The following is a draft working towards a proposal for adoption as a Wikipedia policy, guideline, or process. The proposal must not be taken to represent consensus, but is still in development and under discussion, and has not yet reached the process of gathering consensus for adoption. Thus references or links to this page should not describe it as policy, guideline, nor yet even as a proposal.

Several prior discussions have indicated editor concerns with elements of deprecation. Currently no guideline describes deprecation and all that it entails. There is:

• a section within WP:RS (Wikipedia:Reliable_sources#Deprecated sources),
• the information page WP:DEPS: Wikipedia:Deprecated sources,
• another information page WP:DEPREC: Wikipedia:Reliable_sources/Perennial_sources#Deprecated,
• and the definitions and information within the WP:DAILYMAIL1 and WP:DAILYMAIL2 RfCs (the first source to be 'deprecated', see History section).

Related discussions at ANI:

• Wikipedia:Administrators'_noticeboard/IncidentArchive1087#Can we decide what the heck "deprecation" means, or alternately, use a different word? leading to this RFC.
• closed ANI discussion Wikipedia:Administrators'_noticeboard/IncidentArchive1087#User:David Gerard violating Wikipedia:Deprecated_sources#Acceptable uses of deprecated sources (of which the immediately above referred to disussion was a part).

Purpose of deprecation

What is the meaning and purpose of deprecation? How is this distinct to general unreliability?

Discussion (Purpose of deprecation)

It seems reasonable. I would add for deprecated, "self-published or user-generated content authored by established subject-matter experts is not acceptable." That is the key distinction between generally unreliable and deprecated. If a source is deprecated, we cannot be confident that the article was written by the named author or that it has not been altered. Private Eye for example had a diary supposedly written by then UK PM John Major. TFD (talk) 14:53, 30 December 2021 (UTC)

I don't think this (If a source is deprecated, we cannot be confident that the article was written by the named author or that it has not been altered) is the (or even a) consideration when deprecating a source. If a source is altering someone else's content, or falsely writing that someone else authored it, then it's probably Spam Blacklist eligible. I'm not even sure the Daily Mail does that. Private Eye is a satirical magazine; I'd have to see the article in question, but I doubt someone reading the entry would seriously think John Major wrote it. ProcrastinatingReader (talk) 14:56, 30 December 2021 (UTC)

I wouldn't say it's a specific consideration in deprecation discussions, but it's definitely relevant afterwards. Once we've established that a source doesn't faithfully report facts in general, that necessarily applies to any individual facts, like ​who authored a particular article - and similarly, they might take excessive editorial liberties for the sake of clicks, and so on. If there are deprecated sources that we do trust to that degree, I suppose we could make a further distinction between deprecation and the spam blacklist, but I'm not sure that there's a useful one to be made, and that might fall outside the remit of the spam blacklist regardless unless we redefine it. Sunrise (talk) 00:07, 31 December 2021 (UTC)

I don't see that this clearly differentiates why we should have "unreliable", "deprecate" and "blacklist". I get there is a big difference between a source that is unreliable because it "reliably" reports the claims of "crap" experts as fact vs a source that falsely attributes claims to an expert. One is unreliable because the messages they report aren't worth reporting while the other might actually be asking good and important questions but is changing the answers provided. My understanding is the reason the DM was deprecated vs just "unreliable" was editors couldn't trust that quotes etc weren't altered. While certainly there is a difference between how editors will be allowed to use an unreliable vs deprecated source, this doesn't make it clear why a source should be deprecated vs marked as just unreliable. Honestly, most of the debates on the topic come across as, "deprecate because a lot of editors don't like it". I will note this comment clearly hits on the Criteria for deprecation section below. Springee (talk) 04:55, 2 January 2022 (UTC)

Current definitions

Generally unreliable: Editors show consensus that the source is questionable in most cases. The source may lack an editorial team, have a poor reputation for fact-checking, fail to correct errors, be self-published, or present user-generated content. Outside exceptional circumstances, the source should normally not be used, and it should never be used for information about a living person. Even in cases where the source may be valid, it is usually better to find a more reliable source instead. If no such source exists, that may suggest that the information is inaccurate. The source may still be used for uncontroversial self-descriptions, and self-published or user-generated content authored by established subject-matter experts is also acceptable.

Deprecated: There is community consensus from a request for comment to deprecate the source. The source is considered generally unreliable, and use of the source is generally prohibited. Despite this, the source may be used for uncontroversial self-descriptions, although reliable secondary sources are still preferred. An edit filter, 869 (hist · log), may be in place to warn editors who attempt to cite the source as a reference in articles. The warning message can be dismissed. Edits that trigger the filter are tagged.

Name

Are "deprecation" and "deprecated sources" the best terms for the concept described in section above?

Discussion (Name)

At WP:V, "source" includes three possibilities, publisher, work and author. The words "unreliable" and "deprecated" are not present, "reliable" and "questionable" appear plus a pointer to WP:RS, where "unreliable" does appear ("No source is 'always reliable' or 'always unreliable' for everything.") and where there is a section for "deprecated sources" (as mentioned in the introduction above). — Preceding unsigned comment added by Selfstudier (talk • contribs)

• Sorry if this is the wrong place, but I should note the description above for deprecated -- "The source is considered generally unreliable, and use of the source is generally prohibited" -- differs substantially from what the linked page actually says, "Deprecated sources are highly questionable sources that editors are discouraged from citing in articles, because they fail the reliable sources guideline in nearly all circumstances" (emphasis mine in both cases). The latter is more in line with the dictionary definition of deprecate and I think we should strive to use words for what they actually mean, not our own internal definition. (I should also note the definition of deprecated given at Wikipedia:Reliable_sources#Deprecated_sources differs somewhat from the other two already cited above.) Calidum 17:31, 30 December 2021 (UTC)

I wouldn't think of them as being different definitions, but rather as different ways of saying essentially the same thing. A general prohibition is a type of discouragement, and the rest of WP:DEPS (including the followup "in nearly all circumstances") should make it clear what form the discouragement takes. Likewise, the third example you mention from WP:RS is simply focused on what that means in practice ("they should not be used, unless there is a specific consensus to do so"). I suppose if you have to pick one to be "official", the language "generally prohibited" is directly taken from the summary of consensus in DAILYMAIL1. Sunrise (talk) 00:07, 31 December 2021 (UTC)

@Sunrise, I think there is a big difference between "generally prohibited" and "discouraged". We discourage a lot of permissible things (e.g., using unnecessary technical jargon in articles, citing journal articles without providing a URL/doi/PMID/other link , adding only a bare URL in an article because you don't know how to format a citation properly, creating articles that are obviously verifiable but for which you didn't post any sources in the first draft...). See also Wikipedia:Strongly Discouraged. In the "generally prohibited" category, I'd put copyvios and circumventing the spam filters. Those are things you normally must not do, i.e., "generally prohibited". WhatamIdoing (talk) 03:44, 2 January 2022 (UTC)

I would agree that it could be interpreted that way if comparing the two phrases in isolation. However, I think the context (including, but not limited to, the second half of the same sentence) makes it fairly clear which interpretation is being used. Sunrise (talk) 05:58, 4 January 2022 (UTC)reply

When we use words that "in isolation" mean something different from what we intend, editors will come to different conclusions about what is actually meant. This is an avoidable source of disputes. WhatamIdoing (talk) 02:50, 8 January 2022 (UTC)reply

• This requires an answer to the "purpose" question first. If the purpose of deprecation is general prohibition, I'd like to see it called "prohibited publishers" or "banned sources" or similar, rather than the WikiSpeak definition of "deprecated" (which is rather out of touch with the dictionary definition, and besides is software lingo). This also helps editors at RSN be more clear on what they're voting for. I don't think sources with some history of fabrication but a lot of history of good publication should be 'prohibited'. It requires volunteer judgement, for sure, but 'generally unreliable' seems more accurate in such a case. ProcrastinatingReader (talk) 19:33, 31 December 2021 (UTC)reply

Criteria for deprecationedit

What is the criteria for a source being 'deprecated' (as opposed to being 'generally unreliable')? What kinds of evidence should editors look for to make this determination?

Discussion (Criteria for deprecation)edit

• One thing I wanted to get at when creating this RfC is inconsistency in accepted evidence. Consider Independent Press Standards Organisation results. In the RfC for The Mail on Sunday 3 editors decided to ignore IPSO outcomes. David Gerard (I'm choosing his comment because he was the one of the only ones to address the 'IPSO argument for reliability' head-on) wrote Because IPSO complaints are not the be-all and end-all of whether a source should be deprecated in Wikipedia, and IPSO is widely regarded as a captured regulator. In other RfCs, we decide IPSO outcomes are highly relevant for deprecation and consider them heavily, sometimes use them as a sole basis for deprecation. This seems like an inconsistent approach to me. I'd appreciate specific guidance on what kinds of evidence, in what kind of quantity, constitutes enough evidence for deprecation. I appreicate it's not an algorithm, but with over 40 deprecated sources now, I think we're at least able to write up roughly what kinds of evidence is considered convincing. Right now, it's somewhat a free-for-all. ProcrastinatingReader (talk) 23:35, 29 December 2021 (UTC)reply
• Consider IMPRESS as well? Selfstudier (talk) 11:47, 30 December 2021 (UTC)reply
• To my mind, we are over-using “deprecation”. It should be reserved for those few sources that go beyond “general unreliability”. Blueboar (talk) 13:13, 30 December 2021 (UTC)reply
• I would expect a relevant reliable source that says the publication routinely publishes false information that is not corrected. That would not be a problem with hate groups and conspiracy theory websites. TFD (talk) 14:55, 30 December 2021 (UTC)reply

  So then what's the criteria for general unreliability? Because I thought that was the requirement for a source to be considered GUR. ProcrastinatingReader (talk) 23:03, 30 December 2021 (UTC)reply

  My impression for GUNREL is that the standard is basically "whoever showed up that time didn't like it". I have seen editors claim unreliability because (e.g.,) the news source seemed "biased", meaning that it chose to run certain stories but not others, or that it presented the views of one side by not the other side, and that viewpoint is not the one familiar to the editors in the discussion. To pick two completely made-up random country names, "China" and "the US", a news source may be called "GUNREL" for everything if mostly-non-Chinese editors feel that it has a pattern of reporting what China's politicians say that China's government's viewpoint is on a particular dispute between China and the US, without giving what they feel is adequate airtime to what the US's politicians say that the US government's viewpoint is on that same dispute. The problem IMO is the "for everything" part; IMO such sources should be considered highly reliable for what they actually report (e.g., the title of the person being quoted; the quoted statement; China's view at that point in time). WhatamIdoing (talk) 03:56, 2 January 2022 (UTC)reply

I also question, at least partially, the entire foundation of deprecation. For example, and I am not saying FrontPageMag should be cited for anything ever, but seven people at RSN in one discussion is enough to decide that a source may not be used on Wikipedia at all? How does that make sense? There are currently 5 RFCs on RSN right now. When did that stop being a place to discuss if a given source is reliable in a given context? It has turned in to a means to attempt to banish sources across Wikipedia based on essentially click-bait (they published this thing I dont like this one time). This should be reserved for truly exceptional circumstances. I cant imagine any professional publication saying, except in extremely rare circumstances (and 5 times right now is not rare), that such and such website should never be used for any purpose regardless of the circumstances or particulars. nableezy - 10:24, 4 January 2022 (UTC)reply

