Wikipedia:Village pump (proposals)

• Comment and Oppose. The ability to instant create edits lies at the heart of this project. And I think many editor's very first foray into Wikipedia is through attempts at a main edit. It's our welcome mat. Seeing your change live instantly provides a thrill and confirms to newcomers instantly that their suspicions of how things work ("anybody can edit") were valid. Yes, this includes lots of immature vandal edits. But those tends to be stupid things like profanity inserted very visibly. But even this helps prompt good new edits to try to fix that. In other words, the "wild west" nature of the front page is an important way to engage new users which we need. And even some of the vandals turn into good editors as they mature. Using PC makes your first edits boring and people tend not to stick around for boring things. Jason Quinn (talk) 02:22, 2 March 2021 (UTC)[reply]

• Oppose using PC for TFA, but support just about any other form of protection (semi, extended confirmed, etc). IMO, Pending Changes is an abomination that shouldn't be used at all, on any article, ever. It increases the amount of work needlessly and makes editors responsible for someone else' edits. It is also extremely confusing and difficult to understand as a feature, which for prospective TFA use makes it unacceptable. TFA will be viewed and edited by many inexperienced editors. We should not confuse the hell out of them with a monstrosity like Pending Changes. Just semi-protect TFA instead. Nsk92 (talk) 02:47, 2 March 2021 (UTC)[reply]

• Oppose. It looks as if the wind is blowing strongly in one direction on this one, but my take is that this seems like an overkill solution in search of a problem--which proposed problem I think is substantially minimal by the very nature of the context. If TFAs somewhat gain additional traffice from IPs and neophyte editors, they are also gain increased scrutiny from a broad chunk of the community over the same 24 hour period that they are up on the main page. Additionally, these articles tend to be slightly more robust and reflective of engaged editing than the average article. In short, I can't see that adding pending changes will make all that substantial a difference to protecting the article against erroneous, bad-faith, or otherwise problematic edits over the course of its day of featured status. Meanwhile, it seems absolutely certain to have an impact on what has traditionally been one of the Featured Article's perceived benefits: editor recruitment.

In fact, I would argue that having this process by which new editors are corralled into a different article each day (which article represents decent editorial standards) and that space becoming the first place in which they can appreciate the pleasure of contributing to the encyclopedia and enjoying the immediate feeling of seeing those changes go live, while simultaneously focusing community oversight on that article space in an organic fashion all sounds like precisely the balance we want to establish and why this is a "it's not broke, don't fix it" type of situation. Additionally, TFAs frequently benefit from the super-crowd-sourcing feature of the attention they get: the vast majority of even IP edits are beneficial, not deleterious, so why gum up the works by having a queue of (potentially overlapping) pending edits. And those queues don't always get addressed with alacrity: I'm a pending changes reviewer myself and have logged a fair amount of work in the area over recent years, and I can tell you that the queue being addressed is quite a variable thing with regard to both the subject matter of the article and the time of day. Lastly, this is just not the typical scenario (a n inherently problematic article) in which this form of protection is typically seen as most justified.

In short, and with respect to the proposers and supporters above, the cost-benefit analysis runs strongly counter to the proposal, as I see it. Snow ^{let's rap} 04:28, 2 March 2021 (UTC)[reply]

It's equally likely that a new editor attempting to edit a well established article (the TFA) will mess something up and get a notice or worse a stern scary looking warning on the their talk page. This also ignores that PC still allows the new editors to edit the article, while conveniently removing the incentive for at least some vandalism/LTAs. Ex. (had to dig for it a bit): TFA from May last year - the revdel'd edits are a well known LTA vandal (the same as for most of the articles on Wikipedia:Today's featured article/May 2020... - I know, this discussion is not a new idea); after the PC protection the LTA stopped, while still allowing other non-AC edits (a fair bit were common vandalism, some were good faith but misguided; but anyway PC does not create the same issues as semi-protecting would). While PC might not be effective on some articles, or maybe even on most, on the TFA it would probably be the best of both worlds: not showing vandalism to our readers, while still allowing new editors to intervene. Cheers, RandomCanadian (talk / contribs) 05:06, 2 March 2021 (UTC)[reply]

• Support. Using PC on TFA is not actually a new idea. Myself and other admins have preemptively applied it many times to guard against long-term abuse, or in reaction to vandalism. These were emergency situations that necessitated stepping outside policy and lack of consensus, but each time to my knowledge we didn't see any of the problems others frequently cite when questing use of PC on TFA. Specifically, there are enough eyes and activity on this article that it doesn't have much detriment on the backlog. Shielding one of the most visible articles from defacement, while still allowing new users to contribute, is a win-win and frankly long overdue. — MusikAnimal ^talk 02:09, 3 March 2021 (UTC)[reply]

Pending changes bugs

(Moving discussion out of archive.) Unfortunately, there are a ton of bugs affecting pending changes right now: autoconfirmed users having their edits held back for review erroneously, administrators not being able to set pending changes protection, and apparently the pending changes codebase is completely abandoned (no active developers who understand the code). See also the discussion at Wikipedia:Village pump (technical)#Pending Changes again. I wrote a comment in the discussion above expressing some support for this idea, but if we can't ensure the software reliability of pending changes, I'm afraid I'm now quite hesitant to support expanding it. Mz7 (talk) 20:06, 12 March 2021 (UTC)[reply]

As one of the more active PCRs and one of the original formulators of the proposal, I've been following the PC bug discourse with some interest. I haven't actually observed a ton of it in my work, so I wonder quite how big the problem is -- that is, what proportion of people who get false-positived go complain somewhere. (It's not just on VPs, I've seen it come up a few times at PERM.) If the majority of people who get it bring it up, it's pretty small; if this is just the tip of the iceberg, it might be a serious issue.

In terms of the proposal -- the thing that stands out to me is most of the oppose votes are saying "We shouldn't do this, we should semiprotect instead", not "We shouldn't do this, TFA should be status quo". I wonder how the vote distributions would stand out for the following two proposals:

"TFA goes under either pending changes or no protection, only two options, semiprotection is completely off the table and will never happen"

"TFA either goes under semi, PC, or none; rank the options by preference"

It'd be an interesting RfC either way. Vaticidalprophet (talk) 03:11, 13 March 2021 (UTC)[reply]

I go through PC occasionally (when I'm not otherwise busy) and I don't encounter the bug too often (non-improvements and vandalism are, alas, more frequent), or when I do it's usually not too much of a hassle to just accept the change - if I have a particularly large amount of spare time I'll leave a notice about this to the affected editor. The fact that the code is unmaintained and apparently some form of a mess isn't a good sign, however, and well we'd have to weight whether the risk of further issues developping is less of an annoyance than potentially doing something about TFA LTAs... Cheers, RandomCanadian (talk / contribs) 05:42, 13 March 2021 (UTC)[reply]

Update: I'm aware there's now a BRFA intended to address this issue. If that goes through, then I think I'll withdraw my hesitation. Mz7 (talk) 19:06, 15 March 2021 (UTC)[reply]

I think hotfixes via a bot are untenable long-term. As more and more bugs crop up that can't be fixed in the extension, we can't keep using bots and other hacks to get around them locally. I'd personally like to see FlaggedRevs get picked up in the code stewardship review before expanding its use. ProcrastinatingReader (talk) 23:26, 15 March 2021 (UTC)[reply]

• Why don't we scrap pending changes reviewer? and grant the rights to extended-confirmed? I mean, the point of PC is a basic flag to stop vandalism/obvious DE, right? It's akin to a less extreme semi-protection. So it doesn't really make sense, imo, to put it to a separate right. The right should be bundled into extendedconfirmed and the separate right removed I think (+ EC is a reasonable level of 'trust' against pure nonsense vandalism + they can already edit areas like Israel-Palestine). That would also fix this particular bug. ProcrastinatingReader (talk) 17:17, 16 March 2021 (UTC)[reply]

  ProcrastinatingReader, honestly I'm not opposed. Back in 2018 I floated this idea on the idea lab village pump, and there was some mixed reception, and my interest in pursuing the change kind of just fizzled out. (Looking back I really wrote too much, heh. Concision was not/is not my strong suit.) Mz7 (talk) 19:42, 21 March 2021 (UTC)[reply]

  @Mz7: afaics there's two main objections there:

  1. That people would get a permission enabled without knowing what it does / directly asking for it. But this is true with autoconfirmed too: you get a bunch of buttons automatically and probably won't use many of them. And for adminship, too. You don't need to use a button just because you have access to it.

  2. That editors would game the system / increase edit count to gain it. I don't think this is true for a few reasons. 1) same could be said about the 'privilege' to edit in ARBPIA. 2) same could be said for pages currently autoconfirmed/ECP protected (gaming edits up means that you can now accept {{

  edit protected

  }} except that it's directly supported by the software interface).

  So neither objection really makes sense imo. ProcrastinatingReader (talk) 19:55, 21 March 2021 (UTC)[reply]

  While we're at it why not scrap pending changes altogether? Is there any reason other than

  Phil Bridger (talk) 19:53, 21 March 2021 (UTC)[reply

  ]

Should we delete succession box content such as the following? Examples: "Oldest-living British prime minister". What say all of you? GoodDay (talk) 21:57, 4 February 2021 (UTC)[reply]

Survey (succession boxes)

• Yes. We ought not to have any of that trivia in succession boxes. There are often many boxes, dozens even, and additional clutter is unhelpful. One would be very hard-pressed to find reliable sources discussing how
  Alex Douglas-Home as the oldest British prime minister, and I dare say that no biography of either of them mentions it. Surtsicna (talk) 22:15, 4 February 2021 (UTC)[reply
  ]
• Yes. Only posts with transitions (or successions) would need succession boxes. -- Kautilya3 (talk) 22:56, 4 February 2021 (UTC)[reply]
• Not here. Trivia should be deleted, non-trivia should not be. Whether any specific succession is or is not trivia needs discussing individually at an appropriate forum - that forum is very much not here. Thryduulf (talk) 02:21, 5 February 2021 (UTC)[reply]
• No As Thryduulf says trivial ones should be, non-trivial ones should not be. That needs to be done on a case by case basis. -DJSasso (talk) 18:03, 5 February 2021 (UTC)[reply]
• Yes I find succession boxes unnecessary in general since they usually duplicate the infobox, a navbox, and/or the text. When it's not duplicative like this example, it's often trivial that doesn't need a box to itself. Reywas92^Talk 19:36, 5 February 2021 (UTC)[reply]
  • You
    dislike them, whereas I find the clear, consistent presentation extremely useful. Succession boxes, infoboxes, navboxes and prose all serve different functions and so the same link appearing in more than one place is a Good Thing. Some succession boxes should be removed, but most should not be. Which is the case for any given succession box can only be determined by a consensus discussion about the succession box in question. Thryduulf (talk) 01:04, 6 February 2021 (UTC)[reply
    ]
• Yes and would support getting rid of them all. Redundant and pointless clutter. Why not have separate boxes for marriages, spouses, and works dispersed througout the article while we're at it? ~ HAL333 02:50, 6 February 2021 (UTC)[reply]
  • What are they redundant to? Links in prose and links in succession boxes have different purposes so they are not redundant to each other. Boxes (for anything) scattered throughout the prose would be completely disruptive to the prose, which is why succession boxes are not placed in the middle of the prose? Thryduulf (talk) 11:58, 6 February 2021 (UTC)[reply]
• Yes - Not sure we even need them for positions, but definitely not for superlatives. Levivich ^harass/_hound 17:22, 6 February 2021 (UTC)[reply]
  • All superlatives? What about tallest buildings and longest bridges? Those seem extremely useful succession boxes to me. Thryduulf (talk) 18:24, 6 February 2021 (UTC)[reply]
• Comment: I feel the validity or triviality of succession boxes should be discussed on a case by case basis on their respective talk page. Not sure what a global decision one way or another on "trivial" succession boxes could achieve, when it does nothing to establish what is trivial and what is not. PraiseVivec (talk) 14:01, 7 February 2021 (UTC)[reply]
• Yes for purely trivia succession box such as "oldest living X" - unless said role is notable in its own right (not sure if this actually applies anywhere).
  talk | contribs) 05:54, 14 February 2021 (UTC)[reply
  ]
• Yes certain ones should be deleted. At a minimum, the spirit of
  WP:NAVBOX #4 seems applicable: There should be a Wikipedia article on the subject of the succession box.—Bagumba (talk) 10:11, 22 February 2021 (UTC)[reply
  ]
• Depends - Yes for "Oldest-living British prime minister" "successions" as trivia, but without a discussion on others is will depend on the case, this should not be used as consensus to delete non-cited examples. KylieTastic (talk) 10:15, 23 February 2021 (UTC)[reply]
• Yes Succession boxes are not needed, in most cases it's a duplicate.Sea Ane (talk) 21:43, 26 February 2021 (UTC)[reply]
• Oppose deciding this specific one here, support a general rethink of succession/navboxes and what they are good for (if anything). Most of the succession boxes at John Major are less trivial, but there are so many of them that they are not a useful way to navigate anything, and are hidden away by default (just like the ungodly amount of navboxes). Once an article has more than three succession boxes, they tend to stop being useful. —Kusma (t·c) 21:50, 1 March 2021 (UTC)[reply]
• Yes - A lot of succession boxes seem to be pure trivia. If there isn't an article for the topic it shouldn't exist as a succession box. Nosferattus (talk) 04:05, 9 March 2021 (UTC)[reply]
• ? Oppose Confused, how can this RFC decide anything? The RFC is vague and the answers equally ambiguous; everyone saying "Yes" is supporting something different. Who decides which ones are trivial? Aza24 (talk) 23:34, 15 March 2021 (UTC)[reply]

Discussion (succession boxes)

• Why? Ivanvector's squirrel (^trees/_nuts) 22:17, 4 February 2021 (UTC)[reply]

  Because, those are not political or party offices. They're merely trivial. GoodDay (talk) 22:19, 4 February 2021 (UTC)[reply]

  Why do we need a global policy or guideline or anything about this? If you think a given succession box is trivia, discuss that specific succession box somewhere (talk page, WikiProject, TfD, there are plenty of venues). I genuinely don't understand what you are trying to achieve here - there is no chance that there will be a consensus that we should have succession boxes for everything (10th place qualifier in a British Touring Car Championship race as an unquestionable example), but with the wording of the discussion you cannot get consensus (for or against) regarding any individual example. Thryduulf (talk) 02:18, 5 February 2021 (UTC)[reply]

  You expect there to be a separate discussion on each individual bio article? Anyways, I don't exactly know what you're posting about. GoodDay (talk) 17:29, 5 February 2021 (UTC)[reply]

  I expect you to get consensus for each individual succession. That could be on an article talk page or a centralised discussion. e.g. if you want to get rid of oldest living British Prime Minister, you need to have a discussion specifically about removing the "oldest living British Prime Minister" succession box from every article it appears on, with notifications to (at least) the talk page of all the affected articles. In other words you need to do exactly the same as you would for any other bulk change. Thryduulf (talk) 22:26, 5 February 2021 (UTC)[reply]

  That's too time consuming. There's too many 'useless' topics in these succession boxes, to go through one-by-one. GoodDay (talk) 22:30, 5 February 2021 (UTC)[reply]

  A centralised discussion about the succ boxes for oldest living British Prime Minister could be held at Wikipedia talk:WikiProject Politics of the United Kingdom. --Redrose64 🌹 (talk) 23:40, 5 February 2021 (UTC)[reply]

  But the British prime ministers example, is just one example of these trivial topics in suc-boxes. GoodDay (talk) 23:41, 5 February 2021 (UTC)[reply]

  The issue is that we have no idea what other succession boxes you regard as trivial so we (and most importantly editors of the articles they appear on) have no way of knowing whether we agree or disagree. While oldest living British Prime Minister doesn't seem very useful, others such as Prime Minister of the United Kingdom is very much not trivial, where to draw the line between them needs to be determined by consensus. Thryduulf (talk) 00:59, 6 February 2021 (UTC)[reply]

  Subjects like "Tallest President of country", "Fattest Prime Minister of country", "Oldest Stock Car racer", "Youngest 100 meter dasher", would be trivial topics for suc-boxes. GoodDay (talk) 02:18, 6 February 2021 (UTC)[reply]

  The first two undoubtedly would be trivial, the latter two maybe - it would depend if there was any significance given to this in reliable sources. However a short list of topics (which a cursory search suggests are not in use) that you consider trivial do not go any way towards addressing the issue in my comment. Thryduulf (talk) 18:20, 6 February 2021 (UTC)[reply]

  I don't know what your complaint is. Perhaps if you would put down examples of suc-box topics that you believe should & shouldn't be deleted, would help. GoodDay (talk) 18:26, 6 February 2021 (UTC)[reply]

  My point is that they need to be discussed individually or in small, closely related groups (e.g. perhaps succession boxes related to British Prime Ministers). No matter how many examples are listed here you cannot get consensus for anything not listed here, and the more examples you list here the greater the liklihood of a trainwreck. Thryduulf (talk) 00:10, 7 February 2021 (UTC)[reply]

  RFC is already in full swing. We'll see how it ends up. GoodDay (talk) 00:47, 7 February 2021 (UTC)[reply]

  Consensus that some but not all succession boxes should be removed, but no consensus for the removal of any one in particular - that will need further discussion. It's pretty much the only way it can end. Thryduulf (talk) 02:20, 7 February 2021 (UTC)[reply]

  Besides, I find Wiki-Projects tend to garner less attention. Was considering moving another RFC to Village Pump, for the same reason. GoodDay (talk) 23:55, 5 February 2021 (UTC)[reply]

