benefits of linked data.

I'd approach any sort of batch creation of redirects with caution, though. What's good for the goose is not always good for the gander, and there's big danger of clutter, if nothing else. --BDD (talk) 15:58, 16 October 2020 (UTC)

RfD and restoring problematic articles

This is a follow-up to a recent RfD discussion. There's a scenario that happens from time to time: a redirect is nominated for deletion, during the discussion it's discovered that the redirect was an article at some point in the past, participants then point out that it's not RfD's job to evaluate content and so proceed to recommend restoring the article, at the end of the discussion the article is restored. So far, so good. But sometimes the restored article will be of such a low quality that, had it been created anew, it would have been picked up at the new page patrol and immediately sent for deletion or draftified. But because it was restored as a result of an RfD, it's marked as reviewed, so it bypasses the page patrol. I guess the assumption here is that if anyone has objected to the new article, it would be their job to send it to AfD. But this assumes that the editors concerned will notice the restored article and it relies on their initiative. I don't think these two should be taken for granted. And really, when it comes to problematic content, the burden of proof should be on those who advocate retention rather than those who argue for removal.

Maybe the two most recent examples are extreme (very short stubs with little or no sourcing, little time spent existing as an article, doubts previously expressed about viability of the topics, a small number of participants in each RfD), but something like that comes up every now and then, and I think we should do a better job of dealing with it. Because in our punctiliousness about correct discussion venues, we end up undermining fundamental content policies.

What is to be done? I've got one suggestion: if an RfD results in the restoration of an article, and there have been reasonable doubts expressed about the existence, encyclopedic viability, or notability of the topic, then the closer of the RfD discussion should proceed to procedurally start an AfD nomination. Would that work? Any alternative ideas? – Uanfala (talk) 20:44, 1 October 2020 (UTC)

I think a pretty easy fix is to just have admins implementing such closes be careful to mark the page as unreviewed after restoring it. Most NACs shouldn't be an issue, as the article will not be marked as reviewed when they restore it. The only edge case is if an autopatroll editor without new page reviewer permissions makes an NAC. _signed, Rosguill ^talk 20:49, 1 October 2020 (UTC)

I've got both a technical and a fundamental objection to that. The technical one: unless I'm mistaken, once a redirect is nominated for RfD, it's no longer counted as a redirect so it enters the NPP queue, and it will get marked as reviewed, typically long before the RfD is closed. If the closer is an admin or patroler, yes, they can mark it as "unreviewed", but any patroler who then picks it up will see the article was the result of a discussion and will probably assume there's consensus for it. More fundamentally, I really don't like the idea of RfDs getting closed with what is essentially The result of the discussion was restore garbage. If a major problem is pointed out, it needs to be addressed, I don't think we should be making a mess and then hoping someone else will come to mop it up. – Uanfala (talk) 21:08, 1 October 2020 (UTC)

I don't think the technical analysis is correct: when the RfD is closed, first the tag is removed (essentially recreating the redirect), and only then does the closer manually restore the article. When I've closed a discussion as restoring the article and go to mark it as unreviewed, it shows up as if I reviewed it (i.e. my autopatrol), rather than whoever checked it off in the queue. _signed, Rosguill ^talk 22:36, 1 October 2020 (UTC)

Oh, I see. But that happens only when the XFDscript is used; if you close the discussion manually (rarely though it may happen nowadays) then it won't go through an intermediate stage of being a redirect again and will remain patrolled. – Uanfala (talk) 22:46, 1 October 2020 (UTC)

There are only two circumstances in which it is appropriate for RfD to delete article content:

• Where the content has been previously discussed at a competent forum (usually this means the article talk page or AfD) and a consensus reached that it should not be an article.
• Where the content would be speedily deleteable as an article

In all other circumstances it is wholly inappropriate for a discussion about a page that was redirected without discussion to be deleted. In some (although obviously not all) this has been used as an end-run around an AfD discussion in an attempt to remove content that would never get consensus for deletion as an article. Even when not done in bad faith, and even when separated by years it is important for the integrity of the encyclopaedia that we give article content a full hearing at a forum where those interested and experienced in discussing and evaluating article content will find it (e.g. through article alerts). Thryduulf (talk) 09:48, 2 October 2020 (UTC)

• And another recent example: Wikipedia:Redirects for discussion/Log/2020 October 8#Fjollträsk. Here an article was created in the course of the RfD discussion, which was a botched translation of a Swedish Wikipedia article. So botched that the result would have probably qualified for G3, and yet the RfD discussion was speedy closed and the concerns about the article were brushed aside, both there and in one of the follow-up discussions. Again, that may be an extreme example, but we clearly need to do better. – Uanfala (talk) 12:30, 10 October 2020 (UTC)
  • While the follow-up might have been botched, the closure here was not. RfD is not the place to discuss article content - that is the job of AfD and it should have been nominated there. Thryduulf (talk) 14:07, 10 October 2020 (UTC)

I am sorry, I was following this advice, which I deemed competent, in that case. Any and all efforts to prevent anything like that article happening again to circumvent a redirect deletion would be very constructive. I had a gut feeling there was something wrong with the user who, as I saw it, took advantage of a loophole to create nonsense. Again, I am sorry for any and all of the animosity involved. --SergeWoodzing (talk) 20:57, 10 October 2020 (UTC)

• "Not RfD's job to evaluate content" does not, and should not, mean "RfD can't evaluate content". WP:NOTBURO is my lodestar here: if there's consensus for an action, we can follow that action. Deleting a former article is no more inherently problematic than RfD resulting in a page move. I don't think anyone would call such an outcome illegitimate because it wasn't through WP:RM.

Now, the usual caveats apply. If it's clear someone is forum shopping or attempting an end-run around some other process, of course we shouldn't allow that. If such a redirect is deleted, there may be some ground for reversing it if someone later points out there was significant page history and that wasn't addressed in the discussion. And to be very clear, no one should be converting an article to a redirect and immediately taking it to RfD.

But otherwise, as a closer, I'm always going to assume competence on the part of RfD participants. "Don't delete because there's page history and we should be at AfD" and "Delete, even though there's page history" are both perfectly legitimate arguments. --BDD (talk) 16:07, 16 October 2020 (UTC)

The very significant difference between a move and a delete is that the former can be done without (formal) discussion the latter cannot. Article prose that is not speedy deletable and/or has not been discussed in an appropriate location should never be deleted at RfD. This isn't about bureaucracy it's about making sure that relevant editors are aware that article content is being discussed and AfD's greater prominence plus article alerts and deletion sorting is always going to be vastly superior to RFD at enabling this. This protects against both intentional and unintentional bypassing of AfD. Thryduulf (talk) 20:12, 16 October 2020 (UTC)

{{R without mention}} discussion

Hi fellow redirect folks. I'd appreciate any feedback at Template talk:R to article without mention#Refining/clarifying usage. --BDD (talk) 15:51, 16 October 2020 (UTC)