I whole heartedly agree with this concern. I think this is a problem that is born out of the RSP list. In the past editors were a lot more likely to use RSN to argue *this* article is reliable/unreliable for a particular claim. With the advent of the RSP list the objective seems to be to secure a ranking as that ranking now carries a lot of assumed value. This is fine with some of the big sources we have discussed. The rankings for things like Fox News, CNN, Daily Beast etc are likely to be rather stable since we have a lot of discussions and a lot of editors who participated over time. However, for a smaller source we might have only a few editors who weigh in. After that it's easy for someone to add a source to the RSP list based on the opinions of just a few editors in one or two discussions. In the case of deprecation, if we can't get at least say 10 editors (no idea what a good number would be) to weigh in how can we say there was enough input to apply such a drastic restriction to a source? This is why I'm concerned that we don't have a clear definition of when/why a source should be deprecated. Even if we agree what to do with deprecated sources (lock them up!) if we can't agree on the crime we moved away from reason. Springee (talk) 13:00, 4 January 2022 (UTC)reply

I think that "a lot of discussions and a lot of editors who participated over time" is a key point. It would be useful to have a central archive for subjects that have been discussed repeatedly. It is not so useful to spend time on major discussions that could be handled as simple questions. WhatamIdoing (talk) 03:05, 8 January 2022 (UTC)reply

"What kinds of evidence should editors look for to make this determination?" is a great question. In a recent, well-attended deprecation RfC, the closer discounted evidence presented that the source was unreliable. They reasoned that since the choice was between "generally unreliable" and "deprecated", evidence of unreliability was not relevant. My view is that there is "generally unreliable" covers a broad spectrum of bad sources, and that evidence that a source is on the stinky end of the shit stick should be a part of deprecation proceedings. Firefangledfeathers 06:40, 5 January 2022 (UTC)reply

WP:QUESTIONABLE (RS) "Questionable sources are those with a poor reputation for checking the facts or with no editorial oversight....blah

WP:QS (V) Questionable sources are those that have a poor reputation for checking the facts, lack meaningful editorial oversight, or have an apparent conflict of interest....blah

Both versions are similar, idk why they are not identical.

Is gunrel intended to be different from this? (Def above says "Editors show consensus that the source is questionable in most cases."

And what is it exactly that turns questionable/gu into "stinky", fabrication? And? Selfstudier (talk) 11:28, 5 January 2022 (UTC)reply

+1 to "When did RSN stop being a place to discuss if a given source is reliable in a given context?"

If we want to have a vote-out-stinky-sources-wholesale page, let's have that page, but let's not have those discussions dominating RSN, which used to be a useful and interesting noticeboard. WhatamIdoing (talk) 03:03, 8 January 2022 (UTC)reply

Valid uses of deprecated sourcesedit

• Does/should deprecation apply to publishers and/or authors and/or single works?
• What is the difference in valid use between generally unreliable vs deprecated sources?
• When is it valid to cite an article in a deprecated and/or generally unreliable publication? (is it valid to cite an established subject matter expert who is published in a deprecated publication?)

Discussion (Valid uses of deprecated sources)edit

Wikipedia:Reliable_sources/Noticeboard#De-deprecate CounterPunch and Wikipedia:Reliable sources/Noticeboard#RFC: Counterpunch discuss the third point above. If consensus were possible that articles by experts a la WP:SPS may be cited even if published in a deprecated source, that would seem to be the simplest approach? The ONUS would remain with wouldbe citers to justify inclusion, probably it ought to be discussed first at a talk page. Selfstudier (talk) 12:09, 30 December 2021 (UTC)reply

I really oppose this if so called expert choose depreciated source to publish then we assume more respectable outlets didn't want to publish their views so their views are WP:UNDUE. Depreciated source should be used only WP:ABOUTSELF but only on the page source for example Daily Mail could be used only on Daily Mail page and nowhere esle Shrike (talk) 13:10, 30 December 2021 (UTC)reply

If the consensus is instead like that then the criteria for deprecation assume more importance, in other words people may be less likely to consider deprecation in the first instance since expert opinions are allowed for "merely" unreliable sources. Also WP:SPS stipulates that expert designation requires "work in the relevant field has previously been published by reliable, independent publications." Selfstudier (talk) 13:29, 30 December 2021 (UTC)reply

That assumption is based on what exactly? nableezy - 15:26, 30 December 2021 (UTC)reply

If we allow experts' articles in deprecated sources to be used, then there is no difference with generally unreliable. It then isn't a useful distinction for editors, but merely a collective opinion of Wikipedia editors. Experts may not have the same view of which sources are deprecated and may allow The Daily Mail or CounterPunch to publish their articles, although I expect they would draw the line at Alex Jones. But I wouldn't use an experts' article published in a generally unreliable source because it would lack weight. Weight after all is determined by coverage in reliable sources. TFD (talk) 15:03, 30 December 2021 (UTC)reply

The question is if an expert writes something in the area of their academic expertise is that a reliable source regardless of where it is published. Because if it is a reliable source, then it does have (some) weight. There are also examples of columns by experts in deprecated sources being cited repeatedly by other reliable sources. Can somebody argue that such a source does not have any weight? nableezy - 15:25, 30 December 2021 (UTC)reply

At that point the weight would come from the coverage in the reliable sources instead, and that coverage is what we would use for article content. Even if all they do is repeat the same information, it will have gone through their editorial and fact-checking processes, which makes it different from the same information published in the deprecated source. More generally, I suppose this could be a valid candidate for an exception to deprecation, but naturally consensus would need to be established in favor of it before it could be used. In addition, discussion of such an exception would have to specifically focus on information that is not reported in the reliable sources, as otherwise we could just use those sources instead. Sunrise (talk) 00:07, 31 December 2021 (UTC)reply

The language at WP:RS, "..should not be used, unless there is a specific consensus to do so." does suggest that exemptions are possible (beginning with local?).Selfstudier (talk) 10:54, 31 December 2021 (UTC)reply

My 3¢:

• Does/should deprecation apply to publishers and/or authors and/or single works?
  • I think it isn't pointful to declare any single work (e.g., a single book, a single news article including a multi-part series) to be "deprecated" or "generally unreliable". Individual publications should use the ordinary process of comparing the individual work against the specific material to be included in an article, and determining whether (a) that work is reliable for that material and (b) that material meets both DUE and WP:BALASP for that Wikipedia article.
• What is the difference in valid use between generally unreliable vs deprecated sources?
  • "Deprecated" should be considered closer to the spam blacklist, including the possibility of "whitelisting" an individual use. It should not be considered "banned". "Generally unreliable" should be considered more like saying that this is not a great source, but if you know what you're doing, and you are writing very carefully, it may be perfectly reliable for the purpose. Consider, e.g., the difference between writing "There is no good scientific evidence for using parachutes when jumping from an aircraft" and "The BMJ ran an article describing the evidence levels for parachute use", or the difference between "Chris Cultic is the Grand High Master of the Universe" and "The religious organization issued a press release saying that they had adopted a new title for its founder, 'Grand High Master of the Universe'."
• When is it valid to cite an article in a deprecated and/or generally unreliable publication? (is it valid to cite an established subject matter expert who is published in a deprecated publication?)
  • There are and should be different rules for the two categories. In both cases, it can be valid. The only exception is if the publication has a documented history of lying about who wrote their articles (e.g., writing articles and claiming that a Nobel prize winner wrote it – at the borderline for this, I might include wholesale copying of Wikipedia articles).
  • The main thing that needs to be determined is what to do if someone cites a source that you think is inappropriate to support content that you think is otherwise appropriate. Should you blank the good content? Blank the poor source but leave the good content? Go to the trouble of finding a better source yourself? Tag the source in the hope that someone else will do the work that you don't want to do yourself?

WhatamIdoing (talk) 04:18, 2 January 2022 (UTC)reply

Who is an expert?edit

There is another piece to this puzzle - the question of who qualifies as a “subject matter expert”? When we wrote the SME exemption, we were really thinking of academics who had previously published in respected journals … but that concept seems to have been extended beyond academia… especially to media columnists who have previously published their opinion/analysis in reliable news outlets. Have we taken the concept of SME too far? Blueboar (talk) 15:35, 30 December 2021 (UTC)reply

Wikipedia talk:Verifiability#WP:SELFPUB and the definition of a subject-matterexpert. is a current discussion.Selfstudier (talk) 15:44, 30 December 2021 (UTC)reply

Ah… thanks for the link. Blueboar (talk) 16:24, 30 December 2021 (UTC)reply

Removal of deprecated or unreliable sourcesedit

At present, the insertion of new references to deprecated sources is slowed down through the use of edit filters and bots, which is relatively uncontentious. However, how should existing references to deprecated and unreliable sources be dealt with?

The ultimate trade-off is between speed in removal vs judgement on a long-term solution. That is, is it acceptable to indiscriminately remove references to deprecated/unreliable sources and replace with {{cn}} tags? Or should each reference be judged on its own merits, and either a new reference found to support the underlying text, or the text removed along with the ref? Or something else?

Discussion (Removal of deprecated or unreliable sources)edit

• I don't think the "ultimate trade-off" is between speed and judgment. Or, rather, "speed" and "judgment" are both so subjectively perceived that complaining about choosing one over the other is destined to be fruitless. Removals will always be too fast for somebody and too slow for somebody else. Scattered examples of debatable removals will always loom large, because nobody is going to gather statistics on what the proportion of questionable removals actually is — and if they did, anecdotes would still win out over data. The underlying issue is the sheer number of deprecated links still lurking about. XOR'easter (talk) 23:33, 27 December 2021 (UTC)reply

I think that the concern is a bit overblown here. Your average removal of deprecated source looks like this. Does anyone really think that Robert Jungk did not continue working as a journalist after the war? Btw in this case the {{cn}} was added, which is the right thing to do.