  Your subjective opinion that something is "Too time consuming" does not grant you the right to ignore

  WP:CONSENSUS. If you propose a course of action on a WikiProject and advertise the discussion to the relevant talk pages but nobody opines after a reasonable time (~week) then you can go ahead and remove the succession boxes listed in the discussion. Thryduulf (talk) 00:59, 6 February 2021 (UTC)[reply

  ]

  If you want to advertise this RFC on the related WikiProjects? then by all means, do so. GoodDay (talk) 02:18, 6 February 2021 (UTC)[reply]

  If this discussion was actually about specific succession boxes that would be appropriate, but as the discussion is actually a vaugie and unfocused attempt to get rid of an unspecified list of succession boxes you (or presumably any other editor) personally dislikes then it is impossible to know which projects and articles are relevant. On the other hand, if you did deign to list those boxes you deem trivia, then it would I suspect rapidly become a trainwreck due to the large number of disparate boxes editors will have different opinions about. Thryduulf (talk) 11:55, 6 February 2021 (UTC)[reply]

  Since my concerns grew out of the discussion at James Callaghan, one would link to Political WikiProjects. GoodDay (talk) 18:41, 6 February 2021 (UTC)[reply]

All should be delete as links spam due to duplication of links and undue because of overwhelming size. Never understood why we have giant boxes with very few links in them overwhelming the sections. It's definitely a point of contention for content editor that these undue boxes are spamed automatically without consideration of over linking or unwarranted linking.--Moxy 🍁 00:04, 6 February 2021 (UTC)[reply]

Very simply because many (not all) of the succession boxes are very useful for readers. Thryduulf (talk) 00:59, 6 February 2021 (UTC)[reply]

Not sure how duplication of lnks and overwhelm sections is good for readers. We have protocols for these 2 points....just ignored by template spammers.— Preceding unsigned comment added by Moxy (talk • contribs) 01:19, 6 February 2021 (UTC)[reply]

See my reply in the section above for why duplication is not a problem, and I disagree that the presentation is overwhelming. That some succession boxes are trivia does not mean every succession box is spam. Thryduulf (talk) 01:41, 6 February 2021 (UTC)[reply]

We will simply have to disagree. By placement practice alone indicates there very low value to the community. See also links dumped at the bottom of articles because they duplicate existing links and the format is not responsible anywhere else in the articles. It's horrible that these boxes are more prominent than the actual topic-specific navigation boxes. Odd they were not omitted from mobile view as load junk.--Moxy 🍁 02:01, 6 February 2021 (UTC)[reply]

How does placement indicate they are low value? Of course the format is not appropriate anywhere else in the article - see also links and links in succession boxes have a completely different purpose to links in the prose. You need to explain why the duplication is problematic. Thryduulf (talk) 11:55, 6 February 2021 (UTC)[reply]

It's why we have a guideline on this..... distracts readers from the links that are actually important Wikipedia:Overlink crisis. They are so overwhelming that editors hide them in collapsible templates Abraham Lincoln#External links. In many other cases the amount of them is simply overwhelming...mass link spam John A. Macdonald#External links.--Moxy 🍁 17:59, 6 February 2021 (UTC)[reply]

You've explained why you think having too many boxes is an issue, and reiterated your opinion that some boxes are low value (which basically nobody is disagreeing with). However you have not explained why having any boxes is problematic or why their position in the article indicates this. Thryduulf (talk) 20:24, 7 March 2021 (UTC)[reply]

Should we use custom DISPLAYTITLE? We have dozens of titles where custom DISPLAYTITLE would be useful, such as thatPower and C Sharp (programming language). Using custom DISPLAYTITLE would allow us to bypass these restrictions and allow us to use any DISPLAYTITLE and deprecate the template {{correct title}} Aasim (talk) 20:22, 9 February 2021 (UTC)[reply]

• No. --Izno (talk) 22:19, 9 February 2021 (UTC)[reply]
  • @Izno: care to elaborate? —El Millo (talk) 22:21, 9 February 2021 (UTC)[reply]
    • Consider the abuse possible with arbitrary titles. --Izno (talk) 22:23, 9 February 2021 (UTC)[reply]
    • Never mind abuse, consider how hard it would be to identify the correct title so you could do things like basic linking of the page. Yeah, no. --Izno (talk) 22:36, 9 February 2021 (UTC)[reply]

There would need to be some system to only allow "valid" alternate names—currently only modifications that result in the same text (ignoring formatting/normalization) are allowed. For C#, you'd need to allow adding "#" (maybe reasonable to code in as an unsupported character), but also removing "Sharp" (harder to code in, because it already bans hiding arbitrary parts of the page title). One possibility is a separate software feature so the title could only be edited by users with a certain permission, but I don't think this will get much support and it would be a lot of work for fairly minor gain. And there's still the problem that the displayed title won't work if you copy it into the search box or try to link to it. — The Earwig ⟨talk⟩ 22:40, 9 February 2021 (UTC)[reply]

@The Earwig: Removing chars from the title is already supported via the font-size=0 trick, though I think it's generally seen as bad practice. – SD0001 (talk) 14:03, 10 February 2021 (UTC)[reply]

From Wikipedia:Page name#Changing the displayed title, it was formerly possible to hide text with display: none; but this is no longer allowed. It makes sense there are other workarounds, but I'm certain this is a bad idea. For one, I'm not sure how screen readers will handle that. — The Earwig ⟨talk⟩ 14:39, 10 February 2021 (UTC)[reply]

In theory there could be a list of valid substitutions defined by regex somewhere or something, so the complete word "sharp" can be substituted for "#" or the like. --Aquillion (talk) 21:56, 15 February 2021 (UTC)[reply]

• No basically per all the issues Izno already raised. — xaosflux ^Talk 00:10, 10 February 2021 (UTC)[reply]
• I never knew C# isn’t at the correct title. That’s weird. If this proposal is approved, would this currently technically possible with a config change or something, or would code need to be written up? Anyway, it seems to me that ideally # should be supported as a character in normal titles, not using a display title hack. ProcrastinatingReader (talk) 00:16, 10 February 2021 (UTC)[reply]

  @ProcrastinatingReader: Well, to enable # as a legitimate page title without DISPLAYTITLE hacks, you would still have to treat it as a special case because # has a special meaning in URLs. davidwr/(talk)/(contribs) 00:27, 10 February 2021 (UTC)[reply]

  Yep. If a # was a supported character in titles, should

  Foo#Bar take you to the article "Foo#Bar" or the section "Bar" in article "Foo"? – SD0001 (talk) 14:04, 10 February 2021 (UTC)[reply

  ]

  How do articles handle it? eg

  HC Zemgale/LUA? Or is / (ie subpages) not valid in article space? ProcrastinatingReader (talk) 18:48, 12 February 2021 (UTC)[reply

  ]

  Subpaging is turned off in main space (see example Fate/stay night). --Izno (talk) 19:02, 12 February 2021 (UTC)[reply]

  This is technically possible as a config change (setting $wgRestrictDisplayTitle to false) * Pppery * _{it has begun...} 00:37, 10 February 2021 (UTC)[reply]

• Not arbitrary changes carte blanc, but I am open to allowing it on a case-by-case basis after a discussion similar to that of a controversial page-move discussion. I'm also open to allowing whole new classes of "DISPLAYTITLE" to be allowed e.g. "allow #", again, after an RfC for similar discussion for each proposed exception. This could be implemented by allowing DISPLAYTITLE when there is a 1-1 mapping from page titles to DISPLAYTITLE arguments in a "whitelist file" maintained by administrators. Of course, we are talking about a code change, and developer time isn't unlimited. I don't see this as a priority, but if developer- and tester-time is available, I'd favor it. davidwr/(talk)/(contribs) 00:24, 10 February 2021 (UTC)[reply]

IMHO I think abuse is a bit unlikely given that most people dropping by Wikipedia to edit it are here in good faith and probably know nothing about how "DISPLAYTITLE" works. Also it appears DISPLAYTITLE is hidden deep in VisualEditor anyway, so I do not think abuse is likely to happen.

And messing with the "DISPLAYTITLE" of a page could just be treated as disruptive editing, just as we treat page move wars, tendentious comments, etc. Aasim (talk) 00:55, 10 February 2021 (UTC)[reply]

• Oppose It's too annoying and error-prone if the displayed name cannot be copy-pasted to a wikilink or the search box. It can also be confusing if you click a link in search results or somewhere else and come to a page with a different title. PrimeHunter (talk) 15:38, 10 February 2021 (UTC)[reply]
• Oppose. When I worked for a music tech startup, one of our pains in the butt was artists like Ke$ha and NIИ. We were constantly special-casing things like this so people could find them in searches. Yes, it is annoying that typing "C#" into a wiki-search box gets you to
  MOS:DABPIPE). But, I think the alternative would be worse. -- RoySmith (talk) 15:50, 10 February 2021 (UTC)[reply
  ]
• oppose The characters that are allowed in the title are limited for good reasons as far as I can tell. Specifically, # already has a meaning in URLs and so would be a real problem for wiki links. — GhostInTheMachine ^{talk to me} 18:08, 10 February 2021 (UTC)[reply]

  This is not about allowing # in wikilinks, but allowing custom DISPLAYTITLE. Aasim (talk) 21:51, 10 February 2021 (UTC)[reply]

  I get that, but imagine the fun people will have trying to type a wiki link to a page that has a # character in the displayed title. Is a wiki link of A#B to a page called A#B or to a link called B in page A? Is the link to a page called A-something-else, but displayed as A#B? If the page title is displayed as A#B, what should the link be? — GhostInTheMachine ^{talk to me} 22:53, 10 February 2021 (UTC)[reply]

• The issue with this is that we don't want title-spoofing abuse. Perhaps, in certain cases, we could have a system for displaying the "technical title" in a less prominent alongside the "display title"? --Yair rand (talk) 06:19, 18 February 2021 (UTC)[reply]

  We could take an approach like Wiktionary, and have an "Unsupported titles" page where we can have all the unsupported titles listed. Then the technical hatnote would be unnecessary and the links would become clearer. Aasim (talk) 02:49, 21 February 2021 (UTC)[reply]

  Please provide a link to that Wiktionary page. --Redrose64 🌹 (talk) 08:28, 21 February 2021 (UTC)[reply]

  What I mean is "unsupported titles" prefixpage and have subpages of it be unsupported titles. See wiktionary:Unsupported_titles/Number_sign for an example. For the page C#, we could have it be located at Unsupported titles/C-sharp and then have JavaScript change the title displayed to C#. Aasim (talk) 08:40, 21 February 2021 (UTC)[reply]

  There appears to be some JavaScript going on: on visiting the page, the title showed as "Unsupported titles/Number sign" briefly, which was then replaced by a single "#" character. --Redrose64 🌹 (talk) 09:00, 21 February 2021 (UTC)[reply]

  Yeah, it's wikt:MediaWiki:UnsupportedTitles.js, which has a central list of titles. Problem is, trying to link directly to the displayed titles will not work. --Yair rand (talk) 11:54, 2 March 2021 (UTC)[reply]

• Support as there are many titles that should have square brackets [] but can't. # is also a problem. But perhaps this could be done with a mechanism to DISPLAYTITLE that enabled checking for abuse. An edit filter could check for mistakes if a standard is set for mapping to the name. Note that there other unicode characters that are similar in appearance to the forbidden ones. Graeme Bartlett (talk) 22:47, 24 February 2021 (UTC)[reply]
• No per all the points raised by IznoSea Ane (talk) 21:46, 26 February 2021 (UTC).[reply]
• Oppose unless the actual page title is also displayed prominently. —Kusma (t·c) 11:39, 2 March 2021 (UTC)[reply]
• No - per Xaosflux and Izno. A lot of potential headaches for very, very little gain. ƒirefly ( t · c ) 13:33, 2 March 2021 (UTC)[reply]

Should non-hyphenated parameters be fully removed from the CS1/2 family of templates? Primefac (talk) 02:14, 10 February 2021 (UTC)[reply]

Background (CS1)

In 2014 an RFC determined that all parameter names in Citation Style 1 templates should have an alias which is in lowercase that has separations with hyphens between English words, between (not within) acronyms, or between English words and acronyms. The documentation is to show this lowercase, hyphenated version as the one for "normal use". This meant, for example, that |access-date= would be shown as the preferred parameter while |accessdate= was shown as acceptable but discouraged from use. In the following years there have been various trends and discussions formally deprecating many of the non-hyphenated parameters, from a small handful (2019 example) to the current list which contains over 70 entries. Many of these are the non-hyphenated variants of the preferred/hyphenated versions, which are being removed to decrease the maintenance burden and increase the uniformity between templates (i.e. "ease of use"). Other changes have included having RefToolbar give the hyphenated params and setting AWB's genfixes to replace some parameters.

In October 2020, all non-hyphenated parameters were added to the "current list" linked above. In November 2020, a

The issue for Monkbot 18 arises from the number of edits it is/was required to make; a conservative estimate would put the number of edits it has made for this task over a two month period (Nov 2020-Jan 2021) at around 1 million edits; as discussed on the task's talk page, this has essentially removed all but five of the non-hyphenated parameters, but another 2-3 million edits taking up to four additional months will still be required to fully "clear out" these parameters. Additionally, those opposed to the bot also expressed concern that the relatively small-scale discussions to deprecate these parameters had not reached a wide enough audience to merit what they felt were sweeping changes.

Proposal (CS1)

There are three main options with regard to the CS1/2 family of templates, and by extension Monkbot 18's task.

• Option A: Non-hyphenated parameters should be deprecated and removed; the bot is free to continue its work.
• Option B ("status quo"): Non-hyphenated parameters are formally deprecated, but should not be immediately removed. Deprecation can be bundled into genfixes or performed along with other non-cosmetic changes, but (ignoring a possible
  Cosmetic Bot Day
  ) should not be done on its own by a bot.
• Option C: Non-hyphenated parameters should not be deprecated; deprecation should not be continued and bot approval should be revoked. This will also mean that the deprecated parameter list will need to be updated to remove the non-hyphenated parameters.

Please note that this discussion is primarily about the CS1/2 template parameters and whether two full sets of parameters should be kept/maintained. It is not the place to re-litigate the various discussions about Monkbot 18's task; Option B is provided for those who feel the task should not proceed.

Survey (CS1)

• First choice A, second choice B. I'd be happy to see
  AWB's genfixes take on some of this burden, and I'd be happy to see this happen a little more slowly, but it should happen, even though it's occasionally inconvenient for me. Also, when any individual parameter reaches a sufficiently small state (e.g., potentially still thousands of uses, but not hundreds of thousands of articles), the template should be updated to disallow that particular parameter (not merely advise against it in the documentation), so that they won't keep creeping back in, because, hey, it still works, so why should I bother? WhatamIdoing (talk) 19:12, 10 February 2021 (UTC)[reply
  ]

  @

  WP:AWB/RTP, so manual edits and other bots can help whittle down the list while making other (non-cosmetic) edits. GoingBatty (talk) 04:12, 11 February 2021 (UTC)[reply

  ]

  WP:AWB/RTP? I think it was removed. – Jonesey95 (talk) 04:33, 11 February 2021 (UTC)[reply

  ]

  @Jonesey95: You're right! While the functionality exists (and other parameters are still replaced), editors have removed some hyphenation replacement rules. GoingBatty (talk) 04:59, 11 February 2021 (UTC)[reply]

• Option A. I support completing the nearly finished move away from unhyphenated multi-word parameters. See below for more details about this process, which is being questioned now by a very few editors after seven years of work, and when it is more than 90% complete. With any other template, it would have been the work of a few days to standardize on one style of parameter name and convert all of the transclusions. The only reason that the process for CS1/CS2 templates is different is that there are millions of transclusions instead of hundreds. – Jonesey95 (talk) 04:33, 11 February 2021 (UTC)[reply]
• Option C. This is pointless make work; see extended comment in discussion section. Espresso Addict (talk) 07:08, 11 February 2021 (UTC)[reply]
• Option C If it was only confined to a few thousand pages, it wouldn't be a big deal. But when it's upwards of 3 million pages, maybe it should be the other way round - IE no hyphen. Lots of pointless bot cloggage in going through millions of pages for a trivial change. Lugnuts ^{Fire Walk with Me} 08:20, 11 February 2021 (UTC)[reply]
• B if not C As per comments above; many editors are clearly quite happy with the unhyphenated forms, so why not allow both? Changing is pointless. Lots of other templates allow aliases as parameter names. I fail to see the problem. But we are where we are, so I favour B rather than C. Peter coxhead (talk) 09:04, 11 February 2021 (UTC)[reply]
• Option A (second choice B), per Jonesey95. Let's just get everything simplified, as it should be. Get on and finish the job. --NSH001 (talk) 10:18, 11 February 2021 (UTC)[reply]

  Update Add a qualification: I still very much hope that the bot will be allowed to resume, but subject to this condition: only if Monkbot 18 drops its removal of entirely blank lines within citation templates. For the reasoning behind this, see my conversation with Floydian in the discussion section (this is quite an easy change to make to the bot). While I'm here, I'll add a second qualification, also arising from the discussion: the list of articles on which the bot runs should be filtered to remove articles that have already been visited by the bot. This is in order to reduce the alleged problem of "bot spam" objected to by some editors (who says I don't listen to objections?). --NSH001 (talk) 06:52, 12 March 2021 (UTC)[reply]