I support removal of CP and other deprecated sources in most cases but let's not pretend that every single reference to them is hugely damaging and justifies extreme measures to remove it. Alaexis_¿question? 08:32, 28 December 2021 (UTC)reply

Removing every CP ref on sight, without even examining it, is what led to this discussion. By the way, 'pretend' in the sense of claim, is something of a Frenchism or Italianism, isn't it? I don't think pretence in the traditional sense has anything to do with the frenzy of elisions seen recently. It was one editor's honest conviction that deprecation demands such mechanical weeding out. Nishidani (talk) 22:31, 28 December 2021 (UTC)reply

After some thought it would be nice to see something along the lines of grandfathering deprecated sources. That is, in the immediate days after a source has been made deprecated, it would be inappropriate to remove such in a rapid fire bot like fashion, but in a year or so after that point, it should be okay. This is prevalent that the deprecation is well announced (VPP in addition to RSN or RSP and that editors should be away that the onus is on them to replace or remove deprecated sources before this de facto grandfathering period is up. After that year, then no editor should be able to complain about rapid fire removals. This balances the avoidance of mass disruption across a large number of articles in the wake of a deprecatuon determination, and actually achieving that deprecation in a reasonable time frame. And if a source warrants a faster removal (eg DM and BLP) that should be established during the deprecation debate. --Masem (t) 23:01, 29 December 2021 (UTC)reply

Thats only part of the problem, the other part is the unwillingness to engage in discussion about individual sources in a deprecated source that may or should be used. Is "its deprecated" enough to remove a source when evidence is offered to suggest it is reliable? nableezy - 23:07, 29 December 2021 (UTC)reply

I think that question (how to handle one specific case of a deprecated source), but I will add that what propose is a system when the onus initially is on those seeking removal to justify it (even if that means leaving a cn behind when no replacement sources can be found), to the long term onus on thise wanting to keep a cite to a deprecated source to justify its usage. --Masem (t) 00:11, 30 December 2021 (UTC)reply

As Aquillion pointed out to you when you proposed the same thing at WP:VPP a couple of years ago, an RFC that directly contradicts policy in this manner would be invalid - David Gerard (talk) 00:59, 30 December 2021 (UTC)reply

I have the impression that selecting specific events like the above, when it's a rather standard procedure to tag or remove unreliable sources, may smell of WP:POINT and not be very convincing. I too have removed and tagged some, there are others but an example I immediately remember was creationist homeschooling textbooks by DI imprints used in WP (liberaly mixing preaching editorials everywhere including math). Those did not require specific consensus deprecation and better textbooks were easy to use. When a source does, isn't it an even better indication that consensus is against their general use by current policy? Context-specific requests can, and are posted at RSN and often the answer is that in some specific cases it may be usable (also by existing policy, like limited WP:ABOUTSELF). Also by policy, material only covered in unreliable sources and that cannot be written about from the point of view of better analysis is generally considered WP:UNDUE. When it's a WP:BLPRS issue, even generally reliable sources that are not considered by all of the community to be independently published, are routinely removed (an example being Skeptoid that some editors target even if it's technically not necessarily published by the guest and the article is borderline BLP, i.e. it has been retargetted to a particular claim more notable than its proponent; I could, too, cherry-pick many examples of source removal). It seems to me that policy is already good enough to resolve most situations and that when unsure, case-specific discussion can take place. —PaleoNeonate – 03:37, 30 December 2021 (UTC)reply

I forgot this but it's important and beyond deprecation: guides like RSP can also be more flexible than yes/no and specify circumstances where a source is infamous, or can be used, depending on its reputation and the conclusions made at RSN. A great example is Fox News that hasn't been deprecated even if the shows and editorial stance are unreliable. —PaleoNeonate – 03:50, 30 December 2021 (UTC)reply

I went through the VPP from 2019 (#150) but can't find that discussion that you are refering to (at least, where Aquillion is said to have responded), but 1) policy is not law 2) my point is simply to minimize disruption in the wake of a source being labeled "deprecated", and minimizing disruption is also part of policy. Once a source is marked deprecated, the end goal is to eliminate all uses of deprecated sources outside the limited exception set (which is part of this larger discussion but not important here). But particularly with a source that has had wide use before deprecation was applied can't just be stripped out of WP instantaneously without potentially harming the work. (When we blacklist a source, on the other hand, we've established that removal is a higher priority than minimizing disruption). Hence having a brief grandfather period before mass remove of deprecated sources would be the normal way that disruption is minimized while still driving to meet the end-goals of WP:V and other policy. That's why it is important to understand and or establish the distinction between deprecated from blacklisting. --Masem (t) 15:46, 30 December 2021 (UTC)reply

I support the "grandfathering" idea, I think removal of "questionable" sources should be somehow subject to a (semi) automatic procedure that allows time for exceptions to be argued while getting shot of the rest in a reasonable time frame.Selfstudier (talk) 15:57, 30 December 2021 (UTC)reply

@Selfstudier, WP:QUESTIONABLE sources are officially permitted, so they shouldn't be subject to any automatic or time-based removal process. A "questionable" source is any source that at least one of these four qualities:

• expressing views that are widely acknowledged as extremist,
• promotional in nature,
• relies heavily on rumors
• personal opinions

Think about that second item in particular, and how often we cite a subject's official website. If we remove questionable sources, then we can't cite: any manufacturer's website about their products (Special:LinkSearch tells me that there are 80K links to microsoft.com on wiki – you volunteering to remove those?); any book/film/author website; any politician's campaign website; basically any press releases at all; and the list goes on and on. WhatamIdoing (talk) 04:44, 2 January 2022 (UTC)reply

At Wikipedia:Village_pump_(proposals)/Archive_164#Discussion_on_Proposal_3 - Aquillion points out that "'We want to deprecate this source' does not mean 'we want to provide special protections for existing usages of this source'", in fact it means the opposite. Your proposal is to provide more protections for sources explicitly deprecated by a broad general RFC than are provided for merely unreliable sources, which can be removed summarily.

He also notes: "WP:RS is core policy and not subject to consensus; therefore, you can always remove an unreliable source on sight with the reason of 'unreliable source', no matter what, without exception" - though actually WP:RS is a guideline included by reference in WP:V, which enforces that provision. An RFC can't actually find against that, and an RFC proposing to do so would be invalid. Your proposal would require a policy change to enforce, or at the least an RFC to alter all previous deprecation RFCs - David Gerard (talk) 19:02, 30 December 2021 (UTC)reply

What prior, formal RFCs (not just proposals discussed at VPP or elsewhere, but actual RFCs) have there been on the nature of what deprecated sources mean and how they should be handled. We as a community have talked around the matter for a few years, but this is the first actual RFC that I'm aware of to try to nail down those facets. So there would not be any alteration of prior results. Also, I would point out that there's a difference here of what I'm talking about, specifically a grandfathering period after a source becomes deprecated via an RFC, and not the issue of dealing with long-term deprecated sources as that link above was more focused on. Long-term, I tend to agree we can be more indiscriminate about removing deprecated sources well after the source has been established as deprecated ,but the short-term disruption -as well as wider notification that a source is deprecated and will be removed indiscriminately by X date - is what needs to be handled better. That's not really anything discussed in that prior VPP page. --Masem (t) 02:16, 31 December 2021 (UTC)reply

Perhaps some stats on the usage levels of deprecated sources might be useful here? If it's an issue, I think it would be fine to establish something like: starting X months after deprecation, deprecated sources can be removed indiscriminately, except where there is consensus to retain it, to be indicated by a specific "Acceptable use of deprecated source" template. The challenge would be in making sure that doesn't get interpreted as a way to protect those sources for X months, but if those sources are being left in articles regardless then perhaps it would still be an improvement over the current system. Sunrise (talk) 00:07, 31 December 2021 (UTC)reply

This is sneaking in Masem's assumption that the current removals are "indiscriminate", which is something he's consistently failed to show to the satisfaction of others. Remember that he's been trying to put this plan for special protection of deprecated sources into place for a couple of years now, and it's failed to achieve consensus every time - David Gerard (talk) 12:25, 31 December 2021 (UTC)reply

I have shown on several pages and the hatted examples above that your reverts are indiscriminate. A fresh one, among many today, is instanced here, with the edit summary:'Rv repeated deliberate addition of deprecated sources to wikipedia - do not do this').

This revert took place after David was notified on the talk page that the policy he cites allows for exceptions. Nishidani (talk) 13:56, 31 December 2021 (UTC)reply

Oh please. Two or more removals per minute does not allow time for considered judgement (which would simply involve a very basic search for replacement sources in many cases – something that has almost never been done) so they cannot be described as anything other than indiscriminate. _wjemather^{please leave a message...} 16:37, 31 December 2021 (UTC)reply

Once more, indiscriminate, and when contested, indiscriminately restoring his excision in the face of a talk page challenge. The only argument is 'deprecation' means I can removed everything, apparently. See here, here and here persists in removing Uri Avnery as a source for Amos Kenan, though the two were intimately acquainted and this is an obituary written by an insider. He does so by not responding on point on the talk page and against two editors who disagree with his reading of deprecation policy. Encyclopedias are written by people who generally have extensive knowledge of the specific topic, not by editors who have a strong but partisan reading of one policy that appears to give warrant to consecutive blind erasures regardless of content familiarity or expertise. Nishidani (talk) 16:46, 31 December 2021 (UTC)reply

Another reason to automate the process, thereby avoiding such accusations.Selfstudier (talk) 12:32, 31 December 2021 (UTC)reply

The problem is that you have to determine when you have "<content> <citation to deprecated source>" if the content needs to stay (now with a CN behind it) or go, and no bot will be able to do that. But once we're in a mode of doing that indiscriminately (after grandfathering period), it only takes a few seconds for a user to make that determination -eg what DG has been doing reflects the rate and outcome I would expect once we're past grandfathering. I don't think we want all deprecated sources replaced with CNs, given that this would not be appropriate on BLPs, for example. --Masem (t) 18:36, 31 December 2021 (UTC)reply

What I had in mind (if it is possible) is that all "questionable" sources (I am not going to distinguish between unreliable and deprecated because I don't really see that much difference apart from the SPS aspect) get tagged (could even cat them, maybe) with auto removal after a time period unless (talk page discussion/something). So people get a chance to make the argument and if they don't, it's gone.Selfstudier (talk) 18:44, 31 December 2021 (UTC)reply

Using a bot to tag articles creates the problem that tags would never get removed, which is also a problem as identified in these discussions on deprecation. This is different from humanly deciding that removing a deprecated source but likely with a truthful statement that just needs a CN or even keeping a deprecated citation tagged appropriately, compared to outright removal. That all sstill requires a human review. --Masem (t) 18:33, 2 January 2022 (UTC)reply

It is not possible to auto remove within a timeframe? I didn't know that, that's no use then, if there is no time pressure to remove then I agree with you, they will just sit there. Selfstudier (talk) 18:39, 2 January 2022 (UTC)reply

This is what we see in practice. Tagging bad sources results in nothing happening, often for years. Removing them (either bad cite and claim, or replacing with {{cn}} sometimes results in a replacement, but vastly more often they were always bad and stay gone. Functionally, "why don't you just tag them first/instead" is a way to not do anything about the bad sourcing - David Gerard (talk) 20:44, 2 January 2022 (UTC)reply

You are describing problems that are the direct result of indiscriminate mass removals. If WP:BURDEN and WP:PRESERVE were being followed, these things would hardly ever happen. _wjemather^{please leave a message...} 21:46, 2 January 2022 (UTC)reply

Special protection for deprecated sources is a bad idea and not worth further discussion IMO, although I can see where the idea comes from. Thing is, these removals would not be an issue (regardless of speed) if there were actual checking being done on removal, and (per the close of WP:DAILYMAIL1: Volunteers are encouraged to review them, and remove/replace them as appropriate.) an actual 'review' was being done and the 'appropriate' action was being taken. Thing is, many policies break down when applied indiscriminately and at scale, because almost all of them require actual human judgement, and I think that's where the dispute with your style of removal comes in, because you place more value in having an immediate disclaimer that the content is effectively uncited, whereas other volunteers place more value in retaining the encyclopaedic content if possible (or having the content removed otherwise). Regardless, I don't think the solution to this problem is a set of bureaucratic rules which add more protections for deprecated sources. ProcrastinatingReader (talk) 13:41, 31 December 2021 (UTC)reply

What the DM close should have done (and what we've failed to do on all such deprecation RFCs since) is to set a grandfathering period. Using the comp.sci. meaning, users are usually given fair warning how soon a deprecated feature will no longer be supported as to find alternatives, and that's the same thing we should be doing. We should have had consensus set a reasonably time frame for editors to remove/replace them, but we didn't, and now we're at the conflict being for any now-deprecated source "they must go ASAP" versus "giving indefinite time to be fixed, which often means never". Deprecated sources should be removed and not allowed to linger, but we should be accounting for the fact that until the RFC that marked them deprecated they were used in good faith as valid sources, and thus giving a reasonable amount of time for human review to try to remove/replace them without disrupting other content. And to me, six months after an RFC on deprecation of a source is closed and announced to VPP/CENT is more than sufficient time before we go from "careful, human review" of each instance to indiscriminate removal, as well as consist when we have had to grandfather other content due to cchanges in core policy (eg when NFC was required by the WMF in 2008 for example). --Masem (t) 18:36, 31 December 2021 (UTC)reply

Masem. As a content editor I'm still puzzled. David Dean Shulman is a scholar of international standing, who has written 2 academic books on the dispossession of Bedouin in the South Hebron Hills area, of which he has perhaps a unique knowledge, since his Sabbath for 20 years has consisted in being present there to prevent their harassment. Since he has an unparalleled knowledge of innumerable incidents, why should he be deprecated as a reliable source because when not publishing for the New York Review of Books on some incidents, or Haaretz on others, he chooses CounterPunch for some details, whose wiki reputation as a genocide-promoting, Holocaust-denying, conspiracy-mongering antisemitic website is unknown to him (as it is to peon tillers of the wiki patch like myself who have read it (selectively for the abundant quality articles it carries) for almost two decades), or if known, would be probably dismissed as risible, since like so many other Jewish scholars, he publishes there. I can live with a general deprecation ruling, but not one that allows of no exceptions and endorses instantaneous erasures of anything regardless of the authors' global reputation. There is, in such a case, no alternative source to substitute for the CounterPunch material, because these incidents are rarely covered by mainstream sources (and most mainstream sources are written by journalists with, unlike Shulman, no direct knowledge of these incidents). Nishidani (talk) 18:55, 31 December 2021 (UTC)reply

PR asks (in effect) what is the solution? Dedeprecating CP isn't winning at RSN although it has some support. I suggested SPS exception or some sort of autoreview process. DG (and Amigao) want it all gone pronto. No front runner yet, any more ideas? Selfstudier (talk) 19:04, 31 December 2021 (UTC)reply

One can forget predictably partisan scumbags like myself in these deliberations. People with a far better record of impartial and authoritative commentary exist among that significant minority, meaning that my perception that we have created a serious problem with deprecation as warrant for automatic erasure is not solipsistic but shared by saner wikipedians. A a notable number of high quality and irrepeatable sources of encyclopedic value are being junked or trashed (the Daily Mail analogy is sand-in-the-eyes, scores of top-ranking scholars do not contribute to it). Since most commenting here are policy wonks (I'm out of my depth) some collegial solution that addresses the nature of those policy grey zones is required. Nishidani (talk) 21:50, 31 December 2021 (UTC)reply

What the DM close should have done (and what we've failed to do on all such deprecation RFCs since) Yes, but (a) it didn't do that (b) your attempts for literally years now to make it mean that have consistently been rejected. You could mount an RFC to retrospectively make all deprecation RFCs mean that, but it absolutely doesn't mean it now, and going on and on and on as if it really should mean that is just you refusing to accept that broad general consensus has been against you on this point. At this stage, do you think you could craft such a proposal as an RFC with a chance of passing? - David Gerard (talk) 00:12, 1 January 2022 (UTC)reply

Is this about the correct construal of Holy Writ/fundamentalist approaches to the literal sense of the Founding Fathers or about problem-solving (what do we do when a deprecated source regularly publishes articles by scholars/researchers/journalists of distinction. Deprecate or define exceptions more lucidly?) If any issue or crux arises is to be talked down by simply saying 'it is written', then we are behaving like theologians, for whom the answer to anything is already inscribed in traditional sacred texts (policy) and not problem-solving analysts. Nishidani (talk) 09:37, 1 January 2022 (UTC)reply

Based on past deprecation discussions and currently what's in here, an RFC to establish a grandfathering approach for a limited period after a source becomes deprecated would actually have a good chance of passing. But before starting an RFC would be to see what this whole discussion (on the other areas) comes out to be. The broad consensus has been split, not outright against this. --Masem (t) 18:35, 2 January 2022 (UTC)reply

• Two ideas (ideally done together):
  1. Upon deprecation, a bot will change all usages of deprecated sources to include {{deprecated source}} or similar, which functionally acts as a disclaimer and behaves equivalent to {{citation needed}}. However, this has the effect of still leaving the link to the source present, which will mean deprecated source usages are categorised separately and the article can be easily accessed to aid a volunteer who actually wants to 'review/replace' the reference. It might look something like ^{1unreliable source} and adding a category Category:Articles with references to CounterPunch.
  2. Codify in the hypothetical guideline page the 'handling deprecated references' guidance from WP:DAILYMAIL1: Volunteers are encouraged to review them, and remove/replace them as appropriate. Specifically, clarify what "review" and "remove/replace as appropriate" means, and institute a de-facto WP:BEFORE-like requirement as exists for AfD. Something like Existing references to deprecated sources may be handled by any editor. Editors reviewing existing citations should review whether the reference is appropriate in the context of its usage. If it is, it should remain. If it isn't, the volunteer should see if the reference can be replaced with another source. If a source cannot be found, either leave as-is for another volunteer to deal with, or remove the content supported by the ref along with the ref. In the case of a dispute about whether the reference is appropriate in the context of its usage, editors should seek WP:Dispute resolution and discuss on the talk or another appropriate venue, giving consideration to WP:REPUTABLE and considering the criteria individually. (subject to word-smithing) This makes a deprecated reference (and its removal) functionally identical to a {{citation needed}} (and their removal). After all, you wouldn't just remove a {{citation needed}} tag without either adding a reference or removing the underlying content, that'd just be absurd. You would only remove the content and tag in line with the provisions in WP:BURDEN. This offers good guidance to editors, and makes the expectations more transparent for ANI to deal with individual cases of mishandling.
  ProcrastinatingReader (talk) 16:52, 1 January 2022 (UTC)reply

  Support in principle. I would submit for consideration that the "hypothetical guideline page" should amalgamate both unreliable and deprecated under the rubric "questionable" as per V resulting in a substantial simplification of existing procedures. In other words, unreliable sources lose the SPS exception, apparently the only meaningful distinction between the two, unless editors make the argument per the outlined procedure (as potentially wordsmithed).Selfstudier (talk) 22:17, 1 January 2022 (UTC)reply

  That ("amalgamate both unreliable and deprecated under the rubric "questionable") doesn't work.

  First, no source is always unreliable; see the FAQ at the top of WT:V. There is no "SPS exception" for an unreliable source; if the source is unreliable for the material it's meant to be supporting, then being self-published by an expert does not change the fact that it is unreliable for that material.

  Second, "questionable" is its own jargon, and the use of questionable sources is expressly permitted "per V" (also "per RS", where I think you'll find the slightly clearer of the two explanations). We shouldn't conflate "questionable", "unreliable", and "deprecated".

  Third, "deprecated" includes sources that we deem to have altogether too strong of a "reputation for checking the facts" (to make sure that they line up with the official version) and far too much "meaningful editorial oversight" (to make sure that the censors don't object), so they don't really fit the definition of "questionable". Perhaps the solution there is to re-write the definition; it happens that our idea of reliable sources also includes sources that do absolutely no fact-checking at all (peer review is not fact-checking) and sources that have obvious conflicts of interest (e.g., almost every single academic journal article ever published about any pharmaceutical drug during the ~30-year-long span between the initial discovery and the date when the last patent finally expires).

  The more I read about the active disputes, the more I wonder whether the problem lies less in differing definitions of the terms and more in a lack of clarity about editors' behavior. Is it okay to mass-remove sources from articles? Exactly how much WP:BEFORE-style effort do I need to go to? Engage the brain fully, or just enough to make sure that I'm not blanking dailymail.co.uk from the article about Daily Mail itself? Is it better for Wikipedia to have uncited information with no idea of its provenance, or is it better to have a flag on it that says ^{better source needed} so that editors will know where that potentially dubious material came from? WhatamIdoing (talk) 05:57, 2 January 2022 (UTC)reply

  I get where you are coming from but the editor behavior is actually being driven by the definitions because everyone can find a WP:something to support what they are trying to do. The definitions need clarifying at a minimum and personally I still think simplification would be a good idea (although I am not sure if RS RFCs going 1,3,1,3... is an improvement over 1,4,1,4..).Selfstudier (talk) 10:40, 2 January 2022 (UTC)reply

  This also seems to contradict policy, in that it attempts to make special rules to preserve unreliable sources. This is quite apart from the bit where you're proposing something to apply to those sources found to be the worst sources, worse than merely unreliable sources.

Look, think of it this way: if you see a clearly garbage source in an article, can you just remove it? Yes, you can. It's frankly implausible to make it harder to remove the sources that a broad general consensus has found to be garbage than to remove ordinary garbage sources. If your proposal does that, it's ill-thought-through. Deprecated sources are those that are known-bad, and there are no grounds to assume a priori that they are not - David Gerard (talk) 23:44, 1 January 2022 (UTC)reply

The aim isn't really to make it harder. I agree with you that proposals to add novel restrictions on the removal of deprecated sources are untenable, hence why I disagree with Masem's proposal and particularly any kind of 'time limit' proposal. AFAIK there's no novel restriction in either of my ideas. Plus, #2 is exactly how you would handle statements that currently have no reference and just a {{citation needed}} tag (or no tag at all); indeed WP:BURDEN says more or less the same if one reads the full section.

The aim is to make clear the process for removal and that WP:DR should be sought in disputes. The specifics of deprecated sources being not entirely agreed on, it's not immediately clear what grounds editors would have to challenge the removal in WP:DR, but I think other sections in this RfC (namely: 'purpose of deprecation', 'criteria for deprecation', and most importantly 'valid uses of deprecated sources') need to be answered first to determine such grounds. ProcrastinatingReader (talk) 00:27, 2 January 2022 (UTC)reply

So long as "it's deprecated" is not grounds to remove source reinserted and a good faith discussion about any particular source can take place to examine any individual source on its own merits, I dont have a problem with removing sources deprecated. So I agree, valid uses of deprecated sources is the biggest thing to be decided. nableezy - 00:33, 2 January 2022 (UTC)reply

There's also the problem of how to signal to editors who are less than fully informed that it's an acceptable use. You've probably seen this happen before: Someone changes an article; there's a discussion that results in a consensus. Then someone else changes the same thing, and everyone yells at the latest editor for not having magically guessed that there had been a discussion about that exact thing.