• Option C. It is a totally pointless exercise. The unhyphenated ones are better in my eyes as they do not wrap in the edit window and can be underlined as a typo so are clearly visible in the edit window. Keith D (talk) 13:01, 11 February 2021 (UTC)[reply]
• C (first choice) or B (second choice). I've read many of the discussions about this issue and I've never yet seen anything the convinces me that deprecation actually benefits the encyclopaedia in any way. Even if we assume for the same of argument that it somehow does, the real and evident disruption caused by the bot so far and the extent of the changes the bot operator notes will be required to complete the task, very clearly and very significantly outweigh that benefit. Thryduulf (talk) 15:20, 11 February 2021 (UTC)[reply]
• Option A this change is a bit painful, but better for editing in the long-run. It's better to make this change than to not make it. The best time would've been fifteen years ago - but the second-best is now. Allow the bot to continue its operation, get rid of all the parameters, and once all are removed, start generating cite errors.
  talk | contribs) 17:09, 11 February 2021 (UTC)[reply
  ]
• Option A. Unfortunately, the visual editor exists... but isn't useful for large editing projects and thus many people still use wikitext editing - condensing to one form (specifically the hyphenated one as it is easier to read for editors) will help ensure consistency between articles at least in the CS1 templates. Obviously this won't make every article easier to edit as there are articles with non CS1 style citations, but it'll help the millions that do use CS1 look the same to editors instead of having a hodge-podge of hyphenated names. I further agree with the bot continuing to run now, and then running maybe once per week or so after this initial run to fix any CS1 non-hyphenated parameter names. -bɜ:ʳkənhɪmez (User/say hi!) 17:13, 11 February 2021 (UTC)[reply]
• Option C. The point of templates and bots is that they should work to make editors' lives easier, not that editors should change the way they do things to make template and bot creators' lives easier. This bot has it completely topsy-turvy, and if the bot-approvals group has approved this then that is a problem with that group, not with a few unhyphenated parameters. I can't help feeling that that group is looking at the interests of a few bot operators rather than of the many more editors and still many more readers. There's no great complexity in having a few synonyms for template parameters, and there's no problem at all with exporting data - if the synonymic parameters point to the same place in code then they can be exported in the same format with no extra effort. I can't believe that forty years after I went into IT we still expect users to be slaves to the systems that are supposed to help them.
  Phil Bridger (talk) 18:21, 11 February 2021 (UTC)[reply
  ]

  The point of templates and bots is that they should work to make editors' lives easier. Exactly so. That's what this bot is for. It makes the parameter names easier to read (taking them as a whole, not just access-date on its own), reduces the size of the template documentation, makes the parameter names all consistent. All these combine to make learning how to use the cite templates easier. It also makes maintaining the templates easier. a win-win situation. The only reason we're having this discussion is that Mediawiki's watchlist software is so bad at handling bot edits. Otherwise it would be a no-brainer. Some people here seem to be under the mistaken apprehension that this is just to advance the interests of "a few bot operators". Well, I'm quite sure Trappist (the operator of this bot) could do without the stress of planning, writing and running this bot. He does a bloody fantastic job, and deserves a huge amount of credit and appreciation for his work. I can't believe that forty years after I went into IT we still expect users to be slaves to the systems that are supposed to help them. The whole reason for this bot is to make users' life easier. FWIW, I also have experience of IT work more than 40 years ago (seconded for 2 years to work on a (very successful) project - mathematical programming on big data for a life insurance company) and frequently had to liaise with IT people since then. I'm well aware of the problems that can arise when you just allow systems to get more and more complex, so I appreciate Trappist's (and many other people's) efforts to simplify matters. --NSH001 (talk) 23:37, 11 February 2021 (UTC)[reply]

  Clarifying: Re-reading this again this morning, it might appear to readers who don't read carefully that I'm agreeing with Phil Bridger. Nothing could be further from the truth, I still support Option A. Option C makes no sense, for all the reasons set out by SMcCandlish. --NSH001 (talk) 08:35, 12 February 2021 (UTC)[reply]

  And Phil Bridger's hyperbole stance is fallacious. Option B imposes nothing on anyone. "I don't like option A" does not equate to "only option C can work". Option B is the status quo, and it has broken nothing. I'm rather shocked at how stark obvious this is, yet at least 10 editors don't seem to have noticed. I know that we have a lot of populism running around in the world – a lot of "I would feel very strongly about this, if it were true, and it feels good to pretend its true, so I'm going to pretend it's true" behavior. But that stuff needs to be left at the door.  — SMcCandlish ☏ ¢ 😼  04:21, 12 February 2021 (UTC)[reply]

  No, my stance is not hyperbole or due to populism. The reason I reject option B is the word "immediately". It still leaves current consensus that unhyphenated versions of parameters will be removed, just not immediately, and it is that consensus that should change, however many years it has been stable for. I am happy with simplifying the documentation, but not with removal of the option.

  Phil Bridger (talk) 09:19, 12 February 2021 (UTC)[reply

  ]

  As noted above: Deprecation is not synonymous with disabling. You're confusing replacement of the parameters as written in a template transcluded in a page, with removal of the runtogether parameter variant's ability to function in the template. Only option A would do the latter. We have many, many templates that support variant-spelling parameters but do not "advertise" them in the documentation, and it breaks nothing whatsoever to bot-replace them with the canonical version, just as the same bot will replaces calls to a redirect name for the template with the actual page name of the template. I.e., you're having a strong negative reaction to an argument that option B is not actually making.  — SMcCandlish ☏ ¢ 😼  17:50, 13 February 2021 (UTC)[reply]

  Option B very clearly says "but should not be immediately removed". Either the word "immediately" has some meaning, and this option would lead to removal, but just not immediately, or it has no meaning and has no business being there.

  Phil Bridger (talk) 18:10, 13 February 2021 (UTC)[reply

  ]

  SMcCandlish, this was explained in the previous discussion at CS1: ""Deprecation", by the definition used in the context of CS1/CS2, means that if the parameter is used, a red maintenance message will be shown and it will appear in a tracking category. It is a phase before removal of support for a parameter (in which case only an "unsupported parameter" message and optionally a hint on the new parameter will be shown). It is possible to stay in this state for extended periods of time, but the idea is that eventually the functionality will go". So the intention is to error and then remove the functionality of these parameters, not simply not advertise them in the documentation. If you want to propose a variant of option C that removes the parameters from the documentation but still allows them to function, go ahead, but option B is not that. Nikkimaria (talk) 20:09, 13 February 2021 (UTC)[reply]

  What I want to propose is a variant of Option B that removes the parameters from the documentation before letting some bot go around changing the articles on my watchlist as if the parameters are somehow bad. If the params are really bad, and if there's a consensus to that effect, then Step One is to take them out of the documentation (either by dragging them away in the dark of night, silently, like ninjas, or by transparently mentioning them as being deprecated, so everybody knows what's going on). The first phase of real deprecation is telling people to stop using something. Then you can start throwing up red messages and it's not so damned rude or surprising. — JohnFromPinckney (talk) 21:40, 13 February 2021 (UTC)[reply]

  Sure, I would support that variant, too. Or a variant that kept mention of those versions of the parameters but deprecated them. Whatever gets us closer to having consistency in the actual deployed templates being used in citations.  — SMcCandlish ☏ ¢ 😼  04:15, 14 February 2021 (UTC)[reply]

  This stuff should be in a tracking category, if that's what bots and other tools are going to work with. If we don't like the error messages, then we can disable them. This is just template and module code, it's not etched forver into a mountain face like Mt. Rushmore.  — SMcCandlish ☏ ¢ 😼  04:15, 14 February 2021 (UTC)[reply]

• Option C I agree strongly with Phil Bridger here. In other words, I fail to see what exactly is accomplished by making millions of cosmetic edits and then deliberately breaking things that would otherwise have worked. * Pppery * _{it has begun...} 20:21, 11 February 2021 (UTC)[reply]

  Option B breaks nothing. You, et al., are providing an argument against option A, not an actual argument for C, and just ignoring B. Also, a !vote for C is an !vote for overturning a status quo that has been stable for years, in favor of chaos, yet without an actual rationale to do so. The closer should take that into account, per

  WP:NOTAVOTE. In the event of a "no consensus" result we still end up with the status quo. Things like this are why I keep telling people they need to write RfCs better. "Anything that can be misinterpreted will be."  — SMcCandlish ☏ ¢ 😼  04:21, 12 February 2021 (UTC)[reply

  ]

  SMcCandlish, option B supports disabling of the non-hyphenated parameters - that is a breaking change. The difference between A and B is speed, not outcome. Your !vote below suggests you want the non-hyphenated variants to remain supported, but of the options provided only option C accomplishes that. Nikkimaria (talk) 16:18, 13 February 2021 (UTC)[reply]

  Nope. See my response to same claim by Phil Bridger, above.  — SMcCandlish ☏ ¢ 😼  17:50, 13 February 2021 (UTC)[reply]

  Yep - see above. You can also look at what has happened with previously deprecated CS1 parameters such as |authorfirst= - they don't work. Nikkimaria (talk) 20:09, 13 February 2021 (UTC)[reply]

  Which is perfectly fine, given enough time that people stop actually using them. Thus remove them from the template documentation and replace them in deployed template translcusions. Eventually the "monkey see, monkey do" effect of being exposed to deprecated parameter variants goes away, through decreased and eventually zero exposure. I'm mean, come on, that's what the point of deprecation is. It seems to me that you [plural] are coming from a "give me C or give me death" perspective, artificially conflating A and B because you just will not tolerate the idea of any parameter name variants ever going away for any reason. If I'm wrong about that perception, then please explain.  — SMcCandlish ☏ ¢ 😼  04:18, 14 February 2021 (UTC)[reply]

• Option C per Phil Bridger. Ealdgyth (talk) 21:28, 11 February 2021 (UTC)[reply]
• Option C - Commenting here probably falls under my doctors' lists of "things HF shouldn't do while he has a concussion", but I don't want to miss this discussion. While I understand that maintaining the citation templates is not a particularly easy job, in the end, rigidness in the citation template is not desirable. The citation templates should be easy to use, and having a couple aliases for the most common parameters makes it easier to use. And a trout to the bot guild for approving a bot task that was designed to deprecate template parameters with no consensus for that deprecation. Hog Farm _Talk 22:15, 11 February 2021 (UTC)[reply]
• First choice A, second choice B. Wikipedia source text is becoming increasingly complex and difficult to manage making it less accessible, except for the type of experienced editors participating here. Anything we can do to reduce complexity is a win, fewer options the better when plainly redundant. -- GreenC 22:41, 11 February 2021 (UTC)[reply]
• Option C per Phil Bridger, personally I prefer non-hyphenated parameters, and I find deprecating them to be an absolutely pointless exercise that breaks things for absolutely no reason other than to satisfy the cosmetic preferences of a few editors. Devonian Wombat (talk) 00:14, 12 February 2021 (UTC)[reply]
• Option C. Largely per Hog Farm. Keeping extremely commonly used aliases is helpful for editors using these templates. Nikkimaria (talk) 01:17, 12 February 2021 (UTC)[reply]
• Support option C, neutral on option B, strongly oppose option A. My fingers are used to typing many of these parameter names without hyphens. Deprecating them and the accompanying gnomework of replacing them with the un-deprecated versions is already causing me significant hassle, both in trying to remember that really now they should be hyphenated, and in trying to pick out the important changes on my watchlist among all the pointless gnomery. Removing the unhyphenated forms altogether would be worse. —David Eppstein (talk) 01:41, 12 February 2021 (UTC)[reply]
• Option C. I'm with Phil Bridger and Hog Farm. I don't know if this is bike shedding or yak shaving, but it's just not productive and a bad look. Alternatives exist because it's simpler than remembering which of two common possibilities are acceptable, rather than forcing one or the other (as the other options do). Discussing efficiency because of a off-row character on the other hand (oh, so we're making this choice based on everyone using QWERTY?) is the kind of ignore-practical-facts reasoning that yield platypus-shaped end results. -- Mikeblas (talk) 01:52, 12 February 2021 (UTC)[reply]
• A is better in the long run, but B for now, and have the conversion covered by AWB genfixes and similar.
  b} 02:29, 12 February 2021 (UTC)[reply
  ]

  Comment Unfortunately, in this case, leaving it to AWB genfixes alone would be neither effective nor desirable. Because there are likely to be so many of these parms on any given page, the genfixes are likely to swamp the main, intended change, causing understandable annoyance. Much better to leave it to this excellent bot, which is clear and open about what it is doing – no nasty surprises. That's why I prefer option A to option B, but either of these is vastly preferable to option C. --NSH001 (talk) 08:31, 7 March 2021 (UTC)[reply]

• Option C per Phil Bridger. SarahSV ^(talk) 02:51, 12 February 2021 (UTC)[reply]
• Option C. This "everything must be hyphenated" approach doesn't work well with the text editor. For some common parameters like |accessdate=, it is simply better being unhyphenated because the source text is quicker and easier for a human to parse because the parameter doesn't word wrap in the editor's textbox. Jason Quinn (talk) 03:09, 12 February 2021 (UTC)[reply]
• Option B. We should stop listing the nonhyphenated ones in the documentation at very least, so that between editorial shift and AWB/bot genfixes cleanup, we get more consistent over time. It's too soon for option A, if ever, because the templates serve us, we don't serve the templates. It's perfectly fine for templates to quietly support non-hyphenated variants so they don't just break if people try them. But we should not continue listing those variants in the docs. It's antithetical to the purpose of templating, for us to perpetuate inconsistency (without good reason, like an ENGVAR color vs. colour distinction). And it pointlessly makes the documentation longer and more complex for no gain at all; no one looking for how to specify the Archive.org URL needs to know anything but |archive-date=, and telling them |archivedate= also works is just stuffing pointless trivia into their head. Yes, do continue converting to hyphenated versions in genfixes and other automated edits (when doing something more substantive at the same time). Finally, option C is rather pointless. We've regularly been (gradually) deprecating various old parameter names, and it has worked just fine. Option B will not break anything, and will have (already is having) positive results.  — SMcCandlish ☏ ¢ 😼  04:21, 12 February 2021 (UTC)[reply]
• Option B Option C. These are all valid aliases, as there is zero confusion between say |access-date= and |accessdate=. The status quo works fine: hyphenated is preferred because it's easier to read, but unhyphenated is acceptable because there is no ambiguity and evidently plenty of people type it that way. Cosmetic changes should adhere to policy. – Finnusertop (talk ⋅ contribs) 05:53, 12 February 2021 (UTC) Changed from B to C as I am opposed to the implications of the "formally deprecated" part of these valid aliases. – Finnusertop (talk ⋅ contribs) 09:26, 12 February 2021 (UTC)[reply]
• Option C first choice (B second). Editors should be allowed to use either hyphenated or non-hyphenated versions. Consistency is not better than flexibility here: the only people reading the parameter are editors, so let the editors decide whether or not they want to use a hyphen in their template parameters. I share the general concern about the disconnect between code writers and content writers, and the frustration with some template and bot maintainers imposing decisions on everyone without consensus. Fait accompli editing across millions of edits or transclusions is kind of a big deal. Levivich ^harass/_hound 07:20, 12 February 2021 (UTC)[reply]
• I Like C: I especially echo Levivich, Thryduulf|, and Phil Bridger's comments regarding these perennial, periodic, new surprises where editing articles is concerned. That is disruptive and we should just stop it. C, therefore, will bring the sanity back. GenQuest ^"scribble" 07:42, 12 February 2021 (UTC)[reply]
• Option C. These are templates for use by editors, they have no impact on readers, and it's a complete waste of time making rules and regulations about them and then writing bots to enforce those pointless rules. Not to mention that when AWB goes through "fixing" all these in an article, it can drown out the genuine edits and make it harder for people to track what's going on. Get rid of this ridiculous rule, delete it from AWB, and then maybe we can get on with actually building an encyclopedia.  — Amakuru (talk) 09:13, 12 February 2021 (UTC)[reply]
• Option C (which is the
  busywork edits clogging up watchlists for no good reason. We don't have a problem with template aliases, so why the concern about parameter aliases? None has been convincingly articulated. Modest Genius ^talk 12:10, 12 February 2021 (UTC)[reply
  ]
• Since I have been pinged, neutral all round. I'm have nostalgia for the status quo for some reasons already presented (muscle memory, line breaks), but I recognize that once we pass through the valley of the shadow and emerge into the bright uplands yonder of a cleaner implementation, we'll have forgotten the pain. Mind you, I'll probably be 90 years old and beyond caring. But I have some implementation concerns in the Discussion. David Brooks (talk) 16:54, 12 February 2021 (UTC)[reply]
• Option C - Standardization of this sort may be useful to researchers or developers, but not to regular users. Adding citations should be as easy as possible. To that end, I want to minimize the chances that the interface will not know what I'm trying to do, or that I'll get an error because I entered an underscore instead of a hyphen. The rest of the internet is trying to increase flexibility to make user experience easy and intuitive. We do that too in many ways. But here we seem to be doing the opposite: removing flexibility and requiring users to know how we've worded/arranged things.
  I don't care if there's a bot that goes in an standardizes them afterwards or if someone runs AWB behind me. Again, I get the appeal of standardization. We should just be doing everything we can to make the experience intuitive for regular users. — Rhododendrites ^talk \\ 23:04, 12 February 2021 (UTC)[reply]
  • Alternative: Option D - let the bot and/or AWB standardize, but never disable the parameters. Standardization is good; degrading usability is bad. — Rhododendrites ^talk \\ 23:12, 12 February 2021 (UTC)[reply]

    I would certainly support your option D if it was on the table, but cannot support anything that says that this functionality should be removed, as option B (despite the illiterate claims above that it doesn't) clearly mandates.

    Phil Bridger (talk) 18:10, 13 February 2021 (UTC)[reply

    ]

• Option A with B as my second choice. Maintaining our complex citation templates is not an easy task, and if the people who are putting in the work want to do this to make their jobs easier, I'm in favor of it. Legoktm (talk) 23:48, 12 February 2021 (UTC)[reply]
• Option C with Option B as second choice:
  WP:COSMETICBOT so overt. I don't accept that option B reflects past consensus in practice in that I can't remember anyone ever approaching me about my use of |accessdate= or changing it to |access-date= manually. Past local consensus among technical groups who don't work in article space, sure. But enwiki community consensus? No. — Bilorv (talk) 01:07, 13 February 2021 (UTC)[reply
  ]
  • @Bilorv: if I turn bot edits off I will miss acts of vandalism, unconstructive changes or cleanup tagging that are hidden by a subsequent bot edit -- this always takes me aback. Just curious, but why not enable the "Expand watchlist to show all changes, not just the most recent" + "Group changes by page in recent changes and watchlist" settings? I see no bot edits, and they don't cover anything up. — Rhododendrites ^talk \\ 02:50, 13 February 2021 (UTC)[reply]
    • @Rhododendrites: appreciate the suggestion! It's never occurred to me that combining those would produce that functionality (there are so many complex combinations of Watchlist filters and Preferences options), but I've enabled those two and kept "Latest revision" on and enabled "Human (not bot)" as the Preference option "Hide bot edits from the watchlist" didn't seem to do that and now I think (small sample size in my watchlist atm) I see the latest change only, including if it's a bot, but only if at least one non-bot edit has been made (which is exactly desirable). — Bilorv (talk) 11:34, 13 February 2021 (UTC)[reply]
• Option A with B as my second choice. Maintaining the complex citation templates is not an easy task. AManWithNoPlan (talk) 02:52, 13 February 2021 (UTC)[reply]
• Option A on the basis that standardization is good. I don't actually care whether the hyphens are there or not, but it's better one way or the other rather than having a mix. In general I'd rather see us migrate away from using wikitext for reference information, though, it's like doing your finances in a text document. Thanks. Mike Peel (talk) 18:26, 13 February 2021 (UTC)[reply]
• Option A I have always preferred the hyphenated form personally because it allowed the spell checker to verify that there was no typo, whereas the unhyphenated form is always flagged as a typo, although the preview now informs me of errors. I disagree with the contention that humans can parse |accessdate= more quickly than |access-date=; spaces were invented for this reason. I also know the overhead that permitting multiple parameters that mean the same thing in templates entails. I'm aware that this has been cluttering my watchlist with Monkbot edits, but my !vote is to continue. Hawkeye7 (discuss) 22:12, 13 February 2021 (UTC)[reply]

  Note that this is mostly an

  WP:ILIKEIT rationale and dismisses the views of those who indicate that they can parse "accessdate" without issue as wrong without any evidence. It also dismisses the very real disruption caused to others as necessary to reduce overheads, whereas this overhead, even if it is a significant as some claim (for which no convincing evidence has ever been presented) it's still trivial compared to the disruption caused by Monkbot's edits and by the unnecessary disruption to editors using the templates. Thryduulf (talk) 15:29, 14 February 2021 (UTC)[reply

  ]

• B or C. I've no strong preference between accessdate or access-date, both seem useable, if one is formally preferred then fine. However, the massive ongoing bot spam for something that has literally no effect on readers, and barely any on editors, is unwarranted. In addition to being everywhere on the watchlist, Monkbot makes it much harder to disentangle various series of diffs in the edit history for little benefit. A user making an edit inserting "accessdate" isn't an egregious issue that should cause a bot to come running, and such a bot action then obscures the edit in question from watchlists. CMD (talk) 18:18, 14 February 2021 (UTC)[reply]
• Option C. Removing a common way to type parameters in templates reduces ease of use for the end users. Having a bot going around and "fixing" these has a negative impact on the readability of page history and watchlists . No one in this conversation has demonstrated that the maintenance burden on the template is so significant that it would justify all of these downsides. It also seems that the maintenance burden is not something that would be difficult to track like accidental blue links to primary topic articles when non-primary topic articles were intended, since the template code are on the template pages.---- Patar knight - ^chat/_{contributions} 18:53, 14 February 2021 (UTC)[reply]
• Option C. The previous discussions linked above (a six-year-old RFC with seven people participating in it, which specifically promised that nothing would be depreciated, followed by a handwavy argument about maintenance burden), are not remotely sufficient to justify such a sweeping change. Yes, maintenance burden is a pain, but it affects a relatively small handful of bot authors; removing the most widely-used version of a template parameter affects a huge swath of editors, who need to be given deference here on account of being a larger group. And obviously there is not a standing consensus that maintenance burden justifies such changes (at least not one on a scale necessary to justify this), or this discussion wouldn't be so clearly split. Therefore, the unhyphenated version should never have been depreciated, which makes the bot's edits pointless clutter at best and an attempt to push through a controversial change without sufficient consensus via fiat accompli at worst. Furthermore, if maintenance burden is the only concern, obviously the solution is to reverse the 2016 RFC (which, again, had only seven people participating it and agreed merely to create the hyphenated versions as alternatives) and remove the hyphenated version, which currently sees little use and would therefore be far easier and less disruptive to discard. --Aquillion (talk) 22:12, 15 February 2021 (UTC)[reply]
• Option B-ish. I think it's reasonable to have the hyphenated forms be the canonical version of the parameter, but I see no reason to make mass-edits to change from one form to another or to change the usual rules about cosmetic bots here. I see no harm in Option C, but implementing Option A will trade current problems for new ones. --AntiCompositeNumber (talk) 02:15, 16 February 2021 (UTC)[reply]
• Strong Support A: standardization for template parameters is important & useful.

Mild Support B: the # of |accessdate= per page is too damn high (much of the time), so much so as to interfere with checking regular

WP:GenFixes (i.e. many single-screen diffs become many-screen diffs) — I would Strong Support B IIF Monkbot

was allowed to continue & hyphenate at least this parameter.

Strong Oppose C: antithesis of A.

~ 

dgaf)  19:23, 16 February 2021 (UTC)[reply

]

• To whom is standardization for template parameters important & useful?
  Phil Bridger (talk) 20:39, 16 February 2021 (UTC)[reply
  ]
• How and why is standardisation more important and useful than editors being able to improve the encyclopaedia without needing to know the exact format of parameter names and deal with watchlists and page histories full of cosmetic edits? Thryduulf (talk) 20:45, 16 February 2021 (UTC)[reply]

• Are y'all suggesting we bring back all deprecated parameters, and adding more so that every user may choose to use the parameter names that are most comfortable/understandable/intuitive to them? I'm ok with soft-deprecation - allowing both but discouraging/converting one, in bulk once via bot, then gradually/passively via
  dgaf)  12:16, 17 February 2021 (UTC)[reply
  ]

  Yes, I am suggesting that all the two-word parameters should accept hyphenated and non-hyphenated varieties. It's fine for one to be preferred over the other in documentation but both should work and continue to work as this is by far the least disruptive to editors and allows them to spend their time producing/maintaining content rather than worrying about finicky syntax. I don't have massively strong objections to general fixes substituting one for the other when making non-cosmetic changes to the page, but I wouldn't actively encourage it as it will clutter diffs for no benefit. I strongly oppose bulk bot runs and making one option non-functional in the future. Thryduulf (talk) 15:29, 17 February 2021 (UTC)[reply]

  Re "Are y'all suggesting we bring back all deprecated parameters" - yep, that's spot on. They should never have been deprecated in the first place. This is a template, which sits in the background, and exists for editor convenience, nothing more. Deprecating parameters reduces that convenience. And it's well-established that bots and editors shouldn't be running through making cosmetic changes to wikitext that do nothing to alter the page, so those need to stop as well.  — Amakuru (talk) 12:48, 18 February 2021 (UTC)[reply]

• Option C, I guess, as I've not been persuaded as to the marginal utility of the hyphenated versions. Without that clarity, we shouldn't be doing these kinds of changes. (And if we had consensus, then B, but with documentation changes as the first step.) — JohnFromPinckney (talk) 01:48, 17 February 2021 (UTC)[reply]
• Option C If it works don't fix it. Andrew🐉(talk) 13:01, 17 February 2021 (UTC)[reply]
• Option A, strongly oppose option C: anybody working regularly on templates or modules will appreciate the value of settling on a single style for issues like parameter names. I don't care what the agreed style is, as long as it's consistent, and the argument about whether it should be hyphenated, underscored, camel-case or run-on has been settled with hyphenated as the preferred style. It is then nonsensical to fail to implement that style, and I'd prefer it was done as soon as possible. This whole debate is reminiscent of the date-linking wars where strong objections were made to unlinking dates, yet within a few months of a binding decision being reached to delink dates, everyone had moved on, and nobody today would even consider linking dates. Once we have standardised on hyphenated parameters, future editors will look back and think how lame and time-wasting this sort of debate is. --RexxS (talk) 19:46, 19 February 2021 (UTC)[reply]
• A or B but not C per Scott, Hawkeye, and RexxS. Usingahyphenismoreuserfriendly, andwhilethereissomeupfrontworkonourparttomaketheswitch, itisbetterinthelongruntohavesomedelimiterbetweenwordsinsteadofrunningthemtogether. Just like we stopped linking dates and no longer SmashWordsTogether, this is worth the temporary inconvenience. — Wug·a·po·des​ 00:46, 20 February 2021 (UTC)[reply]

  Notwhen it'sjust twowords :-P Levivich ^harass/_hound 08:10, 20 February 2021 (UTC)[reply]

  ...is what your comment would look like if we allowed people to use whichever method worked for them and still made it work. If we turn off those parameters, your comment wouldn't have appeared. If we allow people to use whatever works and use a bot to clean it up afterwards, your comment would look just like everyone else's after some period of time (during which your comment would still be functional, even if pairs of words were sometimes combined). :) — Rhododendrites ^talk \\ 20:15, 22 February 2021 (UTC)[reply]

• Option A for all the arguments already made at length in the CS1/CS2 talk pages over the years and the good arguments brought forward above. Definitely not Option C. --Matthiaspaul (talk) 02:06, 20 February 2021 (UTC)[reply]
• Option C. Such an absurd waste of time and energy and an enormous source of watchlist spam, all to achieve something that will make editing more difficult. Toohool (talk) 02:26, 22 February 2021 (UTC)[reply]
• Option A as standardizing is better in the long term. Let the bot continue its work. If people have a problem with their watchlist getting spammed, they have the option to filter out bot edits.
  talk 05:59, 23 February 2021 (UTC)[reply
  ]
• Option C if not B While I love consistency, I'm struggling to see what the actual issue is here. I guess they can be gradually changed by the bot along with other more useful changes, but this is just cosmetic to the code and clutters edit histories. Reywas92^Talk 06:21, 24 February 2021 (UTC)[reply]
• Option C. Benefits for bot or template builders don't outweigh the inconvenience for other editors.
  Fram (talk) 08:13, 24 February 2021 (UTC)[reply
  ]
• C. I don't see the argument for mandating hyphens. Just let people use both, consistency on this does not matter. Fences&Windows 17:00, 24 February 2021 (UTC)[reply]
• Comment - my default is Option D... I refuse to use citation templates, and format my citations by hand. It means that I can ignore all the silly debates about parameters and what not. Blueboar (talk) 17:49, 24 February 2021 (UTC)[reply]

  Yeah and we can instantly clear the WP:PHAB backlog by writing articles with paper and pen. We can solve many technological problems by abandoning technology. Levivich ^harass/_hound 00:52, 25 February 2021 (UTC)[reply]

• Option C. Bots do some useful work but their code can be changed as needed. Here I am much more concerned about regular editors who use citation templates when making manual edits. These are live human beings and there are many more of them than bots. Making the template syntax too rigid will make their life much more unpleasant. Extra flexibility is both useful and helpful for regular editors. Nsk92 (talk) 02:26, 25 February 2021 (UTC)[reply]
• Option C. I've been typing accessdate for the last 15 or so years, sometimes while reaching for my drink with my other hand (Qwerty keyboard). Why wreck my muscle memory and deprive me of a sip of tea?-gadfium 04:04, 25 February 2021 (UTC)[reply]
• Option Cper
  Phil BridgerSea Ane (talk) 21:53, 26 February 2021 (UTC)[reply
  ]
• Option C, stop the craziness hitting watchlists (although most of that damage is already done). SandyGeorgia (Talk) 00:58, 2 March 2021 (UTC)[reply]
• C sounds best, followed by B. Removing parameters that were widely used in the past will break old revisions of articles for very little practical advantage. —Kusma (t·c) 11:44, 2 March 2021 (UTC)[reply]
• Option C. Others have already hit on why – watchlist spam, waste of time and energy for no genuine benefit, unnecessary imposition on editors, etc. I'm increasingly warming to the idea of writing out citations manually (as someone else here mentioned), just to avoid the constant tinkering around that seems to take place with these templates. JG66 (talk) 11:51, 2 March 2021 (UTC)[reply]
• Option C per Aquillion. Monkbot should never have been approved --In actu (Guerillero) ^{Parlez Moi} 20:22, 2 March 2021 (UTC)[reply]
• A or B This seems like a choice between having unnecessarily having several different ways to write a given parameter, and between standardizing after the fact with a lot of minor edits. Many of the arguments in favour of C appear to presuppose that editors will land in trouble for writing the unhyphenated parameters, but I don't see evidence of that. Jo-Jo Eumerus (talk) 10:30, 4 March 2021 (UTC)[reply]
• Option A. From real world experience, I understand how difficult it is to maintain code without occasional deprecation. Opponents' fears of unacceptable disruption and inconvenience don't match what I encountered when the bot was running, and typing the hyphen quickly became second nature. --Worldbruce (talk) 05:07, 7 March 2021 (UTC)[reply]

  You may not personally have encountered disruption and inconvenience, but I did and there are a great many other editors in this and other discussions that have reported unacceptable disruption and explained exactly why the inconvenience is not justified. Thryduulf (talk) 20:29, 7 March 2021 (UTC)[reply]

• Option A. Long term benefit is worth the short-term disruption. There is too much inconsistency in templates that causes me to waste far more time (e.g. is it "image=" or "Image=" or "photo=" in this particular template); we should move towards more consistency. I would support keeping the parameters functional for a few years (but undocumented and with a warning) for those who really are troubled by typing the hyphen, knowing the bot will come by and change them. MB 15:23, 7 March 2021 (UTC)[reply]

  The better solution to "is it "image=" or "Image=" or "photo=" " is simply to support all of them. That way there is no need for any disruption, short-term or otherwise. Thryduulf (talk) 20:31, 7 March 2021 (UTC)[reply]

• Option A—standardization helps simplify code maintenance, and that means more time can be devoted to future improvements. Imzadi 1979 → 17:34, 7 March 2021 (UTC)[reply]
• The question before us is: Should non-hyphenated parameters be fully removed from the CS1/2 family of templates? Yes, absolutely, for most if not all of the reasons enumerated above. Consistent parameter naming should have been implemented c. 2007 when the various, independently-developed, cs1|2 templates were converted to use {{citation/core}} as the common citation rendering engine. In early 2013, en.wiki migrated to the Module:Citation/CS1 suite. Since then, approximately 180 other-language MediaWiki wikis plus some number of non-MediaWiki wikis have also migrated to the module suite. In the time since en.wiki switched to the module suite, we have added new parameter names to support new functionality while at the same time, we have pared away quite a few parameter names because of redundancy, peculiar name-style, non-use, and other reasons. This reduction includes most of the nonhyphenated multiword parameter names so that today, the only remaining nonhyphenated parameter names are |accessdate=, |archivedate=, |archiveurl=, |authorlink= (and its two enumerated forms), and |origyear= (there are 263 basic parameter names and 77 enumerated parameter names). The worldwide adoption of the cs1|2 module suite has caused us to add support for internationalization. Non-English wikis employing the cs1|2 module suite should retain all of the English parameter names because, very often, articles developed at en.wiki are exported and translated to those other languages. That means that a fully implemented module suite at a non-English wiki must support the 340 English parameter names plus 340 local parameter names. It is best, I think, to have a single consistent style for multiword parameter names so that translating editors don't have to learn about or deal with redundant parameter names (this same applies to beginner editors at en.wiki). The cs1|2 templates are complex enough, we don't need to add to that complexity by maintaining lists of synonymous parameter names that don't have semantic meaning (for example, |chapter=, |article=, |entry=, |section=, etc, are treated by cs1|2 as synonyms but, to editors, convey different meanings – the inclusion or omission of a hyphen conveys no meaningful distinction). So yeah, non-hyphenated parameters [should] be fully removed from the CS1/2 family of templates and since we can't go back to 2007 to do what we should have done then, we should do it now, we should do it quickly, and we should get it done. A is my preferred option but, if needs must, then (sigh) B; never that other option. —Trappist the monk (talk) 00:50, 8 March 2021 (UTC)[reply]

  Thanks for weighing in, Trappist; I just wish you had done it a month ago. Your explanation is persuasive, but I still think Option A is silly if we don't at least simultaneously change the documentation to tell users, "don't use that parameter anymore". Having the bots fight against the human editors is silly and inefficient and possibly even dangerous. — JohnFromPinckney (talk) 02:03, 8 March 2021 (UTC)[reply]

  So to summarise Trappist's arguments: It's much easier for developers to only have a few names so we should disrupt editors and thus the wiki to make their lives easier, regardless of the costs (it's been explained at length previously how having two names do the same thing is not at all a problem for editors). Sorry, but that is not persuasive in the slightest: templates exist to support editors not the other way around. Thryduulf (talk) 03:06, 8 March 2021 (UTC)[reply]

  No. Trappist said no such thing. He helpfully explained why this change is necessary. With the possible exception of watchlists, this change does not seriously "disrupt editors" (there's no way having to type in one extra character can be said to be a problem of that magnitude); on the contrary, in the long run it makes life easier for editors, because there is only one simple, easy-to-remember rule: all muti-word parameters use a hyphen to separate the words. On watchlists, see #Worth noting below, which seems to be a promising solution. And if the worst comes to the worst and that solution doesn't work for you, then I sympathise, but at least the "disruption" is only temporary until the bot is finished. --NSH001 (talk) 10:05, 8 March 2021 (UTC)[reply]

  No, the disruption is not temporary. Not only are people likely to continue using the "old" parameters, but removing support for these parameters from the templates means that millions of old revisions will have errors in them instead of simply displaying the references like they used to. Creating errors in the references of old versions of countless pages, so that you only have to maintain 340 instead of 349 parameters (well, not even that, 340 parameters + 9 synonyms)? That seems like a lot of disruption for little gain.