As an example, we had a couple of years of edit-warring over whether the WP:LEADIMAGE for Pregnancy needed to have a naked woman in it. Editors (often less-experienced editors) were regularly surprised by the rather WP:GRATUITOUS nudity at the top of the article, so they would remove it, and they were regularly reverted by a couple of guys claiming WP:NOTCENSORED. It took a couple of RFCs to settle the matter, and there is no more edit warring over the images now, but I still wouldn't expect any editor to know that these prior discussions took place (except for the participants in the discussion, and even some of them might have forgotten by now).

If you are just working your way down Special:LinkSearch or using WP:AWB, hitting hundreds or thousands of articles, you can't possibly know that there was a discussion about this particular URL being used in this particular article. WhatamIdoing (talk) 06:13, 2 January 2022 (UTC)reply

I think that generally, in an encyclopedia, editors who are less than fully informed should refrain from editing. nableezy - 22:16, 2 January 2022 (UTC)reply

If inexperienced editors refrain from editing, no editors become experienced and Wikipedia dies. I think the onus should be on those seeking to include deprecated material rather than on those seeking to remove, but it's good to give inexperienced editors as much guidance as possible. BobFromBrockley (talk) 19:02, 3 January 2022 (UTC)reply

If I had said editors not experienced that might make sense. I said editors not informed however. As in if you have no understanding of the topic or the source, maybe dont remove it? nableezy - 15:19, 4 January 2022 (UTC)reply

OK, that I agree with (but extended-confirmed doesn't mean informed, and we don't really have a mechanism for identifying who is informed, except using community consensus to correct un-informed editing). BobFromBrockley (talk) 12:19, 5 January 2022 (UTC)reply

While noting that the process for new inclusions differs somewhat from the process of removing old ones, the onus is of course with those arguing for inclusion but that is the case for every single source not only deprecated sources. Selfstudier (talk) 19:07, 3 January 2022 (UTC)reply

If we decide that editors who are less than fully informed should refrain from editing, then we all need to quit now. Nobody is fully informed about everything, and most of us are not fully informed about anything. WhatamIdoing (talk) 04:05, 8 January 2022 (UTC)reply

The issue I see is that this appears to put the onus on the editors seeking to remove the depreciated source, rather than on those seeking to restore it - that onus has already been met by the depreciation. I think it would be better to state that once a depreciated source has been removed, it should not be restored without consensus. BilledMammal (talk) 03:52, 2 January 2022 (UTC)reply

When I started the RfC I intended for the questions to be answered sequentially. I do partially think it's a bit silly to answer this question without having the foundation gained from the previous ones. To the point, any response I can give to you will be based on my (unwritten) opinions to the previous questions, rather than anything we've established. I did comment on this to try re-focus the discussion on something actionable, but I still think it's more ideal to answer the issues sequentially so we're not talking over each other or starting from different foundations. ProcrastinatingReader (talk) 04:13, 2 January 2022 (UTC)reply

I did have a similar thought when writing that reply, as answering this question without an answer to the others will not resolve the issues that started this discussion, but I am still working through my position on the other questions, and I felt it was important to state my position here, both because we have a clear proposal, and because I see the onus being against those who wish to retain as a fundamental aspect of deprecation that should not be changed even if we clarify other aspects.

To expand on why I see it this way, it is because we need a practical method to address broad historic use of inappropriate sources, and to switch the onus would stop deprecation being that practical method. BilledMammal (talk) 04:28, 2 January 2022 (UTC)reply

@BilledMammal, since you are, compared to most of this in the discussion right now, a relatively new editor, I think it might be informative for the rest of us if you explained what you mean by "inappropriate sources". Most of us remember when the idea of a list like Wikipedia:Reliable sources/Perennial sources was laughed at as pointless, because reliability depended so much on the specific material in the article that you were supporting. But I wonder if you've "grown up", as it were, with a model in which the article content is less relevant than something else (the publisher's reputation, maybe?). What do you think makes a source "inappropriate"? WhatamIdoing (talk) 06:22, 2 January 2022 (UTC)reply

Nothing more than that "inappropriate sources" are those whose publisher has been deprecated by consensus. What should be deprecated (and thus made inappropriate) is a question I am struggling to answer, but I have some thoughts that I'll go and post in the relevant section. BilledMammal (talk) 06:54, 2 January 2022 (UTC)reply

Unfortunately, I still don't have a definition of "inappropriate sources" or "Generally unreliable" sources, but I am not sure we need to produce one here; the consensus process of determining whether a source is "Generally unreliable" appears to be functioning fine, with the issues being related to the lack of ability to distinguish between sources, and because rather than including nuance in the discussion and summary, we appear to use the easy method of deciding whether we think a source is "bad" or "really bad" and classify the first as "Generally unreliable" and the second as "Deprecated". BilledMammal (talk) 15:46, 2 January 2022 (UTC)reply

@BilledMammal, your description of "the easy method" fits with my impression, too. The RSP process results in far more emphasis on picking a red–yellow–green color than on producing nuanced understandings. Sometimes nuance isn't necessary – as an example of this, I remember someone copy-pasting the contents of an AIDS denialist website into articles; it was both factually wrong and a copyvio, neither of which require careful consideration of details – but for most news media, nuance should be the main goal. WhatamIdoing (talk) 05:34, 8 January 2022 (UTC)reply

I said something similar a long ago, I think of the first times this blew up. In general, it's fine to remove a detail sourced to a deprecated source if an editor feels this is justified, just as you can with anything unsourced. I think doing it en-masse is likely to be seen as a problem just as deciding you're going to remove anything tagged with CN throughout Wikipedia but that's mostly a separate issue which frankly I don't care about. But it's a major mistake to remove deprecated sources and replace them with CN. It's making things worse not better. While deprecated sources may be completely unreliable, they still help provide details such as a rough time frame or maybe other related details in the deprecated source which may help find a reliable source.

As a far better alternative, personally I suggested the better souce needed tag back then but if editors feel this isn't clear enough then we can make a new tag e.g. that suggested by ProcrastinatingReader. As I think I also suggested, if editors really want, we can even make the tag hide the source so it doesn't appear to readers. Alternatively editors could put the source into a hidden comment although this isn't ideal since someone trying to source the detail may look for a source before editing. Whatever there are surely so many solutions that don't involve damaging Wikipedia by making it harder for editors to improve it, which is what we're doing when we remove a deprecated source and replace it with an CN tag.

P.S. Why do people keep saying that we replace unreliable sources with the CN tag? There's nothing in WP:RS that supports this practice and just like replacing deprecated sources with a CN tag, it's a bad idea unless you certain the source provides no useful information. Sometimes where the source is so bad, it may be better to remove it and bring it to the talk page but frankly this is rare, mostly it's better just to tag it. The exception would be where the source itself is blacklist or harmful e.g. may have malware.

P.P.S. I admit, I've probably personally removed some forums and other non RS and replaced them with CN well generally the fact redirect, in the past. I don't think recently but I'm not sure. Most of the time, I shouldn't have done so. I do appreciate in some cases it's questionable if the source provides any useful information for finding another source that isn't obvious from what we say in our article so I acknowlede in some cases we're probably not making things worse, however as this requires careful analysis and you could easily be wrong, it just seems like a bad idea in general when there are much better options available.

Nil Einne (talk) 09:07, 3 January 2022 (UTC)reply

There are multiple possible actions - replacement of the bad cite, remove the claim and the bad cite, remove the bad cite but not the claim - and advocates who will insist that any of these must be required. But every case is unique - David Gerard (talk) 15:13, 3 January 2022 (UTC)reply

The fact that every case is unique is why removal of deprecated sources should be human-based and not done by a bot or made through bot-like, rapid fire actions at least without prior consensus that supposed the use of bot or bot-like actions (per WP:MEATBOT) I would figure with my Google-fu that judging if there's another source out there or if the claim has some potential to be valid but needs a better source to affirm would take 2-3 minutes for a given citation, and that's assuming its a topic within the 21st century. The result could still be outright removal, or leaving a CN behind, or replacing the source but there's no way I can do that in less than a minute per cite. Now, as I proposed, after 6 months from a formal announcement, then we can talk about using high-speed bots or tools to remove or handle such cites without the same human-review care. (Personally, creating a special CN tag that includes in hidden wikitext the original citation, and flagging said page to note its past use of deprecated sources, would be a good step). --Masem (t) 01:31, 4 January 2022 (UTC)reply

Your 2-3 minutes claim is completely made up, and not at all supported by the near-uniform badness of deprecated links - and mainly shows that you complain about deprecated links being removed while having spent negligible time going through the backlogs yourself. Have a look through my edit summaries for the past couple of days - I took a cue from Bobfrombrockley and made each summary much more explanatory. If you can claim with a straight face that this editing is indistinguishable from a bot, then you have the programming skills to radically advance Wikipedia bot editing, and you only need to take care not to tell the bot what a good idea paperclips are - David Gerard (talk) 23:48, 4 January 2022 (UTC)reply

The ones that I see of late are removing opinions from Counterpunch under the claim of UNDUE. That's extremely subjective (RSOPINION can still apply to deprecated sources, but it becomes a matter of discussion if that person's opinion is UNDUE or not, not simply because it was in a deprecated source (it depends on the nature and reasoning for deprecation), and of course, making your own decision on the outcome of this page's very discussion, which is basically not appropriate per WP:FAIT. But even those that are removing deprecated sources to claimed facts, I know that it takes a bit of time to do a skim of results from Google, Google Books, and Google Scholar, and that's at least a 2-3 minute process just to satisfy that a replacement sources is not immediately obvious (eg on this diff 4, it took me about 1.5 minutes to locate this as a replacement source from Deutsche Welle, at least for the first part of the removal). The rate your edits are going and lack of any attempt to find replacements (instead just more interested in removal) is overtly disruptive, particularly given wholly separate issue about possible sock involvement in its deprecation discussion. What you are doing is again against what ArbCom has said not to do in WP:FAIT, regardless if at the end of the day the goal is to remove all uses of deprecated sources. There are far too many open questions on the process that you should absolutely not be continuing in the mode you are in until the issues raised on this page are resolved. --Masem (t) 05:26, 5 January 2022 (UTC)reply

I get your point Masem but the Arafat poisoning example is actually a good instance of the system working isn't it? David Gerard deleted the cite but left the claim, which looks plausible on the face of it, and added a CN tag; another editor then provided a reliable source; overall the encyclopedia was improved by a stronger source being used instead of a poor one. The problem is only when the whole claim is deleted too quickly, but most instances of that happening seem to be defensible (and can be corrected by other editors watching the pages, if we work together collegially). BobFromBrockley (talk) 12:19, 5 January 2022 (UTC)reply