  Fram (talk) 11:13, 8 March 2021 (UTC)[reply

  ]

  Doesn't really matter, it's relatively uncommon to cite old versions of pages, and the error messsages can safely be ignored – only the latest page matters. They also suffer from deleted templates (which can be much more serious for the page as actually displayed to the user) and wikilinks which have gone red because the target page has been deleted. No doubt other things I haven't thought of yet. One just accepts these imperfections. Doesn't seriously "disrupt" anything. --NSH001 (talk) 14:05, 8 March 2021 (UTC)[reply]

  A redlink because the target doesn't exist isn't an issue (it is even an improvement). The others are problems, and knowingly adding much, much more of the same is a serious issue. A version of Salvador Dalí from 5 years ago now already has 10 or so errors: the lang text ones are the most annoying. But the planned defenestration of e.g. accessdate will suddenly add 24 extra errors here. Basically, every edit that MonkBot does for these, will mean one or more errors in every history revision, or millions upon millions of old versions which will become worse, and no current of future versions which will become better.

  Fram (talk) 14:33, 8 March 2021 (UTC)[reply

  ]

  "Basically, every edit that MonkBot does for these, will mean one or more errors in every history revision, or millions upon millions of old versions which will become worse". That's false. Monkbot 18's edits make no difference whatsoever to the display of error messages; in fact it will reduce (drastically) the number of error messages that are eventually displayed when and if the remaining are formally deprecated, where "formally deprecated" means changing the CS1/1 module suite to actually generate the error messages. --NSH001 (talk) 09:37, 11 March 2021 (UTC)[reply]

  Not false, just shorthand. One can see the number of errors that will exist by looking at the number of edits that Monkbot does for this. No, Monkbot doesn't cause the errors directly, the deprecation does; but the two belong together. If option C is chosen, then there will be no "when and if", the remaining will not be formally deprecated, and no error messages will be generated, not in historic versions, not in live versions. None.

  Fram (talk) 10:56, 11 March 2021 (UTC)[reply

  ]

  "One can see the number of errors that will exist by looking at the number of edits that Monkbot does for this." That's very misleading. The number of edits by the bot has nothing whatsoever to do with the errors displayed in the live version (except that, as I've already explained, it reduces them). You're trying to mislead readers here by implying that every single edit by the bot causes an error message, when in fact it only does so for historic versions. Error messages in historic versions don't matter. --NSH001 (talk) 12:00, 11 March 2021 (UTC)[reply]

  No, that's not misleading if one reads the actual conversation, which is about historic versions. "removing support for these parameters from the templates means that millions of old revisions will have errors in them instead of simply displaying the references like they used to." and "Basically, every edit that MonkBot does for these, will mean one or more errors in every history revision, or millions upon millions of old versions which will become worse, and no current of future versions which will become better." That's what you replied to, and which we are still discussing. We can disagree whether errors in old revisions are a problem or not, or whether the maintenance cost of keeping a few synonyms alive is negligible or significant; but please don't start accusing people of "trying to mislead readers" (a bit rich coming from someone claiming that "access-date" is easier to use and more meaningful than "accessdate").

  Fram (talk) 12:12, 11 March 2021 (UTC)[reply

  ]

• Option C And this issue is largely why I am of the current position BAG is not suitable for purpose and needs nuking from orbit. Only in death does duty end (talk) 15:31, 8 March 2021 (UTC)[reply]

  The BAG has a few functions. It does a good job of some of them - notably it's pretty good at ensuring bots do only what their authors intend them to do before they are unleashed. The main issue is with ensuring that only tasks with consensus to be done and consensus for a bot to do the job get approved - I can't recall an instance when it failed to approve something that should be approved, and it does stop the truly egregiously bad ideas from proceeding, but there are multiple instances (including this one) where bar for consensus has been set too low and a small local consensus has allowed a bot without wide community approval to operate. If we have bots then we do need some sort of approvals process, so don't nuke the one we have without coming up with something better first. Thryduulf (talk) 16:54, 8 March 2021 (UTC)[reply]

  "only what their authors intend them to do before they are unleashed." On a technical level this is correct - what BAG doesnt do is any sort of even half-decent job of confirming that bots *continue* to only do the tasks they were initially approved for, or that that approval in the first place has consensus amongst those editors who it will *affect* that the task is wanted or needed, or that there is a clear benefit to those effected by it. The problem with automated editing is rarely the technical aspect, its the business case for it. If we did a review of current active bot tasks (and editors using automated editing to perform mass edits) do you genuinely think they will all be able to point to a discussion that shows a level of consensus proportionate to their effect on the wider editing populace, rather than the walled garden (and the term I would actually prefer to use here as its more accurate is circle-jerk) of 5 or 6 data & machine readable obsessed editors who want it? Only in death does duty end (talk) 08:30, 11 March 2021 (UTC)[reply]

  On the technical vs business-case aspect, I agree, that was much of the point I was trying to make. I also agree that a regular review of bot and automated editing tasks should be undertaken. There are some that I'm sure still have community consensus (e.g. the anti-vandal bots) but I can't say that will hold true for every task. I also agree that there must be an active consensus of editors that will be affected before a task goes ahead - in the case of monkbot almost all affected editors were entirely unaware that it was even proposed. Thryduulf (talk) 14:20, 11 March 2021 (UTC)[reply]

• Option B - I can easily see both sides of this debate. As someone who works in IT, I fully agree that IT should serve the needs of users first wherever possible, rather than simply making life easier for the people who maintain the systems. I see far too many systems that make life harder than the non-digital alternative, which is just ridiculous. There are of course times when things need to be deprecated because they're either incompatible with other things, have security holes, or take up too much development time to maintain. This case is one of the latter, although I'm not sure what extra burden the deprecated (non-hyphenated) paramaters actually add. I would recommend that we formally deprecate the unhyphenated parameters and clean them up with AWB genfixes but I can't support continuing to run Monkbot 18 given the strength of feeling here against such a task. Personally I do not care about 'watchlist spam', but clearly many people do, and we cannot simply dismiss their opinion because "it'd make life easier for us techs". ƒirefly ( t · c ) 07:42, 10 March 2021 (UTC)[reply]

  As I understand it, there's a framework for supporting parameter aliases, which are configured in Module:Citation/CS1/Configuration. As there are many other aliases supported, the framework would remain in place. isaacl (talk) 15:23, 10 March 2021 (UTC)[reply]

• Option A / B per above.  Spy-cicle💥  ^Talk? 18:50, 10 March 2021 (UTC)[reply]
• Option A, strongly oppose Option C. Hyphens should be standard and we should discourage them. I do not mind continuing to support the non-hyphenated version per B, but B also implies the bot isn't free to do its job, so A it is. SportingFlyer T·C 12:57, 11 March 2021 (UTC)[reply]

  Why should one version over the other be discouraged? What benefit to the encyclopaedia does it bring that outweighs the disruption that removing support for a long-standing and well-used feature causes? Why should the bot be free to continue to do a job it didn't have consensus for? Thryduulf (talk) 14:22, 11 March 2021 (UTC)[reply]

  We're trying to gain consensus here, right? I support standardising the reference tags, and I don't think the opponents have good counter-arguments. I'd appreciate if you just let me (and others) have our opinions without the need for commenting - you're not "correct" and we're not "wrong." SportingFlyer T·C 15:01, 11 March 2021 (UTC)[reply]

  I'm not trying to stop people having opinions. I'm trying to get people to explain them and to actually counter arguments left explaining preferences different to their own. Why is standardising reference tags more important than the disruption standardisation causes? Why is a desire to avoid this disruption "not a good counter-argument"? I may or may not be "correct" but unless you can explain why forcing standardisation against the wishes of a very significant number of editors (who will not see any benefit from such standardisation but will experience significant ongoing disruption due to it) is somehow desirable then there is no hope of getting consensus for your views. Thryduulf (talk) 16:02, 11 March 2021 (UTC)[reply]

  You're chiding me for not countering your argument, even though your argument above is simply "I don't see the point." We've been working on this for awhile, it's been discussed on talk pages, it makes it easier to maintain the encyclopaedia, and it's a good idea to finish the project as opposed to having it held back by users who don't like it. I also thought I was just !voting here, I think this is obvious and I don't want to be drawn into a long argument about this, so not only did I not appreciate your response, I'd appreciate it if you left this conversation alone going forward. SportingFlyer T·C 16:49, 11 March 2021 (UTC)[reply]

  As with many of these types of questions, the conclusion each person draws depends a lot on the weighting they give on the relative priorities of different factors. We lay out our lines of reasoning with the tradeoffs, and others can use it as they wish to figure out what tradeoffs they would like to make. isaacl (talk) 17:13, 11 March 2021 (UTC)[reply]

• Option C Unless there is an issue with bots having to read accessdate and access-date as the same thing (and there doesn't appear to be a problem from what has been described, as bots can do this) then I'm not convinced that forcing users who already write accessdate to make an error when they continue to do so after its use is stopped that bots will then flag up for humans to solve is going to be a useful thing to do. Creating a situation which is going to frustrate and demotivate volunteers for no valid reason other than "conformity" doesn't sound like a good plan. Indeed, it was stressed in the original 2014 RfC that "Establishing this uniform parameter name convention does not preclude the existence of any other alias for a parameter, merely that a lowercase, hyphenated version will exist for each parameter." And some of those supporting did so because: "This will significantly reduce editor confusion. They don't have to think about: "Is this the parameter where the words are mushed together, or is it one where they are separated by an '_' ?" Hopefully this will make the templates easier to use." - User:Makyen. What we should be looking at is making bots read the varying ways that users may write a word or phrase in a template. I hate it when, for example, I use a capital letter by mistake, and the template doesn't work, and I have to work out what the problem is. I don't follow how frustrating, demotivating and alienating volunteers by changing how a template works and so creating problems for them will "decrease the maintenance burden" on those users. It is sometimes things like functionality being changed, that drives users away. SilkTork (talk) 01:09, 12 March 2021 (UTC)[reply]
• Option C. I haven't checked whether this move was done under a six-year old RFC with seven people, or under a seven-year old RFC with six people. But no one disputed that it was an at least six years old RFC involving at most seven people. Now, it seems that there are a large number of users who disagree. Perhaps telling them they "don't see the larger picture" will be enough to silence the dissenters. Maybe the "Visual Editor reception" will happen again. Who knows the future? Pldx1 (talk) 15:07, 12 March 2021 (UTC)[reply]
• Option A. I do not see any real argument for why citation templates should be a mish-mash of two different parameter names. It's an eyesore, it's a pain in the ass, and it brings zero utility. There are a couple ways of fixing it: for the new parameter names to be integrated into every tool, and every adjacent task, and every automated editing tool, and then maybe in ten years 90% of them will be gone (but the remaining 10% will be scattered willy-nilly across the entire project and impossible to hunt down). Or we could just have there be one short run and be done with it. I, for one, would be glad for the issue to be concluded. jp×g 05:33, 15 March 2021 (UTC)[reply]

  Have you actually ready any of the comments left by others? There are at least half a dozen different reasons given for the utility of multiple forms of parameteres. Even if option A is selected, disruption will not be short term but will continue for years for the reasons explained multiple times elsewhere on this page. Thryduulf (talk) 12:27, 15 March 2021 (UTC)[reply]

• Option A. Over time, we evolve different conventions. Once we have done so, we need to clean up the uses of the old conventions. Option B will be too slow and leave this issue dragging on for many more years. So, firstly ensure that all documentation is unambiguous about using only the new names. Secondly, make sure that all editing tools use only the correct parameter names and all genfix rules are up to date. Wait for the bot to complete a full pass and then treat the old parameter names as solid errors — GhostInTheMachine ^{talk to me} 15:36, 15 March 2021 (UTC)[reply]
• Option C. The encyclopedia should be run for the benefit of readers, and subject to that, for the benefit of editors. Creating busywork and random conventions for the sake of it is not useful. Stifle (talk) 10:19, 16 March 2021 (UTC)[reply]
• Comment: The argument made here and elsewhere about how much easier things would be for template- and module-writers under Option A, is a completely misguided and arrogant one. It seeks to benefit a small subgroup who are neither the consumers of the encyclopedia, nor its principal creators. Namely, it calls for optimizing ease of work and convenience for the very small number of template/module writers, rather than optimizing for (the much larger number of) article editors. The trade-off must benefit the largest number; if you make it harder for editors, then the article readers will suffer, and they are the largest group of all and the reason the encyclopedia exists. This should be axiomatic, but the encyclopedia isn't here for the convenience of template and module writers. Personally, I'm happy writing and improving templates and breaking my head against squirrely, confusing, and sometimes infuriating template problems, just to make things slightly easier for editors. If it's too inconvenient or too much work for template/module-writers, then they should abstain from "helping", and just improve article content, rather than seek to make their own lives easier. Mathglot (talk) 23:42, 20 March 2021 (UTC)[reply]

  Thryduulf said it better (and briefer) than I, here: "Those who do care about template mechanics should always be acting in ways that actively put readers first, editors second and programmers third." Precisely. Mathglot (talk) 00:04, 21 March 2021 (UTC)[reply]

• Option C (preferred) or Option B (seoncd choice) -Beyond My Ken (talk) 03:06, 23 March 2021 (UTC)[reply]

Discussion (CS1)

Some suggestions regarding the options:

• Pedantry: "Non-hyphenated parameters" should read "Non-hyphenated multi-word parameters" in all three options. Parameters that contain only a single word or acronym do not need a hyphen.
• In Option C, there is no point in suggesting that "the deprecated parameter list will need to be updated to remove the non-hyphenated parameters"; there are no instances of those deprecated parameters left in the affected namespaces, so support for them will be removed shortly, just as support has been removed for dozens of unhyphenated multi-word parameters already.

Some history and a status update, for those here on VPP unfamiliar with the long history of updates and changes to the Citation Style 1 templates ({{cite web}} and its siblings): As far as I can tell, there are only six unhyphenated multi-word parameters left – |accessdate=, |airdate=, |archivedate=, |archiveurl=, |authorlink=, and |origyear= – out of an original population of many dozens. So far, through the work of scores of editors and bots over the seven years since we standardized on hyphenation of multi-word parameters, we have deprecated, removed from pages in affected namespaces, and then removed from the CS1 templates themselves (as WhatamIdoing suggests above), many dozens of different unhyphenated multi-word parameters in CS1/CS2 templates. All new multi-word parameters during that time have been introduced using only a hyphenated form. This RFC is essentially asking: should we finish the job, or leave it at over 90% done, in a sort of limbo state, with six parameters as exceptions to the overall pattern, just because those parameters are used in a lot of articles? – Jonesey95 (talk) 04:38, 11 February 2021 (UTC)[reply]

• Problem with "Option B": Option B is not the status quo. If it were, I'd be much happier. Since it's not, I don't know whether to !vote for A or B. The documentation at Cite web, for example, makes no mention of accessdate being deprecated. This means that users will be quite justified in blissfully adding and readding templates with accessdate, even after the bot has come through already and changed accessdate to access-date in 20 places. Each iteration tends to address fewer instances, but each run is a separate entry in my watchlist. Watchlist entries seem to be part of the complaint against this kind of work, so the first step should be turning off the faucet at the source, before spending energy to, um, bail out the boat. — JohnFromPinckney (talk) 06:37, 11 February 2021 (UTC)[reply]
  • The first sentence is true; all but six of the unhyphenated multi-word parameters have been formally deprecated and removed. As for "turning off the faucet at the source", that would be Option A, but in the past, when we have made red error messages appear in large numbers of articles as a first step in standardizing the citation templates and noting errors in them, there has been significant pushback. In order to avoid that pushback, the bot was acting to fix 90+% of the non-standard parameters before turning on the error messages, but that work has been stopped as well. Option A will allow that work to be completed. – Jonesey95 (talk) 15:09, 11 February 2021 (UTC)[reply]
    • I feel we are not understanding each other. "Deprecation" involves communication as a first step; removal is a second, later step. If we are going to have a bot changing pages (e.g., accessdate to access-date), causing some distress to the populace (and this is the status quo), then we should at least change the documentation to say that accessdate is deprecated, and is not a usable alias. I am strongly against a bot going around to change parameters to the "good" names when we don't ever tell the humans they shouldn't use the "bad" ones. That's just an endless cycle of watchlist-cluttering edits causing great irritation. If you want to avoid pushback, tell the people not to use the unhyphenated versions, then run the bot, then remove support some time later. — JohnFromPinckney (talk) 15:41, 11 February 2021 (UTC)[reply]
      • Deprecate in terms of the CS1 templates means to emit a red error message indicating deprecation. When we emit any red error messages for parameters in CS1 that are used often (or lack thereof for parameters that should be used more often), a lot of people get very irate. We do not change the documentation to indicate deprecation until after the module begins telling people inline about deprecation. So yes, you are not understanding each other, but it's a question of terms of art. --Izno (talk) 16:44, 11 February 2021 (UTC)[reply]
        • Why do people get irate? Because suddenly millions of readers go from nothing being wrong to seeing a load of red error messages all over hundreds of thousands of articles (and you couldn't seriously expect someone to debug a CS1 error as their first contribution). These citation templates aren't for us. They're for the reader and they need to be functional with low rate of error messaging 24/7 with no exception, because even small amounts of downtime have significant reputational damage. — Bilorv (talk) 02:54, 15 February 2021 (UTC)[reply]
          • I would be more sympathetic to such an argument of "significant reputational damage" if there weren't over 100,000 pages with active errors on them (and that's ignoring the hidden Category:CS1 errors: missing periodical) and another 250k in its sister (also hidden). --Izno (talk) 20:22, 15 February 2021 (UTC)[reply]
    • That "we" (who is that, some class of super-techies whose opinions count for much more than those of us "normal" editors?) get significant pushback when creating error messages is simply a demonstration that the wider community does not agree with what "we" have decided. The answer is to stop doing what you are doing, not to do it more stealthily.
      Phil Bridger (talk) 18:37, 11 February 2021 (UTC)[reply
      ]
      • You are welcome to participate at
        Help talk:CS1, the same as any other editor. Decisions are made by consensus there, inline with how consensus is practiced everywhere else on wiki. Don't like a decision "we" made? Get involved. --Izno (talk) 18:42, 11 February 2021 (UTC)[reply
        ]
        • I saw absolutely no discussion of the removal of the freetext editors field on the talk page. Where did that happen? I don't think it's coincidence that all the editors whose names I know well from content creation/editing are voting B/C and all the editors voting A are those I don't recognise at all. Espresso Addict (talk) 20:08, 11 February 2021 (UTC)[reply]
          • The difference is between those people who think this is an encyclopedia and those who think that it's a playground for techies.
            Phil Bridger (talk) 20:43, 11 February 2021 (UTC)[reply
            ]
            • Your ad hominem has no welcome here. Move along if that's the best you can offer to the discussion. --Izno (talk) 20:47, 11 February 2021 (UTC)[reply]
              • That is just the kind of reply that people who are here to create an encyclopedia tend to get from the techies. No response to genuine concerns but just an order to "move along". I have no wish to monitor whatever pages are used to ignore end users, but I have a right to expect that any decisions taken will respect the interests of everyone, not just a self-appointed clique of people who "know" what is right.
                Phil Bridger (talk) 19:53, 15 February 2021 (UTC)[reply
                ]
                • It's the kind of reply to people who needlessly create category A and category B and then line themselves up in one or the other categories. If you (specific/personal) don't want to monitor whatever pages, that's your prerogative. Do know that it is your choice and you are responsible for that choice, and choices have consequences. Changes are always announced ahead of time and consensus is sought for non-obvious changes (and even obvious changes with non-obvious implementations), so you have no excuse not to tune in at least once every couple months when the regular "Shit is Changing" post gets made. Secondarily, the scare quotes are not indicated by this discussion, nor any prior discussion that I can see, whatsoever. I have not seen any such 'clique' nor any of the users who would prospectively be in such a 'clique' claim they "know" what is right. As I said, if you have nothing to contribute but smears and attacks and divisiveness, move on. --Izno (talk) 20:22, 15 February 2021 (UTC)[reply]
          • In reverse chronological order: the patch notes indicating removal, the removal discussion, the patch notes indicating deprecation, the deprecation discussion. 4 times mentioned over a period of half a year on that talk page, a pre-existing maintenance category indicating a soft deprecation, and my removal of over 4k instances of the parameter over the year and a half preceding that deprecation discussion. Basically by hand (my hand, as it happens). --Izno (talk) 20:45, 11 February 2021 (UTC)[reply]
          • As for I don't think it's coincidence that all the editors whose names I know well from content creation/editing are voting B/C and all the editors voting A are those I don't recognise at all., I invite you the same as I invited Phil Bridger. You may watch and edit
            Help talk:CS1 at any time, the same as me. "I don't recognize these people" is a trash association to make and is the same kind of ad hominem that I have asked Phil so kindly to stop employing. It is certainly not sufficient cause to say "I don't like this change". --Izno (talk) 20:49, 11 February 2021 (UTC)[reply
            ]
            • I think there's a genuine problem here that there's a disconnect between the editors maintaining the citation templates and the editors employing them to write and maintain articles. I didn't make the decision to abandon years of using citation templates (CS2 actually, but the same parameter changes are happening there too, even less announced) lightly. Espresso Addict (talk) 21:35, 11 February 2021 (UTC)[reply]

              The disconnect is real and significant. Someone whose skills and interest lie in writing or maintaining article content shouldn't need to care about what happens at pages like

              Help talk:CS1 (and how on earth is a new editor even meant to know that page exists?), let alone have to pay attention to discussions there. Those who do care about template mechanics should always be acting in ways that actively put readers first, editors second and programmers third. The only time there should ever be breaking changes or a need for cosmetic edits is when after detailed examination there are literally no alternatives available. We've ended up here because that hasn't been happening and a local consensus has ignored the needs of end users. Thryduulf (talk) 01:38, 12 February 2021 (UTC)[reply

              ]

              I almost commented last night, and then put that version into a text file and went to bed. Now I have been spurred by a comment above to reply here. The editors using these templates to write and maintain articles are the same as the editors maintaining the citation templates. Get that through your head. If we prioritize different qualities versus this other supposed separate group, it's because we have the experience to do so. If you don't, let me reiterate, get involved. Come and say "I don't like this" or "this is a painful interaction" or X, Y, and Z, and provide a suggested fix. That's how consensus starts forming, like every other page or process on this website. Others will say "I don't want this to work like that suggested fix because A, B, and C". Then you discuss the tradeoffs and make a decision. If you're unwilling to step up and discuss the paper cuts, then we end up having mass RFCs or AN drama or what have you over what would otherwise be small issues or ones that could have been discussed informally with the ordinary "let's figure it out" consensus view of the world. That's a disconnect too, and blaming editors interested in one page versus those apparently disinterested is not the right way to move forward. Want something to be better? Ask. Propose. Cajole. Make the effort to put forth the minorest of social interactions required to start a discussion in the one place where everything about that thing is discussed. We're all volunteers and our heart is in just a right a place as yours, but if we should have disagreements, then we talk to each other and find out how to fix the problem or agree to disagree on the points of interest. Not have constant complaints of "they didn't listen to me". Consensus is not unanimity.

              The only time there should ever be breaking changes or a need for cosmetic edits is when after detailed examination there are literally no alternatives available. This is quite frankly an opinion lacking any community consensus whatsoever. We've recently approved an RFC allowing cosmetic edits on a regular basis and from what I could see in that discussion (and murmurings elsewhere), cosmetic edits might be closer to having consensus than not, it's just inertia that leaves us not performing them regularly. Moreover, hundreds of templates, and certainly all those which go through

              WP:TFD, have a process applied which causes breaking changes or which results in more-or-less cosmetic edits. Are you claiming that TFD does not have consensus as a process? That templates which are being cleaned up because of a talk page discussion on that template's talk page should be stopped? You want your cake (to not care about how things work) and eat it too (have all of your opinions and claims listened to, when they aren't even articulated in the first place). Get real. --Izno (talk) 20:22, 15 February 2021 (UTC)[reply

              ]

              I fear this just demonstrates the truth of what Thryduulf writes. "The editors using these templates to write and maintain articles are the same as the editors maintaining the citation templates. Get that through your head." is self-evidently false, as well as impolite. Speaking only for myself, what I actually want is to be able to get on with writing & improving articles, rather than having long side discussions on matters that aren't directly relevant. Espresso Addict (talk) 00:10, 16 February 2021 (UTC)[reply]

              get on with writing & improving articles Then do that. No-one is stopping you from not caring. I am just calling you hypocritical for raising a fuss when you don't and then changes happen elsewhere that affect you. Don't like it? Change your behavior. (I won't reply to you again.) --Izno (talk) 00:34, 16 February 2021 (UTC)[reply]

              The editors using these templates to write and maintain articles are the same as the editors maintaining the citation templates. Get that through your head. Both uncivil and not even close to correct. While it is possible that some, maybe even many, of the editors maintaining templates also write and maintain articles there are literally hundreds of editors writing and maintaining articles who have never been near a discussion about the template code, let alone written any code. Writing encyclopaedic prose and writing computer code are very different skills; nobody is or should be required to do both. However given that the purpose of this project is to write an encyclopaedia everything that is not writing an encyclopaedia should be done for the benefit of (first and foremost) readers of the encyclopaedia and (secondly) those who write and maintain the encyclopaedia. The convenience of those dealing with tools is least important. Thryduulf (talk) 00:23, 16 February 2021 (UTC)[reply]

              Writing encyclopaedic prose and writing computer code is a false dichotomy and a blatant misrepresentation of what the two paragraphs I had to say. I did not ask you to do both. Nor do I serve you (general) in your supposed role as an article writer. There are no (formal) hierarchies on this encyclopedia, and even considering yourself as supposedly above me or my efforts is the actual uncivil statement. Lastly, I place myself fully in both supposed groups given the thousands of articles where I have bettered their citations. (And as I said to Espresso, I will not reply in this section again.) --Izno (talk) 00:34, 16 February 2021 (UTC)[reply]

              My role here is not as an article writer (I do very little of that), I spend the majority of my time on the project trying to ensure that readers can find the content they are looking for and those that do write and improve articles without needing to worry about nonsense like this that will needlessly make their job harder. There might not be a formal hierarchy but their should be: (1) Readers are head-and-shoulders above anyone else. (2) Those who write and maintain the encyclopaedia a short way above (3) Those who support those write and maintain the encyclopaedia are a long way above (4) Those who hinder any of the above. I place myself in the third category alongside many of those who maintain templates. Thryduulf (talk) 21:02, 16 February 2021 (UTC)[reply]

• Comment. What is the merit of citation templates? Let alone the increasing creep to make them more and more rigid and harder and harder to use? The only reason I can see is to make it easier to export and sell data. No-one ever seems to give a thought to those who type citations manually; it is far harder to type "access-date" (which requires two hands and a hand movement) than "accessdate" (one hand, no move). I don't understand what the benefit of this change is at all. In fact, broadening this discussion, I don't understand why the freetext editors field was suddenly withdrawn this January. Personally I've decided to meet the latter change by reverting to writing out references by hand, which is easier, more flexible, and seems to have no downsides. Espresso Addict (talk) 07:04, 11 February 2021 (UTC)[reply]
  • Citation templates make standardizing citation style and look much easier. They can, for example, automatically standardize date display. Machine-readability is also a good thing, imo. I suppose access-date is slightly harder to type - but editors editing wikitext manually would probably not mind. Otherwise, tools for inserting these citations exist.
    talk | contribs) 17:12, 11 February 2021 (UTC)[reply
    ]
    • It's not slightly harder to type, it's a great deal harder to type for a touch typist. And here's an editor editing wikitext manually who does mind, enough to bother responding here. There's no tool I know of that creates citation template code in the form I prefer for ease of wikitext editing afterwards; they all make code soup that takes longer to fix than retype. Espresso Addict (talk) 20:08, 11 February 2021 (UTC)[reply]

      A bit slower, yes, but I personally don't find it a great deal harder—touch typists generally have practised it a lot while learning. Nonetheless, I don't think ease of typing by some metric is the primary issue. isaacl (talk) 20:40, 11 February 2021 (UTC)[reply]

      Speaking as one of those touch-typists, and as a person who learned to type on an actual typewriter, I don't find the hyphenated version any harder to type, and I do recall my typing teacher saying that it was faster and easier to type words that were split evenly between hands, instead of all in one hand. WhatamIdoing (talk) 22:49, 11 February 2021 (UTC)[reply]

      Yeah, personally I imagine it has a greater effect for a hunt-and-peck typist. When I said a bit slower, I was thinking compared with a word of equal length that consisted only of letters. Of course, in this case, the hyphenated name has one more character which would comprise most of the difference for me. isaacl (talk) 23:02, 11 February 2021 (UTC)[reply]

      For me, not an issue on a regular keyboard, but a bit of a pain on a tablet. Then again, so much is a bit of a pain on a tablet... · · · Peter Southwood ^(talk): 05:39, 2 March 2021 (UTC)[reply]

    • I use citation templates myself, but I respect the wishes of those who prefer not to use them. It is precisely the fact that I use them that leads me to the opinion that obvious synonyms for parameters should not be removed, as proposed here. What on Earth is the problem with allowing both hyphenated or unhyphenated forms? All it means is a few extra bytes of storage, and no extra code if it is done properly. And the work has already been done, but this proposal is to undo it. Do we still live in the days when I started in IT working for one of the world's leading business information companies whose UK operations were all run from a computer with 1MB of memory and where all online programs were limited to 12KB? I thought we had moved on from there.
      Phil Bridger (talk) 20:34, 11 February 2021 (UTC)[reply
      ]

      I am reminded of a story about a lady, who (decades ago) bought a large supply of personalized stationery and then discovered that the post office was renaming her street. She convinced the local postmaster into letting her continue to use the old address until her stationery was used up. This saved her some money and effort, but it had external costs. For many years to come, every mail carrier had to learn that "123 Main Street" wasn't on Main Street and wasn't number 123, but instead had to be delivered to the other side of town.

      Aliases are often low-cost in storage and computational terms, but they are not actually free. "Not free" can add up to a quite significant cost when it is repeated a million times. WhatamIdoing (talk) 22:56, 11 February 2021 (UTC)[reply]

• Although I sympathize with the desire for all templates to align with one standard (from a user's perspective, I would personally find it easier), I just don't think it's feasible at this point in time. In which case, I sympathize with those who think computers should make our lives easier, and just accept both formats. On the third hand, from an implementer's perspective, I appreciate that it adds a lot of noise to template syntax, if not implemented with a Lua module. Since the CS1-based templates are implemented with Lua, the overhead can be minimized, and so I think both formats should continue to be supported. I don't feel there is much advantage to converting en masse, by any mechanism. isaacl (talk) 20:40, 11 February 2021 (UTC)[reply]
• Pinging some previous participants in related conversations who may not yet be aware of this discussion: @
  Gonzo fan2007, Sariel Xilo, Modest Genius, and SlimVirgin:. Nikkimaria (talk) 01:17, 12 February 2021 (UTC)[reply
  ]

  Thanks for the ping, I was indeed unaware of this discussion. Modest Genius ^talk 12:16, 12 February 2021 (UTC)[reply]

• Does anyone have any evidence that having alias for parameters makes things harder for end users? In my experience as a trainer and long-time user, having multiple ways to achieve the same ends allows editors to spend their time and energy working on content rather than puzzling over making sure that the template parameters are named in exactly the right way. I've never met anybody who was comfortable enough to be dealing with templates in the first place who was not completely comfortable with the idea that "you can write it as either access-date or accessdate, it doesn't make a difference." (or similar). Speaking personally, I don't want to have to learn whether it is "accessdate", "access-date", "access_date" or "access date", I just want them all to be accepted so that whichever I input the template works and I can worry about more important things. Thryduulf (talk) 01:29, 12 February 2021 (UTC)[reply]

  The inconsistency across all templates does make things harder. Users have to either remember which templates only support one form and which one, or look it up each time. I appreciate though that there is extra complexity in trying to support lots of aliases, particularly for ones that aren't implemented with a Lua module (either each use is going to become more elaborate and verbose, or the template could delegate the detailed implementation to a helper template, with the top-level one only doing parameter normalization), so personally I wouldn't want to require that every template must support aliasing. Thus I think we're kind of stuck with inconsistency. isaacl (talk) 02:12, 12 February 2021 (UTC)[reply]

  This inconsistency should be tolerated. Some, particularly high-use, templates having aliases is important for the great number of people who use these templates. They don't have to remember was it |access-date= or |accessdate=. If it doesn't work on some other template, well fine. But eliminating aliases and simply making the computer say no on all templates makes things a lot harder for the vast majority of users. That inconvenience is far greater than the fact that a few lesser-used templates don't support aliases and you might need to take a look at the documentation sometimes. – Finnusertop (talk ⋅ contribs) 06:05, 12 February 2021 (UTC)[reply]

  Yes, as I said in another comment, I feel both forms should continue to be supported for the CS1-based templates. I'm pretty sure the vast majority of templates don't support both hyphenated and non-hyphenated parameters—for example, from what I can tell from the code and a quick test, {{Infobox}} doesn't. That's just the way it goes when trying to make as many aspects of Wikipedia markup accessible to as many editors as possible: standardization won't always happen, and often simpler markup will be preferred by more editors than more complex markup, even if it would add more functionality for users. isaacl (talk) 22:22, 12 February 2021 (UTC)[reply]

• In Option C, there is no point in suggesting that "the deprecated parameter list will need to be updated to remove the non-hyphenated parameters"; there are no instances of those deprecated parameters left in the affected namespaces, so support for them will be removed shortly, just as support has been removed for dozens of unhyphenated multi-word parameters already. This argument directly goes against
  WP:FAITACCOMPLI. Also, stop referring to them as "depreciated" - it is clear that there was insufficient consensus to declare them depreciated in the first place. I can understand that it's frustrating to have something you thought was decided fall apart like this, but it is extremely important that sweeping changes get more discussion before being implemented; the only way we can reasonably encourage that is by refusing to allow things like Monkbot's use here to determine policy, ie. it is necessary to make the work it did moot, or people will have a constant incentive to take the easy route and avoid bothering with time-consuming, often-frustrating, and unpredictable (but necessary) discussions before a change of this magnitude. That means, yes, allowing and even encouraging people to resume using non-hyphenated parameters even though you thought you were finally done with them; if you want to avoid that happening next time, seek larger-scale discussions first before making large-scale changes to avoid unexpected blowback if it turns out you don't have the consensus you thought you did. No matter how well-intentioned this was, we absolutely cannot risk rewarding the use of a tool to make a wide-spread change without sufficient consensus - which means, yes, as painful as it is, very deliberately taking the wrench to the kneecaps of the intended improvement this change was meant to establish, and intentionally ruining it even after the "cost' has been paid. It's not ideal, but it shows why it's so important to have broad discussions involving many users before making wide-spread changes; having policy effectively set by bots performing mass edits and establishing things as fait accompli is just not acceptable. I can sympathize with the amount of work that went into this that will be wasted as a result, but reaching a clear, unequivocal consensus involving a large number of editors should have been the first step for something of this magnitude, so you wouldn't suddenly run headlong into an RFC like this and find the consensus wasn't what you thought it was. --Aquillion (talk) 10:07, 4 March 2021 (UTC)[reply
  ]

• Thanks to Nikkimaria for the ping based on my earlier involvement. I'm appallingly neutral (or torn) on the main topic but I earlier did some analysis that gives me concern about impacts on Template space in particular, if deprecation goes ahead fully. Here I'll use |authorlink= as a proxy for all six, because I personally type it more often. There are three areas that concern me: templates that indirectly invoke CS1, those that transclude CS1-invoking templates, and clones. I'm concerned about where and when red error messages appear, and documentation, in which I include both /doc and TemplateData.