I mean, I am not denying that the ultimate net result (where a deprecated source is replaced with a higher quality source) is right. However, if we go off the DM RFC, the key is that deprecated sources should be replaced or removed, and there is no sign that David is making any good faith effort to find replacement sources, creating the onus on others to fix up when they are just removed (removing material completely or leaving a CN), which is a MEATBOT problem. And this is still within days of Counterpunch, having been previous assumed a reliable source in good faith across numerous articles it seems, being declared at RSN to be deprecated (note: not widely announced to the rest of WP), giving no time for editors on affected pages to try to find replacements first. That's a FAIT problem and is disruptive to the work. Now, if David wanted to run a community effort to stamp out deprecated sources, both announcing it widely (VPP/CENT) and seeking consensus to make sure this was what the community wanted (without any grandfathering period), then hey, that's all good. But David's taken on their own interpretion of the results of the various deprecated RFCs to imply they have to be removed immediately (no such urgency was mentioned in any RFC outside of DM + BLP) and is against the current PAG pages related to deprecated sources (what we're here on this page to resolve), even while the community is trying to resolve that. You can be right as to the end goal, as to eliminate deprecated sources, but the methods of getting there can be against standard practice or community expectations, and that's the issue here is David's practice, not the end goal. --Masem (t) 13:11, 5 January 2022 (UTC)reply

While deprecated sources may be completely unreliable, they still help provide details such as a rough time frame or maybe other related details in the deprecated source which may help find a reliable source. Exactly. I don't see the obsession with removing links. It's not like visiting the website will give you a virus or something. There's the same unreliable content in both scenarios, but in one it's easier to detect the source of unreliability and easier to find a replacement. It hurts my head thinking about how some folks are doing removals currently, because it just seems quite clear to me that this approach doesn't make sense regardless of what your POV is on the source in question. ProcrastinatingReader (talk) 18:49, 3 January 2022 (UTC)reply

There has been at least one unique situation where deprecated links needed to go post-haste - that being Daily Mail for BLP, but that was clearly defined during the DM RFCs leaving no question of rushing through to remove them. But most of the other deprecations RFCs have not discussed speed or urgency for removal. I have to agree with David on that if there's no deadline set, these will linger forever, and we want to promote a means to replace/removed, but unless we have a DM BLP situation, urgency to remove is not established. Hence why I've suggested this grandfathering as to actually create an explicit default timeframe for that purpose. --Masem (t) 02:37, 4 January 2022 (UTC)reply

I agree with @ProcrastinatingReader about the verifiability problems created by removing sources. Content with a bad source (preferably a tagged bad source) behind it is easier to evaluate than content with no source. WhatamIdoing (talk) 05:44, 8 January 2022 (UTC)reply

Definition of "Self-published sources"edit

In cases where the creator and the publisher are the same, a work is always considered self-published. In cases where they are different a work is considered self-published if the publisher has minimal direct or indirect control over the content, such that work is the independent product of the author and can be solely attributed to them, rather than having aspects that were influenced by the publisher. There is a presumption that publishers with an editorial process has more than minimal control over content.

Is this a suitable definition for whether a source is self-published, and if not how can it be improved? For additional information, including the reasoning for the specific wording, see this section, though most of information there, including how it might be integrated in the RSN process, is beyond the scope of this sub-discussion. BilledMammal (talk) 17:22, 2 January 2022 (UTC)reply

While the beginning is very clear, I think that such that work is the independent product of the author and can be solely attributed to them, rather than having aspects that were influenced by the publisher. There is a presumption that publishers with an editorial process has more than minimal control over content. is unnecessary and adds a lot of confusion.

• such that work is the independent product of the author and can be solely attributed to them Isn't that the case for all works, though? Even those that have been reputably published? If I cite something from the New England Journal of Medicine, the work is still the independent product of the authors and can be solely attributed to them. The NEJM is only the publisher.
• rather than having aspects that were influenced by the publisher This very absolute phrasing contrasts with "minimal" at the beginning. If you have minimal editorial oversight (limited, for instance, to copy-editing) some "aspects" are "affected", but the work may still be classified as a self-published source.
• There is a presumption that publishers with an editorial process has more than minimal control over content. This is not a very useful presumption, because it shifts the initial problem ("minimal control") to the question of what constitutes "editorial process". It's also a non-sequitur because "editorial process" is by definition a more substantial review than "minimal control". So we don't really gain anything by adding a new concept. I think that the concept of "minimal control" is sufficient. JBchrch talk 00:20, 4 January 2022 (UTC)reply

In a prior separate discussion related to SPS, I've suggested that when we talk SPS, we're distinguishing SPS from sources where there is an entity involved that edits, reviews, validates, and greenlights publishing of the work (either they publish directly, or give the go-ahead to the author), and subsequently handles post-publishing matters (redaction, etc.). This is more than just a having a simple copyedit review before the user publishes (a mechanism that Forbes Contributors use as best we know). This was intended to be clear that an SPS is not necessary a work published by the creator on a site owned by that creator - eg this expands SPS to include YouTube, Forbes Contributors, and Medium, as well as op-eds, short correspondence in most peer-review journals, and open-source or predatory journals. Whereas, with a peer-reviewed journal, there's the whole peer-review process and board of editors that provide the editorial oversight, equivalent to the editorial desk at NYTimes. There was some agreement but the discussion (wasn't a formal RFC) never got to a conclusion. --Masem (t) 02:32, 4 January 2022 (UTC)reply

@BilledMammal, I think your definition might be confusing "platform" or "printer" with "publisher". The publisher is the person/organization that decides whether this source is going to be made available to the public.

The source doesn't need to have "aspects that were influenced by the publisher" to have been properly published. Similarly, merely having aspects that were influenced doesn't mean that it wasn't self-published. You can hire a designer and an editor from a vanity press. They'll definitely "influence aspects", if you get what you paid for. But you're still the one who decided that it was going to be published, and that makes it self-published, no matter how many others influence your source. WhatamIdoing (talk) 05:50, 8 January 2022 (UTC)reply

Discussionedit

This is related to the RFC on CounterPunch and honestly this is the heart of what I was trying to get at with it. We are somehow treating "editorial control" as a detriment compared to actually self-published sources, and that goes against the entire hierarchy of sources we have here. Unless, and only unless, there is some evidence that the "editorial control" involves misrepresenting what a given person has written, then a self-published source is the very lowest rung in our source quality hierarchy. And every exception to self-published sources should exist to any other source, absent that evidence of misrepresentation in which you cannot even trust that the person actually wrote what the source says they wrote. I still do not understand how the counterargument makes any sense. That if somebody had posted the article they put up on a deprecated source on their personal blog it would somehow be a more acceptable source? Seriously, can anybody explain the logic of this, assuming there is no argument that the deprecated source has ever modified a piece submitted by the author? nableezy - 00:38, 4 January 2022 (UTC)reply

The basis that I am working from is the notion that if this is not the same article that would have been published without the influence of the publisher, then it is not self-published, with the exception of copy-editing changes. The reason for this is that the publisher, even if they don't misrepresent the work, has an influence on it, and if the publisher is not reliable then that means the work is not. For instance, an article published in History will be heavily influenced by History's editorial preferences, and so even if the author would be acceptable as an SPS if they published on Medium, they are not acceptable as an SPS if they published on History.

To meet this, and taking into account the comments by User:Masem and User:JBchrch, I propose the following altered text for consideration: In cases where the creator and the publisher are the same, a work is always considered self-published. In cases where they are different a work is considered self-published if the publisher has minimal direct control or indirect influence over the content. BilledMammal (talk) 05:17, 4 January 2022 (UTC)reply

That isnt addressing my point. There is no "influence of the publisher" in any source in which they do not actually "edit" the material. And even if that were the case, why would it matter? Why would it matter if I, video games history professor and widely cited scholar on the mythology of Halo, wrote something for a deprecated source or posted the same column on my personal blog? Exactly how is my blog more reliable? Nobody is saying that it is self-published, and I would appreciate if my actual argument was addressed, not one I did not make. Mine is, given that self-published is the lowest rung on the ladder of reliability, every exception made for that applies to any other rung. Given, again, absence of any indication that the source cannot be trusted to accurately reflect the named authors words. Why exactly would my posting a blog article be citeable but my posting in a source, lets call it UpperCut, that has been deprecated not be? nableezy - 10:14, 4 January 2022 (UTC)reply

The reason is because the deprecated source has influence on the work. We know that the SME is suitable for use when writing without that influence, but we can not assume that they are when they are writing with that influence. BilledMammal (talk) 10:44, 4 January 2022 (UTC)reply

What exactly is that based on? That is, to me, completely irrational and illogical. What influence does a publication have on a scholar writing a piece published under their own name? Where is there any evidence of any influence at all. And, crucially, how would that impact the reliability of the scholar. nableezy - 15:29, 4 January 2022 (UTC)reply

The fact that works are not independent of the publisher, per current policy. And the nature of influence can vary, but it does exist; scholar writing an article about the Holodomor for RT is going to be limited in what they can say, and so the reliability of the piece will reflect RT's reliability, rather than the reliability of the scholar. To be clear, the reliability of the scholar isn't influenced, the reliability of the work is. BilledMammal (talk) 01:07, 5 January 2022 (UTC)reply

This argument doesnt make any sense to me. Yes, things are not independent of one another, but you seem to be taking "deprecated publisher" as a negative as a opposed to "self-published". But that is not true if the pieces are not being modified, if the author is putting their name to a work it is their reliability at issue. The different parts that affect reliability, author, publisher, the work itself, those are cumulative. Think of it as a math problem, with each part adding to the other. And the lowest value for publisher, 0, is self-published. What WP:RS says is Reliable sources may be published materials with a reliable publication process, authors who are regarded as authoritative in relation to the subject, or both. These qualifications should be demonstrable to other people. That or at the end there, means the sentence should be read as that either the author or publisher being reliable may make a source reliable, or, obviously ideally, both. Deprecated should mean that the value of the publisher is 0, and in cases like the Daily Mail where the publisher is responsible for the actual output more so than any named author, that may render a source closer to the blacklist. But for things like opinion pieces or works by experts, that just makes them equivalent to being self-published. nableezy - 02:57, 5 January 2022 (UTC)reply

I disagree that the lowest value is zero; I believe the lowest value is negative. I think we can agree on this in the case of publications like the Daily Mail, where we know that works are directly modified, but I also believe this is true in cases where the work is not directly modified, as might be the case in my Holodomor example above - but I don't think we are going to agree here. BilledMammal (talk) 03:12, 5 January 2022 (UTC)reply

I think Ive been clear where there is evidence of manipulation everything else gets tossed out. But see for example how WP:RS treats predatory journals: The lack of reliable peer review implies that articles in such journals should be treated similarly to self-published sources. It is saying that when you have no faith in the publisher, treat it as though it is self-published and evaluate the author. nableezy - 03:13, 5 January 2022 (UTC)reply

I see it as saying that when the publisher doesn't do anything except publish (when the publisher has minimal direct control or indirect influence over the content) we can treat them as a self-published source. This typically applies to predatory journals, but not to RT.