1. There are many CS1-friendly templates that invoke one of the canonical templates; {{Cite encyclopedia}} is often used for example. Some use parameter rewriting, converting |authorlink= to |author-link= before passing it on.
   1. Does their documentation always give precedence to |author-link=? I think we caught them all during the earlier discussion, but I can't guarantee the quality of the search.
   2. Because the CS1 code never sees the deprecated parameter, should this template emit a red error itself during the substitution? Or maybe just forget about the mapping and have CS1 do it? And how hard is it to find templates with the parameter substitution code?
   3. If |authorlink= ever moves from deprecated to invalid, then all those mappings and documentation need to be trimmed in one fell swoop. Well, I guess the mappings can stay in place although they could trip up inattentive longtime users.
2. There are templates that simply transclude a pre-filled CS1-style template: for example embedding {{Cite book}} as a citation to a specific volume that is relevant to any use of this template. Have they all been fixed? If not, their user will see a red error message (correct?) that has nothing to do with their own usage.
3. There may be templates that are modeled on CS1 but roll their own expansion, although I don't know of any such. If they happen to use |authorlink=, should they also be brought in line?

Two final comments: there are more than six, if you consider variants like |author1link= and |authorlink1=. And in principle the arguments above apply to any page outside Template space that gets transcluded by someone, although I know that feature is rarely used. David Brooks (talk) 16:51, 12 February 2021 (UTC)[reply]

Shoulda said: the above applies to A and to some extent B but you can probably figure that out. David Brooks (talk) 16:57, 12 February 2021 (UTC)[reply]

DavidBrooks: There is a tiny army of editors at the ready to fix these templates. As far as I know, there are no transcludable templates that pass |authorlink=, your example, to any CS1 templates, so using those templates would not generate an error message. |accessdate= and the two or three remaining unhyphenated parameters being passed from template transclusions to CS1 templates were being processed by the bot before the bot was paused to hold this discussion. Once this discussion is closed, the bot will be able to resume that work, with human assistance, if the closure is a reasonable one. – Jonesey95 (talk) 17:13, 21 February 2021 (UTC)[reply]

@

Citeer web}}, {{Alox2}}. But this less restrictive query shows a few more (ignore the "DYK" archive, I think). Not sure if you meant this type of usage in your comment; I may be missing the context. David Brooks (talk) 05:05, 22 February 2021 (UTC)[reply

]

@

Citeer web}} still need to be fixed (I'll do them later today). The documentation of "Cite book (short)" in {{Quicktemplates}} lists the not-incorrect |authorlink=, so that comes under the "prefer the hyphen forms in documentation" rule. I didn't look for |authorlink2= or |author2link= because they are unlikely to appear alone. David Brooks (talk) 19:17, 22 February 2021 (UTC) ... Y, also {{Bach's compositions (sources)}}, which indeed had |author2link =. David Brooks (talk) 21:42, 22 February 2021 (UTC)[reply

]

• I would like to clarify what I meant when I originally wrote the options, since there is some confusion. My intention with "removal" was to refer only to removal of usage of nonhyphenated parameters by transclusions of the template, not removal from the implementation of the template itself (I will call this "disabling the parameter"). This is also distinct from deprecation, which is designating something in documentation and warning messages as discouraged. Remember that the original reason for this RfC is to clarify whether we should have a bot going through the article space removing the parameter usage as its sole action. A, B, and C are respectively continue removing now, remove only as part of other non-cosmetic edits or
  in a situation where cosmetic-only edits are allowed, or do not remove. In case of A or B, I did not intend for them to make a statement about whether we should disable the parameter when this is done. This is an important question, but orthogonal to whether the parameter usage is removed now or later. — The Earwig _alt ⟨talk⟩ 22:11, 13 February 2021 (UTC)[reply
  ]

  Thanks for clarifying (except the part where you formulated the options). Unfortunately, it now means I'm missing the option: Non-hyphenated parameters should be deprecated now and removed later. The actual status quo seems to be: Non-hyphenated parameters should be removed now and deprecated later, and the removal can occur by bot which does nothing else on the page (cosmetic only), but will revisit the page as often as necessary to re-remove the non-deprecated params that keep getting added. — JohnFromPinckney (talk) 23:01, 13 February 2021 (UTC)[reply]

  For context, this is how I originally phrased it; note B and C are swapped. Isn’t what you’re desiring exactly option B (in the current formulation)? It seems describing B as the "status quo" is confusing. Instead, view A as what the bot was doing before it was stopped, and B as what is currently happening with the bot stopped, but with clear, formal deprecation. — The Earwig _alt ⟨talk⟩ 00:19, 14 February 2021 (UTC)[reply]

  Mmm, maybe, and in any case I'm very grateful for your link to the pre-RfC coordination at Primefac's Talk. And I can say that your original formulation was clearer than the formulation we're now using for the RfC proper. However:

  1. There appears to be no consensus for the removal of non-hyphenated multiword parameters. The 2014 RfC determined only that the hyphenated version must exist, and the close explicitly allowed for non-hyphenates.
  2. If this RfC is intended to determine whether non-hyphenates should be deprecated, it's poorly formulated to the extent that it mentions the current bot activities.
  3. I can't view Option A as what the bot was doing before it was stopped, because the parameter has not been deprecated. (The statement by Nikkimaria that deprecation "in the context of CS1/CS2" means a red maintenance message will be shown is not something I can accept, as any reasonable software provider knows that advance notice is the first part of deprecation. Unfortunately, I don't usually work "in the context of CS1/CS2", so haven't argued before against this unhappy definition.)
  4. Option B, as written on this page, is a garbled mess, not only because of the "status quo" mention, but also because of the "Deprecation can be bundled into genfixes..." bit. Deprecation, as above, is (first) a documentation task, followed by a (presumably small) step to cause red messages to be generated (I'm not sure what's involved here). I don't see what would need to be "bundled into genfixes". I had the feeling that many !voting here saw "removal" as meaning elimination of usages, but maybe that's due to my own confusion. Your Option C (and your original formulations in general) were much clearer.

  There should never have been any bot activities, because (a) there's no consensus, yet, to deprecate non-hyphenates, (b) the params are not yet deprecated, and (c) the bots' edits have been largely accessdate => access-date only, so violate

  WP:COSMETICBOT. So I guess I'll be !voting for B, although I'd much rather choose your original Option C. This RfC as written has been too unclear (at least for me). — JohnFromPinckney (talk) 11:20, 14 February 2021 (UTC)[reply

  ]

  Re The 2014 RfC determined only that the hyphenated version must exist: As called out in the proposal, it also said The documentation is to show this lowercase, hyphenated version as the one for "normal use". Hence my comment above: some of us recently tweaked the documentation of a few high-use templates to make the hyphenated version the privileged one, but I can't guarantee we got both /doc and TemplateData for every template that indirectly uses CS1/2. If we end up with both hyphen and no-hyphen equally valid (is that C?), then tweaking the documentation will be the only change visible to current editors. SHould there be a more valiant effort to track them all down? David Brooks (talk) 19:45, 14 February 2021 (UTC)[reply]

  If the consensus is for Option C, then nobody has to track down anything; we can leave the documentation as it is. BTW, the claim written there, This will also mean that the deprecated parameter list will need to be updated to remove the non-hyphenated parameters is another red herring and would, it seems, not apply at all (as accessdate isn't even listed there yet).

  If we chose Options A or B, then yes, the documentation should be immediately changed. I can't believe it will take too much valor to find the Template:Cite web documentation, as it doesn't seem terrible exotic, but it should be included in the tweaking in those cases. — JohnFromPinckney (talk) 20:07, 14 February 2021 (UTC)[reply]

  • The 2014 RFC had a participation of only seven people and was six or seven years ago. It is patiently absurd to update documentation with such a sweeping change based on that alone, and any such changes ought to be reverted pending the outcome of a more clear RFC. Furthermore, the 2014 RFC specifically indicated that nothing would be depreciated, so if you are relying on it as a justification for any changes then obviously no non-hyphenated versions can be depreciated until / unless a new RFC overturns it. --Aquillion (talk) 22:19, 15 February 2021 (UTC)[reply]

  Please read the summary at the top of this Discussion section. Here's the summary of that summary: Dozens of unhyphenated multi-word parameters have been deprecated, removed from pages, and then removed from the Citation Style 1 templates over the last seven years. There are only six left. During those seven years, many new, hyphenated multi-word parameters have been introduced, without unhyphenated aliases. At present, the situation is that unhyphenated multi-word parameters are the standard, and there is just a bit more work to do to remove the final six outliers. Unfortunately, it appears that many editors have not enabled the useful settings "Expand watchlist to show all changes, not just the most recent" and "Group changes by page in recent changes and watchlist" so that bot edits can be hidden without losing visibility of bad human edits, so people are complaining about a bot "clogging" their watchlists. – Jonesey95 (talk) 23:59, 15 February 2021 (UTC)[reply]

  None of that is however relevant to Aquillion's comment in the slightest. That a flimsy consensus (at best) has been used to justify removing functionality in the past does not mean that that removal was a good thing or that the bot edits (which clog both watchlists and page histories with unnecessary edits, whether they hide other edits or not) should continue. Indeed, I strongly suspect there would be support for re-enabling the parameters already removed. Thryduulf (talk) 00:27, 16 February 2021 (UTC)[reply]

Please note, folks, that this bot is a one-off; it may be processing some 2 or 3 million pages, but it's still a one-off, that is (unless reverted) it will only ever appear ONCE in any article history. So the idea that it's "clogging up" page histories is a big red herring. If you're bothered about it "clogging up" your watchlist, then please set the options recommended by Jonesey95 and others. So that objection falls by the wayside too. Please also note that this bot finishes the job. Once it's done, that's it – no more bots making this sort of change. To the charge that this bot is "disruptive" or that it's somehow "removing functionality", I can't do better than copy Rexx's observation above: "Once we have standardised on hyphenated parameters, future editors will look back and think how lame and time-wasting this sort of debate is." --NSH001 (talk) 19:30, 21 February 2021 (UTC)[reply]

Except that, no, that's completely inaccurate. Consider List of University of Pennsylvania people. When I look at the last 500 edits just now, I see that Monkbot has been there SEVEN times: here (21:07, 19 October 2020) and here (01:02, 28 December 2020) and here (17:21, 14 January 2021) and here (21:21, 16 January 2021) and here (01:53, 18 January 2021) and here (15:30, 18 January 2021) and here (20:54, 30 January 2021). While that first edit (from October) did nothing with |accessdate=, the others all did, as often as it found new additions to the article. Note that the fifth and sixth edits both occurred on the same day.

I don't ordinarily block bot edits from my watchlist because I want to know what's going on with "my" articles, and the bots generally do useful and interesting work. It's just that the (to me) sudden and unforeseen flurry of multitudinous edits (doing, it appears, nothing which really interests me) was quite irritating. The repetition on List of University of Pennsylvania people really got my blood pressure up, and that's not far off from "disruptive" to me. — JohnFromPinckney (talk) 21:15, 21 February 2021 (UTC)[reply]

Hmm... The October edit is Monkbot 17 (not 18), so doesn't fall into the scope of this RfC. The edit on 28 December is the main application of this bot. That leaves 5 (mostly quite small) unexpected additional edits by the bot. They are all caused by editors adding parms that don't conform to the canonical standard. So yes, I should have qualified my statement by allowing for that possibility, sorry for that. Understandable that this should happen in the absence of a clear deprecation (so far) of the unhyphenated form. Perhaps the editors concerned are using some cite-generating tool that doesn't generate the canonical form, in which case the tool should be updated. Whatever, that strengthens the case for deprecating the non-canonical forms ASAP, and for moving forward as fast as possible with the main task. It's late now, and I will be going to bed shortly. Will comment on Phil Bidger's intemperate remark below in the morning. --NSH001 (talk) 00:06, 22 February 2021 (UTC)[reply]

Naw, the editor concerned is/was just copying/pasting whatever they found in whatever articles they were looking at. (The user also hasn't learned that refs go after punctuation, or that a named ref copied from another article might not be declared on the target page, or that Wikipedia has a Preview function to allow checking for errors. Not that I'm bitter.) Agreed, without actual deprecation, nobody knows what they're supposed to use or not use, so watchlists get plagued with unnecessary repetition. And I can't point such a user to the deprecation or consensus not to use the non-hyphenated forms, because there aren't any yet. — JohnFromPinckney (talk) 00:21, 22 February 2021 (UTC)[reply]

Thanks for that. If it is the case that editors on that page are simply copy-pasting cites from the original wiki bios, then that's good news – the problem will go away if the bot is simply allowed to continue and fix the problem in the source articles. I don't think it's true that there is no consensus for this change: the desired style was settled in an RfC several years ago, and has been 90% implemented in the years since, with no substantial objection. So there is an effective consensus, and it makes no sense not to carry the task through to its logical conclusion. The objections here amount to a dislike of large numbers of bot edits appearing on watchlists, not on the actual merits of the case. --NSH001 (talk) 07:34, 22 February 2021 (UTC)[reply]

The RfC in question only had seven supports, and only concerned making sure a hyphenated version existed and was presented first in the documentation. I don't know how that could be read as effective consensus for deprecating a parameter that, prior to the bot run, seems to have been more commonly used than the hyphenated variant - and even if it could, it would be a

limited one. The objection to the bot is not just that it edits a lot, but that it "fixes" something that isn't a problem. Nikkimaria (talk) 14:01, 22 February 2021 (UTC)[reply

]

Exactly this. — JohnFromPinckney (talk) 02:05, 23 February 2021 (UTC)[reply]

Nikkimaria, the first fallacy in your argument is that you are assuming the wider community, if it participated in the original RfC, would disagree with its conclusion. That isn't (quite) the case – I'm pretty sure that, out of the various choices available for multi-word parameters (runon, camelCase, under_score, hyphenation, etc), hyphenation would still be chosen, simply because it is the easiest to read in wikitext. Possibly underscore might be better (it won't line-wrap) but most people, apart perhaps from those with a programming background, will be much more comfortable with the hyphen, so that RfC came to a very sensible conclusion. The detail of its implementation is another matter, though. The second fallacy in your argument concerns the reason for the objection to the bot. Firstly, the observable fact is that some editors are now kicking up a huge time-wasting fuss about this bot, but said nothing about the many other bot runs for all the other CS1/2 parameters doing exactly the same thing. Secondly, and I see this repeatedly in other RfCs as well, that editors tend to focus solely on a perceived short-term problem, without taking account of the bigger picture and the wider context. That context has already been well described in the introduction to this RfC, and in the links given there. It's astonishing to me that people don't see that that the whole point of this work is to make the citation templates simpler and easier to use. Part of that process is to make the parameter names more meaningful and consistent. It's ridiculous to leave the mess inherited from the early days of citation templates, with these few remaining parameters sticking out like sore thumbs. --NSH001 (talk) 08:25, 25 February 2021 (UTC)[reply]

Please explain how allowing e.g. "accessdate" is less meaningful, is less simple, is harder to use for regular editors. That is the bigger picture, the wider coontext: the benefits are actually only for a small group of people, who have every right to present their case and ask if their life can be made easier, but should not be astonished when it turns out that in some cases, their preferred solution is not supported by the larger group of people who don't do (or not as regularly) the template and bot stuff. Putting error messages on thousands of pages for things which are not an error but something which a few people decided is no longer allowed at all is losing sight of the bigger picture as well.