I would also note in practical terms that I don't believe a proposal along the lines of what you are advocating has a snowballs chance of passing; even the more restricted version that I am advocating has opposition on the grounds that it is not restricted enough. BilledMammal (talk) 03:26, 5 January 2022 (UTC)reply

Take that a little further, how would we determine the amount and nature of this influence? Selfstudier (talk) 11:01, 4 January 2022 (UTC)reply

RT has enough influence that they are not a SPS; Medium is the opposite. Drawing a line between the two is difficult, but we will need to do that, and I believe the line above is a reasonable way to do it. The nature of the influence is simply our current assessment of the sources reliability. BilledMammal (talk) 11:59, 4 January 2022 (UTC)reply

Oppose - Self-published means self published. A work is self published if the author is the publisher, anything else doesn't make sense. It's not up to us to judge how "minimal" or not the editorial process is. If someone publishes an article, a chapter or an essay in a medium that s more than a blog, then it's no longer self-published, even more so if the publisher chooses which works to publish or not. --Mvbaron (talk) 13:10, 4 January 2022 (UTC)reply

I thought someone would say that :) Which takes us right back around to the CP rfc, "treat as" SPS which can only mean that it is not SPS but we still want to treat it as if it were in order to keep access to expert opinions. Of course, if the principle were accepted for CP, why would it not be for all others. Selfstudier (talk) 13:26, 4 January 2022 (UTC)reply

My problem with that definition is when I imagine a hypothetical publisher that selectively publishes unmodified submissions - I can't see how this isn't usable as a self-published source. BilledMammal (talk) 14:28, 4 January 2022 (UTC)reply

You don't have to imagine a hypothetical publisher that selectively publishes unmodified submissions. This really does happen. The New Yorker was famous for it. You will find that magazine listed at RSP as reliable, with a note about its "robust fact-checking process". Many old-school magazines accept unsolicited submissions, pick the best of the bunch, and run them with no more than copyediting. They don't need to do more, because they are already looking at finished work, and there is so much good content available that if yours isn't good enough, they can pick something else. The publisher's role in this model isn't to influence what you write. The publisher's role is to pick and choose the things that they want to present to the world. WhatamIdoing (talk) 06:04, 8 January 2022 (UTC)reply

That's a good example, but I can't think of how to address it in our policies - and we need to. BilledMammal (talk) 03:46, 12 January 2022 (UTC)reply

Question I think we all agree that if a person publishes a thing on their personal website that is self published. What if Honda publishes a claim on their website? I'm going to assume that Honda, like many large companies, has some sort of PR group that reviews content before publishing. Is that self published? A person within Honda writes a story about the new Honda Civic. It's almost certainly going to be reviewed before publication so does that make Honda a RS publisher? How should we handle publications that, while certainly edited by someone, are not independently edited or have a clear vested interest in the content of the material they publish? Note, this question is not specifically related to Deprecation but since the question is being asked. Springee (talk) 14:09, 4 January 2022 (UTC)reply

Supposedly that falls under questionable source (per WhatamIdoing, promotional). Then gunrel and dep are carve outs from questionable with their own specific characteristics, although I remain unconvinced of an essential difference between these two. I could understand "fabrication" as a differentiator but the list is presently extended to include all sorts of other factors besides. Selfstudier (talk) 14:41, 4 January 2022 (UTC)reply

I think this question highlights that material that originates from an organization may not have a single author, and doesn't operate in the same manner as traditional publishers. I suppose the question comes down to can an organization be an "author"? And if so, wouldn't that make material authored by an organization and published by that same organization be self-published? --Kyohyi (talk) 17:04, 5 January 2022 (UTC)reply

An organization can be an author (both in law, and in practice), and whatever a corporation writes for itself and makes available to the world itself is self-published. Self-published sources can be reliable. Honda is assumed to be a reliable source for what Honda says about itself (including about its employees, who are not considered "third parties" within the meaning of BLPSPS). WhatamIdoing (talk) 06:06, 8 January 2022 (UTC)reply

Historyedit

• First use: The initial remarks of this page say Daily Mail was "the first source to be 'deprecated'". In fact in Daily Mail RfC (closed on 2017-02-08) the word "deprecate" only appears in a suggestion that some anti-Daily-Mail !votes should be deprecated. On 2017-11-17 the "Deprecated publisher" template was created. On 2018-02-08 an RfC re Fox News mentioned deprecating, but it wasn't. On 2018-08-30 InfoWars RfC was closed with no mention of "deprecate". On 2018-09-25 Breitbart RfC was closed, which appears to be the first time the word is used by an RfC closer, so it would be correct to say Breitbart was "the first source to be 'deprecated'". On 2018-12-17 the WP:DEPS essay was created which incorrectly said Daily Mail was "the first source to be deprecated" (later the word "formally" was added but that doesn't save it). On 2020-01-09 Omer Benjakob in Haaretz said Daily Mail was the first, which unfortunately is the source for the cite in the Daily Mail article. However, the WP:DEPS statement that the Daily Mail RfC was a landmark decision is not obviously false. Peter Gulutzan (talk) 19:49, 27 December 2021 (UTC)reply

  Wasn't Daily Mail still the first to be 'deprecated' (as in, given the treatment we now describe as 'deprecation', even though this was not the term used at the time)? It seems in the Breitbart RfC, Guy/JzG just gave what happened to Daily Mail a name ("deprecate") and it stuck, but that RfC didn't define the concept AFAICS; WP:DAILYMAIL1 and WP:DAILYMAIL2 did. ProcrastinatingReader (talk) 23:29, 29 December 2021 (UTC)reply

  It's justifiable sometimes to apply a term that wasn't used at the time, like a Proleptic Gregorian calendar, provided we accept first that it means exactly the same thing that was meant at the time. But isn't that assuming, in advance, that the definition is strictly what WP:DAILYMAIL1 closers said, so we're done here? Peter Gulutzan (talk) 00:46, 30 December 2021 (UTC)reply

  It is by definition the same thing, isn’t it? The opening statement wrote: Should Breitbart be deprecated as a source in the same was as WP:DAILYMAIL which defines the meaning of what he meant. That RfC might’ve coined the word, but it didn’t define any of the concept. ProcrastinatingReader (talk) 02:57, 30 December 2021 (UTC)reply

  Actually "deprecate" was coined long before Guy|JzG, but even A new meaning of 'deprecate' doesn't resemble what WP:DAILYMAIL1 closers said, so it's not what anyone could have meant then using a dictionary. Instead it's been called a "term of art". Sunrise called it that in June 2019 but I might have missed an earlier use. Peter Gulutzan (talk) 15:32, 30 December 2021 (UTC) Additional note: at the time of that Guy|JzG quote, which was before the decision to redirect, WP:DAILYMAIL pointed to the Daily Mail RfC. Peter Gulutzan (talk) 17:42, 30 December 2021 (UTC)reply

DAILYMAIL1 doesn't specifically use the term "deprecate", but it set the initial precedents for what would later come to be called deprecation. This includes the precedent that consensus found it possible for a source to be "generally prohibited". However, the precise meaning of "generally prohibited" was left open to discussion, and it was the discussions over the following years that established the consensus towards a strict interpretation. That was the same period that the concept was applied to other sources as well. I don't know if I was the first person to specifically call it a "term of art", but I think it was clear by that point that the word "deprecate" had taken on WP-specific meanings that did not fully match the dictionary definition. For instance, this is the revision of DEPS from the time that I made that comment. Sunrise (talk) 00:07, 31 December 2021 (UTC)reply

• Exception for old articles: The Daily Mail closers said it "may have been more reliable historically" and two of them participated without objecting when an exception was accepted for a 1940 Daily Mail use. A closer of a June 2020 RfC accepted the claim that "Daily Mail published a faked version of its own front page" but only changed WP:DEPSWP:RSP. Peter Gulutzan (talk) 18:27, 29 December 2021 (UTC)reply
  • It also changed WP:RSP, and was in fact to do so - David Gerard (talk) 22:52, 29 December 2021 (UTC)reply
    • Indeed it was WP:RSP, I made a correction.Peter Gulutzan (talk) 23:27, 29 December 2021 (UTC)reply
• Edit filter for all: The Daily Mail closers suggested an edit filter to warn about use "as a reference". One of them said in a follow-up thread "Note that in the close of the Daily Mail RFC, we intended the edit filter to warn editors, but not disallow edits. This was a very intentional decision, and I hope the wording reflects that clearly." However, in April 2019 discussion and an RfC determined adjusting the edit filter wording to reflect results of each RfC separately would be difficult so Edit Filter 869 can't mention all possible exceptions or extra restrictions. The current edit filter text is here, most was written by Guy|JzG in August 2019 and May 2020. National Enquirer was excluded. Peter Gulutzan (talk) 21:15, 30 December 2021 (UTC)reply
• Exception for opinions and/or aboutself: The February 2017 Daily Mail RfC close said "it could make sense to cite it as a primary source if it is the subject of discussion", which could be interpreted as: WP:ABOUTSELF applies. The idea that WP:ABOUTSELF is the only important exception -- so not opinions in general -- seems to have first appeared on WP:RSN in November 2018 in an RfC re World Net Daily, the idea was rejected with the closer saying that opinions in general are allowed but this might have been superseded when World Net Daily was deprecated. In July 2019 the only-if-aboutself idea came up again and was rejected by a Daily Mail closer saying (as other Daily Mail closers had already said) that non-about-self opinions are acceptable too though one might wonder why. In December 2021 the only-if-aboutself idea came up again and was rejected by the Breitbart closer, who said that the closing remarks themselves were intended to be clear that opinions in general are allowed. For some other sites (Occupy Democrats, Crunchbase, Epoch Times) the closer remarks suggest some opinion would be allowed, and about 18 say something like "as in the 2017 RfC of the Daily Mail" without being explicit about that corollary. However, for at least four sites (VDARE, Taki's, FrontPage, bestgore), and some fake news sites, closer remarks seem stricter. Peter Gulutzan (talk) 21:59, 31 December 2021 (UTC)reply

  That's interesting. It's not as clear cut as it might be. Would you say that there is a trend to a stricter interpretation over time? Selfstudier (talk) 09:46, 1 January 2022 (UTC)reply

  I can't say that because the sample size is small and might be skewed (I don't know how many RfCs failed). Peter Gulutzan (talk) 18:31, 1 January 2022 (UTC)reply

  I believe that @Iridescent has a story to tell about citing the Daily Mail as a primary source about itself, and having someone declare that the February 2017 Daily Mail RfC prohibited editors from citing it.

  Self, the general rule is that all rules get applied mindlessly. WP:Policy writing is hard. You can write a policy to defend the goal against wikilawyers, POV pushers, people with an axe to grind, etc. (e.g., by deploying a lot of WP:BRADSPEAK), but what really helps is if editors routinely resist overly aggressive interpretation of "the rules" in their everyday editing. WhatamIdoing (talk) 06:31, 2 January 2022 (UTC)reply

  It was here; even though it was sourced to an impeccable academic source specifically discussing the Mail story, someone tried to claim that the RFC meant we could neither mention the existence of the Mail story, nor link to it to allow readers to see for themselves what the story in question had been. ‑ Iridescent 08:51, 2 January 2022 (UTC)reply

  Erm, that’s correct. The removed text does not fall under ABOUTSELF… Mvbaron (talk) 09:15, 2 January 2022 (UTC)reply

  Earlier Iridescent refuted a different someone, me. Peter Gulutzan (talk) 15:17, 2 January 2022 (UTC)reply

  ) Selfstudier (talk) 15:23, 2 January 2022 (UTC)reply

  +1 that’s a very good analysis! It seems that from the original daily mail closers at least 3/5 allowed for opinion pieces, and subsequent closers sometimes allow it and sometimes are stricter. What this means, I guess, is that there's a default assumption that deprecation allows for more than ABOUTSELF, but that default can be overridden by consensus. I wonder what the point of deprecation is then - it doesn't seem to entail a stronger prohibition than that for generally unreliable sources. The straightforward policy-RFC question would be "Are opinion pieces/pieces by established subject matter experts in deprecated sources allowed"? Mvbaron (talk) 10:22, 1 January 2022 (UTC)reply

  That would be the generalization of the Counterpoint RFC running at the moment. Idk whether it is better to run the general question or press on with the attempt to make this guideline which covers a bit more than just the SPS question as well as their being no consensus in here on that as yet.Selfstudier (talk) 12:55, 1 January 2022 (UTC)reply

  not sure it is, the CP RFC is (a) about counterpunch and (b) at least ambiguously phrased and about self-publishing. In my opinion, we definitely need the general RFC. And I thought that we need multiple RFCs on the guideline here anyway. Mvbaron (talk) 13:09, 1 January 2022 (UTC)reply

  If there is enough consensus to draw up a draft guideline, then all the points can be addressed in one RFC on the guideline itself. There does seem to be a consensus that we need a guideline but not what it should say.Selfstudier (talk) 13:27, 1 January 2022 (UTC)reply

  One massive RFC may not be the most effective way to find agreement. WhatamIdoing (talk) 06:32, 2 January 2022 (UTC)reply

  That's perhaps true, it's also true that we are not in agreement here either just in terms of even drafting such an RFC so the alternative? Given that there are currently running open RFCs addressing lesser matters equally without agreement (not to mention the ANI discussion).Selfstudier (talk) 10:25, 2 January 2022 (UTC)reply

  Just as a practical example of whether RSOPINION should allow for deprecated sources: in the UK, many of the top recognized film/television critics are those that work for the tabloids of the UK, including the DM, Mirror, etc. I've postulated that while we'd not use these critics for films produced outside the UK or that generally have an international showing, it would be odd not to use them for their RSOPINION on Brit-made works, as long as consensus agrees that individual critics here are recognized as those experts. I havent followed up on this much as most of these DM reviews were used on TV episodes and wthin the TV wikiproject, they are backing off on articles where there are "routine" reviews and little else to go beyond that as requiring standalone, so there is less a need to get these as RSOPINION, but I am certain there are other cases out there that involve deprecated sources and authors of opinion pieces that are considered appropriate experts to include without being UNUDE on a topic (eg this seems to be an issue with Counterpunch). --Masem (t) 16:16, 6 January 2022 (UTC)reply

I think that is the core issue of recent worries. Of 70,000 articles, we are talking about the retention of 200 or so written by top scholars in their fields, who unfortunately do not perceive that venue in the way a majority of the 30 odd editors who voted for option 4 did. I can handle deprecation, even though I think that the deprecation process was wildly subjective in skewing that webzine as toxically anti-Semitic, conspiracy-mongering, anti-Israeli, genocide-promoting, and Holocaust-denying on the basis of a handful of mostly unexamined diffs. What worries me is automatism in any form of decision-making - 'it is written ergo . .' is not what modernity is about. That we are questioning the use of Norman Finkelstein or Michael Neumann, who are both sons of Jews who either survived the Holocaust or were tipped off to flee before they were nabbed, as source for its greatest historian, whom they knew, i.e. Raul Hilberg, is witheringly uncommonsensical, and extremely doctrinaire in its insistence that there can be no exceptions. Fear, taboos, rote judgment from precedent should never trump the exercise of careful context-embedded analysis in writing encyclopedic material. To the contrary, across-the-board application of an 'exclude-on-sight-rule' is dangerous bait for our laziness. Your point about British film critics is also well-taken. Deprecation is fine, as long as it is clear that we don't write off eminent journalists and scholars from having their opinions known on this encyclopedia simply because they don't share our rather random decisions about what is reliable or not.Nishidani (talk) 18:27, 6 January 2022 (UTC)reply

+1 about keeping scholars who write for the public. That's what scholars should do.

I am not a fan of the idea that if you work for the Daily Mail, then everything you write, even for more reputable publishers, should be banned. I think this is an especially silly idea for pop culture reviews. We don't expect any fact-checking in a film review (or any other opinion piece). A fact-checker just can't do anything with a sentence like "This film has one of the most thrilling car chase scenes since The French Connection in 1971 – five stars!" WhatamIdoing (talk) 06:17, 8 January 2022 (UTC)reply

• WP:DEPS says that deprecation is just 'general unreliability + edit filter + autorevert', which doesn't seem correct. Consider the entry for National Enquirer at WP:RSP: In the 2019 RfC, there was weak consensus to deprecate the National Enquirer as a source, but no consensus to create an edit filter to warn editors against using the publication. It seems current 'deprecation' is just 'super unreliability' (in intent), although in practice the quality of discussion varies so much, particularly in fact-finding, that I don't think the discussions even conclusively establish 'super unreliability'. ProcrastinatingReader (talk) 03:33, 5 January 2022 (UTC)reply

I'm not even sure how we distinguish between "unreliable" and "super unreliable"; I think I've said this elsewhere, but I worry that it only serves to make users lazy and results in those who believe it can be used for certain topics !voting "unreliable", rather than specifying what those topics are so that they can be discussed and recorded as exemptions. BilledMammal (talk) 03:37, 5 January 2022 (UTC)reply

Unstructured discussion and unintended proposalsedit

Removed as too long and not helpful BilledMammal (talk) 03:28, 12 January 2022 (UTC) BilledMammal (talk) 15:40, 2 January 2022 (UTC)reply

What is this though? A continuation of this RFC? An RFCbefore a new one? A basis for a proposal? What do you expect editors to do as regards the above, that's what I am asking? Selfstudier (talk) 15:58, 2 January 2022 (UTC)reply

If I understood your comments elsewhere correctly, this discussion is an attempt to generate a draft proposal for an RFC, and I suppose it is an attempt to move forwards with that by providing specifics as a starting point - I suppose what I expect editors to do is see if the method of considering the issues that RSP is intended to address helps them understand what they believe needs to be done, and from there possibly engage with the ideas that I found to be a natural consequence of considering the issue in that light. If you believe it is premature or otherwise inappropriate, feel free to collapse. BilledMammal (talk) 16:17, 2 January 2022 (UTC)reply

Idk whether the proposal is good or not, I read it but haven't considered it in detail. It's a lot to take on board in one go. Does it all have to be considered at once or can it be broken into pieces. I ask because you can see from the other discussion that there has not been agreement even on matters that can be straightforwardly expressed in a sentence or two. For instance, your definition of a self published source is one such, why not see if you can get some agreement on that part? Or the part about deprecation being time limited. Or some other part.Selfstudier (talk) 16:29, 2 January 2022 (UTC)reply

While each aspect is intended to complement other parts of the change, most of the proposals can function independently. I'll look at breaking them out, starting with the SPS one. BilledMammal (talk) 17:08, 2 January 2022 (UTC)reply

Here is a concrete current question Diff. So CP is deprecated and it has to go, right? Selfstudier (talk) 16:35, 2 January 2022 (UTC)reply

Under what is proposed above, yes. It will not immediately alter the status of CounterPunch or any other source, as I believe any such alteration would derail the discussion - because of this I am also uncertain about the inclusion of the expiry date, as while the functional change to the status will be minimal it may appear to be significant and that alone may be enough to cause issues.

What it will do is open a door for CounterPunch to be considered a SPS, if it can be demonstrated that the influence of CounterPunch on the content of the works published by them is minimal. BilledMammal (talk) 17:08, 2 January 2022 (UTC)reply

And it's gone Diff Selfstudier (talk) 17:47, 2 January 2022 (UTC) Back again:) Selfstudier (talk) 18:46, 2 January 2022 (UTC)reply

Ah, I made the same mistake overlooking the "Beyond Chutzpah" source; with that it is appropriate to keep, both under the above proposal and the current situation, as the only statement sourced to CounterPunch is an WP:ABOUTSELF. What I wrote above remains true in general though. BilledMammal (talk) 00:52, 3 January 2022 (UTC)reply

Reopening the closededit

The discussion that prompted the creation of this discussion, mentioned at the top here, Wikipedia:Administrators' noticeboard/Incidents#Can we decide what the heck "deprecation" means, or alternately, use a different word?, was closed on 26 December last by Ymblanter with comment

This might be a useful discussion, but it should occur either at Wikipedia talk:Reliable sources or at Wikipedia:Village pump (policy). I am closing it here and suggest that whoever wants to proceed to move it to a more appropriate venue.

Here it is, straight copy paste, I propose we continue this discussion (and the "voting" as well).Selfstudier (talk) 15:50, 3 January 2022 (UTC)reply

Should anyone want to reply to anything in here, then make sure to ping whoever made the comment you are replying to (outside of the archive).Selfstudier (talk) 22:54, 3 January 2022 (UTC)reply

@Selfstudier: I think the close was trying to say the point of the discussion was useful to continue, not the specific text itself. The ideas were incorporated into the questions above (e.g. JPxG's criticism of the terminology "deprecation"). The discussion below is too large and unfocused, however. I suggest deleting this section; the discussion is already getting a bit unfocused and cluttered, and focus is necessary to get to any kind of outcome and not just end up going round in circles chasing our tails. ProcrastinatingReader (talk) 17:29, 3 January 2022 (UTC)reply

That's why I bought it over, we are already chasing our tails, at least there was some signs of agreement in this discussion, especially early on, more than there is presently in the other. The closer thought it was potentially useful, let's see if anyone wants to continue, if not then we can close it, with the other in all likelihood following on behind I'm sad to say. Selfstudier (talk) 17:41, 3 January 2022 (UTC)reply

There is one thing below that caught my eye reading the below through again, the idea "....may be as simple as formally adopting Wikipedia:Deprecated sources". Maybe that's the way to go to get a more focused discussion.Selfstudier (talk) 17:46, 3 January 2022 (UTC)reply

Wikipedia:Deprecated sources is largely procedural (and literally an 'information page') and doesn't provide guidance on any of the large issues IMO (some of the outstanding questions above). It's not a solution to these problems. ProcrastinatingReader (talk) 18:02, 3 January 2022 (UTC)reply

I know it isn't a solution, if it was we wouldn't need to do any of this, would we? I am trying to address your concern over focus. Perhaps think of it as a clotheshorse on which we might proceed to hang some clothes. I know the ..writing policy is a hard thing.. but I wouldn't even dream of trying to write a guideline following the above discussion, it would be a complete waste of time.Selfstudier (talk) 18:10, 3 January 2022 (UTC)reply

The discussion below is a part of the ANI discussion Wikipedia:Administrators'_noticeboard/Incidents#User:David Gerard violating Wikipedia:Deprecated sources#Acceptable uses of deprecated sources closed on 4 January 2021 referred to at top.