Fram (talk) 09:24, 25 February 2021 (UTC)[reply

]

you are assuming the wider community, if it participated in the original RfC, would disagree with its conclusion. I don't know that, and neither do you, because the wider community was never consulted. We can be reasonably sure that the discussion would have been far less one-sided, based on subsequent commentary. As to the second half of your point, as Fram notes, it's not at all clear why deprecating widely-used aliases would make the templates easier to use for the average editor. Nikkimaria (talk) 13:33, 25 February 2021 (UTC)[reply]

(replying to both). Yes, that's easy. There's a very simple rule for all the CS1/2 citation templates: all multi-word parameters use a hyphen to separate the words. That's it. Dead easy to remember. Moreover, it's already implemented for the vast majority of the parameters (only 5 are left to do). I don't buy the argument that having to type ONE extra character is an insufferable burden. The only valid objection I can see is that the bot may flood watchlists, but that is temporary until the bot run is finished. So my sympathies to those who feel irritated (I don't – I have over 6,000 pages on my watchlist, and bot spam doesn't bother me), but the irritation will be temporary. If it really does bother you, others have mentioned a way to configure your watchlist to avoid the problem. --NSH001 (talk) 07:35, 8 March 2021 (UTC)[reply]

I'm glad you personally are not irritated by this change - but others are, and others have explained why hiding the problem is not solving it. No one is proposing removing the hyphenated variation, simply supporting the (often more widely used) aliases. It's not appropriate to

justify a change on the basis of it being mostly carried out. Nikkimaria (talk) 02:49, 12 March 2021 (UTC)[reply

]

On watchlists, see #Worth noting below for a possible way of reducing the "disruption". I forgot to reply to Fram's point about the wider context: I was thinking about the overall naming convention, and why the bot makes it easier, but Trappist's latest contribution to the survey section explains also the wider consideration that we need to consider all the non-English Wikipedias that have borrowed our CS1 templates/modules. --NSH001 (talk) 10:40, 8 March 2021 (UTC)[reply]

So the solution to a bot editing against consensus is for those who object to stop watching it? You couldn't make this stuff up. Once again, this is a fucking encyclopedia, not a place for techies to dictate to editors. Find a playground elsewhere, or get a real job and you will find out that people can only get paid to write programs if thay do what their users want them to.

Phil Bridger (talk) 21:27, 21 February 2021 (UTC)[reply

]

Phil, your first sentence is false. This bot was approved by consensus, following standard procedures. Indeed its operator went to extraordinary lengths to shout from the rooftops that it is a "cosmetic" bot. I'll ignore your intemperate and baseless personal attack. Finally, on the question of bot edits and watchlists, I refer you to Bilorv's conribution at 11:34, 13 February above, which looks like a good solution (I'll just add the caveat that I haven't tested it yet). --NSH001 (talk) 08:25, 25 February 2021 (UTC)[reply]

Reading some of the above comments, I'm sorely tempted to add a new proposal here: every editor who deliberately makes a change which adds an error message to at least 10,000 pages is stripped from their template editor right. Excluding hidden cats of course, these aren't a problem; but no depreciation of any parameter justifies such mainspace disruption for readers.

Three responses: First of all, the error messages displayed by CS1 templates are displayed by consensus, not by a single editor. Second, the error messages, such as those displayed in articles within Category:CS1 errors: unsupported parameter, are shown not because a single editor changed a template, but because individual editors made errors when they used CS1 templates. Third, the objection to error messages being splashed across article space is why the bot was operating before the deprecation error messages were displayed. People hate the display of hundreds of thousands of minor error messages, so the bot was fixing the articles before the CS1 modules were changed to display deprecated-parameter errors.

Fram's feedback offer give something to think about, however; if the logical options A or B are chosen here so that the last 10% of the hyphenation of multi-word CS1 parameters can be completed, perhaps we should not display deprecated-parameter error messages (except maybe in preview mode) in articles until the vast majority of parameter name replacements are complete. – Jonesey95 (talk) 16:29, 24 February 2021 (UTC)[reply

]

Thanks, but remember

Fram (talk) 17:06, 24 February 2021 (UTC)[reply

]

Three responses: First of all, the error messages displayed by CS1 templates are displayed by consensus, not by a single editor. One thing that this discussion has made crystal-clear, I think, is that many of the "consensuses" used to make sweeping changes like this are not remotely sufficient in terms of scale per

WP:FAITACCOMPLI. --Aquillion (talk) 10:03, 4 March 2021 (UTC)[reply

]

I wish that people would stop repeating this "seven-person consensus" canard; it is a weak platform on which to base an argument. The consensus about hyphenated parameters has been in place for seven years, and has been reinforced by multiple discussions about deprecation of dozens of individual multi-word parameters during those seven years. The discussion page on which every single one of those discussions has occurred, Help talk:Citation Style 1, has 396 page watchers. – Jonesey95 (talk) 05:48, 11 March 2021 (UTC)[reply]

This page has ten times that, and I'm willing to bet there are more people opposing deprecation here than there are who supported it in the discussions you mention put together. Repeating a local consensus for less used parameters != an appropriate level of consensus to get rid of other parameters from literally millions of articles. Not to mention that the consensus that was supposedly established seven years ago wasn't even for deprecation, just prioritization. Nikkimaria (talk) 02:49, 12 March 2021 (UTC)[reply]

• I wish that people would stop repeating this "seven-person consensus" canard; it is a weak platform on which to base an argument. The consensus about hyphenated parameters has been in place for seven years To be clear, the consensus of that RFC was that non-hyphenated parameters are allowed, subject to the unambiguous restriction that no parameters can be depreciated. The RFC specifically did not favor non-hyphenated parameters over hyphenated ones, and the statement at its start specifically promised that no parameters would be depreciated as a result. There has never been any sort of consensus (not even a weak, highly-local one) to depreciate hyphenated parameters; nor, as a result, have any depreciations made on those grounds ever been valid. And it is clear from this discussion that such consensus would never have been reached if it had been sought (which it was not.) A discussion among a tiny number of people, without an RFC, which directly violated the very RFC they tried to use to justify their actions, with no further RFCs or any effort to get consensus from or even inform the wider community of what they were doing, is not a "consensus" in any way, shape, or form - longstanding consensuses get their weight from the large number of people who have seen and accepted them, and in this case the longstanding consensus is (and remains) to retain hyphenated parameters. --Aquillion (talk) 10:31, 16 March 2021 (UTC)[reply]

This is uh, not a good idea. In some cases many templates are used in error, though they previously did not detect such errors. Detecting and fixing such errors is a good thing. IAR is policy, and if someone is breaking tons of things, sure, remove their rights, but simply displaying error messages isn't a clear-cut issue.

As for non-silent parameter deprecation, yeah, replace first then deprecate. I don't think anyone disagrees with this.

talk | contribs) 20:30, 3 March 2021 (UTC)[reply

]

Well, I certainly disagree (and have, several times, on this page already). If there's consensus to do away with certain parameters, then the very first thing to do is deprecate them, that is, tell everybody not to use them anymore, next, run the bots to replace the usage, then, eventually, show the red messages and, ultimately, completely remove support for the deprecated params. Any other path is crazy. — JohnFromPinckney (talk) 03:54, 4 March 2021 (UTC)[reply]

• One of the objections I raised with Monkbot18 which is not being addressed whatsoever in the "status quo" (option A) argument is that the bot also make other changes, including the removal of whitespace and line breaks from citation templates. This is unacceptable per
  MOS:STYLERET. In my mind, the time I have spent undoing these changes to the articles I focus on has been a far bigger burden than the lack or presence of a hyphen in a parameter. If this bot is tasked with swapping parameters, it should ONLY be changing "accessdate" to "access-date" (and vis a vis similar parameter hyphenations), and absolutely nothing outside of that. Option A should not be construed as approval of the bot code as currently written, but the task for which it was meant to be accomplishing. - Floydian ^τ _¢ 18:33, 7 March 2021 (UTC)[reply
  ]

  Well, that's very odd, because Monkbot18 is very careful to preserve whitespace (with one very reasonable exception), so I don't see how it's going against STYLERET. It's true that it does remove empty/blank parameters, but that's a good thing, as it reduces clutter in the wikitext. It does a few other good things as well, see User:Monkbot/task_18: cosmetic cs1 template cleanup. Can you give specific examples of edits you think it's doing wrong, please? --NSH001 (talk) 19:13, 7 March 2021 (UTC)[reply]

  There's a nice list of reversions if you look at my edits for January 17, but here is an example. - Floydian ^τ _¢ 06:27, 8 March 2021 (UTC)[reply]

  Ah, that's the "one very reasonable exception" that I referred to. All the other changes that you felt you needed to make in your "cleanup and standardise first 150 refs. Revert stupid friggen MonkBot making stylistic changes to thousands of articles on a "consensus" of like 3 people, and block it from making further edits" edit are down to other editors/bots, not Monkbot18. I think a blank line within a cite template is always unnecessary, and uses up valuable space within the edit window, but if it bothers you that much, you can always ask Trappist nicely, and I'm sure he'll remove that tweak for you. --NSH001 (talk) 07:15, 8 March 2021 (UTC)[reply]

  In that case, Your Mileage May Vary, and STYLERET applies. The blank lines make it easier to pick out citations from text, and scroll bars overcome the "valuable space" argument. I tried to ask that behaviour to be removed. I was met with (*paraphrased) "bot code was approved, proceeding". So I just blocked the damn bot from the 300 articles I work on. Now if someone is that concerned about a hyphen in a parameter, they can go manually change it. Simple as that.

  Or, you know, the bot could stick to making the changes to parameters that it is supposed to. I shouldn't need to ask, it's not part of the bot's mandate, remove it. - Floydian ^τ _¢ 15:43, 8 March 2021 (UTC)[reply]

  First, you say I tried to ask that behaviour to be removed. I was met with (*paraphrased) "bot code was approved, proceeding". That is surprising, as I would expect Trappist to be amenable to such a request (it's important to get this job done, so a small change like this one to avoid pushback is worth making). Can you point me to the conversation where you made this request, and received that reply, please?

  (later) - ah, don't bother – found it myself User talk:Trappist the monk/Archive 17#Task 18 taking out reference spacing. Looks like you didn't explain yourself very well to Trappist. FWIW, I can see the rationale for the blank line for in-line cites within the article body, if it stops people from turning it into the (horrible) horizontal format. I touch on this again below, but I agree with you that Monkbot18 shouldn't be removing the blank line. (I hope Trappist is reading this). But the rest of the changes it's making are good, and should stay.(Actually, to be strictly correct, Monkbot should remove it within LDR or biblio listings, but not within the main body. For simplicity, I'd say simply don't bother trying to remove it.)

  Secondly, Or, you know, the bot could stick to making the changes to parameters that it is supposed to. That statement is false. The bot was approved to carry out the tasks listed at the link I've already given: User:Monkbot/task_18: cosmetic cs1 template cleanup, and that's exactly what the bot has been doing. Apart from the blank line issue, the other changes are good, and valuable, and will reduce the need for more bot runs in the future. One of the reasons I like this bot so much.

  The next bit is very interesting (to me) but is wandering mostly off-topic, so I'm putting it in small text. If you'd like to take it further, feel free to discuss it on my talk page. The problem of citation clutter in the main body of Wikipedia articles has been annoying me for years, and especially the huge problems caused by long, horizontally-formatted citation templates, which in my opinion make wikitext difficult or impossible to read and to edit. This is all set out on a very long "thread" (it isn't really a thread any longer): on my talk page. It is talking about a way of setting out citation templates that I call "ETVP" for "easy to visually parse", which is similar in many ways to the citations in your example, but also differs in some respects. The interesting point to mention here is that I had a difficult and bruising experience trying to introduce ETVP-formatted citations into article bodies. The excuse offered for reverting me was mostly "it takes up too many lines" in the edit box (hence "the one very reasoable exception" above), so in the end I gave up on that, and focused mainly on other solutions (ETVP within WP:LDR or ETVP within biblio listings using short-form referencing) which in most cases are actually a better solution anyway. It's a fascinating paradox that you seem to have gotten away with it by using more lines, not fewer, combined with a lavish use of white space – the exact opposite of what one would expect. So I am now thinking of adding a similar option to my ETVP script – and thank you for prompting that thought. Will need some thought, though.

  --NSH001 (talk) 10:02, 9 March 2021 (UTC)[reply]

  The blank lines is the only issue I'm having / pointing out here. I have no problem with updating deprecated parameters, I have no problem with my watchlist having a litany of bot edits. I do have a problem with going through 250 articles that have an average of 50 citations, to reinsert a blank line in each. This is not one of the 6 tasks listed at User:Monkbot/task_18: cosmetic cs1 template cleanup.
  There are also some automated tools that do similar nonsense that I revert on sight (e.g. Regex Citation Formatter). - Floydian ^τ _¢ 15:49, 10 March 2021 (UTC)[reply]

  We appear to be in agreement here, specifically that you support option A, but only if Monkbot 18 drops its removal of entirely blank lines within citation templates. If you could confirm that my understanding is correct, that would be very helpful. Thanks. --NSH001 (talk) 10:00, 11 March 2021 (UTC)[reply]

  You would be correct good sir! - Floydian ^τ _¢ 15:36, 11 March 2021 (UTC)[reply]

Worth noting

Trappist has kindly set out, very clearly, a suggestion on how to configure your watchlist to avoid the problems of large numbers of bot edits in watchlists. I copy it here, in case it is helpful to anybody:

• at Special:Preferences#mw-prefsection-watchlist:
  • check Advanced options → Expand watchlist to show all changes, not just the most recent
  • check Changes shown → Hide bot edits from the watchlist
• at Special:Preferences#mw-prefsection-rc:
  • check Advanced options → Group changes by page in recent changes and watchlist

Note that I haven't (yet) tried this myself, since I'm already satisfied with the way my watchlist is set up. --NSH001 (talk) 09:09, 8 March 2021 (UTC)[reply]

Note to RFC closer: the above instructions are a remedy for all of the people supporting Option C because of "watchlist clogging" and the alleged problems that the bot's edits cause for editors who watch for vandalism. As far as I can tell, no editor using the above settings is objecting to the bot's changes based on watchlist issues. Unless a valid objection is raised, I propose that all Option C support citing watchlist problems be discounted and guided to the above recommendation. – Jonesey95 (talk) 17:30, 8 March 2021 (UTC)[reply]

This is a workaround that (a) should not be necessary, (b) doesn't work for everybody, e.g. those who want to see bot edits (I do for example) and (c) doesn't fix the problem only a symptom. It is additionally highly inappropriate to suggest that large numbers of editor's valid and rationally expressed preferences are discounted. Thryduulf (talk) 17:41, 8 March 2021 (UTC)[reply]

I've been trialing this workaround for the last day or so. Back in December I cleared out my watchlist and took a two and half month wikibreak while the bot ran. I've just come back to Wiki and discovered that the bot has been stopped, but thought I'd try Trappist's suggestion. So far so good, it's a bit different but at least it makes the watchlist saner than during the bot attack! Martin of Sheffield (talk) 20:23, 8 March 2021 (UTC)[reply]

I've always been puzzled that (some) editors get so triggered by large numbers of bot edits. I have more than 6,000 pages on my watchlist (which I have set to show bot edits), and bot spam has never bothered me. I even welcome it, as they sometimes remind me of articles I did a lot of work on perhaps 7 or 10 years ago, and which I really ought to look at again. That said, I do understand the problems of editors who want to deal with vandalism, so this new setting looks to be very valuable, and should indeed remove many (but probably not all) of the "watchlist spam" objections. --NSH001 (talk) 10:40, 9 March 2021 (UTC)[reply]

And on the subject of editors like me who aren't bothered by bot spam, has anyone considered the silent majority who either aren't bothered by the "spam" or who, if they are, aren't concerned enough to come to this RfC to complain about it? Perhaps they ought to be weighed somehow in the balance when considering "consensus"? --NSH001 (talk) 10:51, 9 March 2021 (UTC)[reply]

Well given that most of them will not know that this discussion exists and will just be getting on with adding the parameters, with and without hyphens, as they always have done I don't think was can say one way or the other what their opinion is - especially given that the bot has not been spamming its unnecessary changes for quite a while now (the discussion has been open a month tomorrow, and I think it was stopped a day or so before then, and more than a few editors are of the opinion that arguing against bots/bot operators is pointless). Instead of grasping at straws to discredit or dismiss the opinions you disagree with, perhaps you could instead try listening to why they disagree with the changes, not just the manner of the changes. I also note that you have completely ignored my explanation of why this will not actually solve the problem for everybody and ignored that there is no reason why the problem should need to be solved in the first place. Thryduulf (talk) 11:49, 9 March 2021 (UTC)[reply]

None of this changes the fact that there is no consensus to depreciate non-hyphenated parameters, nor has any such consensus ever existed. Rather than trying to convince the RFC closer, you should be planning how you're going to get that consensus, if you intend to keep doubling down on that. --Aquillion (talk) 10:35, 16 March 2021 (UTC)[reply]

• Comment It's easier for authors if every possible format is accommodated, and recognized as a synonym. There's no reason to delete any unless they;re actually confusing. DGG ( talk ) 05:59, 21 March 2021 (UTC)[reply]