This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

The fine folks over at Talk:Donald Trump currently have a "Current consensus" item on their talk page that disallows including archive URLs for citations that aren't dead (25. Do not add web archives to cited sources which are not dead. (Dec 2017, March 2018)). This runs counter to this guideline, specifically WP:DEADREF, which seems to suggest that it's better to preventatively archive pages than to wait for them to be dead and hope that an archived copy is available (this guideline also notes that even if a link doesn't necessarily die, the content of the link can change and make the source unsuitable for statements it is used to support). My gut says to simply strike that item as a clear WP:LOCALCON and direct those editors here to make their case for an exception, but I wanted to see what the feeling was here before proceeding. Also relevant is this closed discussion: Special:Permalink/1197984238#Reversion_of_archives. —Locke Cole • t • c 07:07, 23 January 2024 (UTC)

You know that it is entirely possible to "preventatively archive pages" without pushing the archive link into Wikipedia, right? Just tell archive.org to archive the page. Then, if you ever need it, there it is on archive.org waiting for you. If you don't yet need it, what is the point of keeping a prematurely frozen archive link here, when archive.org will keep track of all the archived versions that it has and let you choose which one you want when you want it?

I would suggest that, to the extent that WP:DEADREF suggests copying the archive link here rather than merely making an archived copy, that language should be changed. But I note that the actual language of DEADREF is merely to consider making an archived copy; the actual language suggesting copying it here is in WP:ARCHIVEEARLY which does not even have the status of a Wikipedia guideline. Therefore, there is nothing for LOCALCON to be violating.

As for why it can be a bad idea to copy the links here: because sources may still be in flux and the editors may prefer readers to see the current version than an old frozen version. This may be especially true for topics in current politics. —David Eppstein (talk) 07:52, 23 January 2024 (UTC)

(edit conflict × 2) I was about to suggest something similar, i.e. making sure archives exist without actually adding them (if that's possible). Primefac (talk) 07:57, 23 January 2024 (UTC)

WP:DEADREF links to a section titled Preventing and repairing dead links, which is kind of where I got the impression it was more than simply a suggestion (and as to WP:ARCHIVEEARLY, it is literally tagged as a how-to guide). I agree it's possible to create an archive and not link it, but this still places the burden on future editors/readers to find a revision of the page that supports the statement being cited which can be problematic if a source changes (as you note for political content, this can happen frequently). I've also always viewed citations as a point-in-time thing when it comes to people/events, so the idea that an archive link might point to an "old" version is a feature, not a bug. The reasons given at Talk:Donald Trump all seemed to revolve around bloating of the page size which seems like a technical concern that shouldn't be getting used as a means to stifle page development. —Locke Cole • t • c 08:08, 23 January 2024 (UTC)

So your position is that this guideline forces editors to use frozen versions of sources rather than allowing sources to be dynamic? Instead, that seems to me to be the kind of content-based editorial decision that a local consensus is entirely appropriate for. —David Eppstein (talk) 17:30, 23 January 2024 (UTC)

*sigh* If the live source changes after a statement is written, the frozen archive can be used to verify the source as it was originally seen... Nothing is being "forced", I'm just stating plainly that it's better behavior for editors to preserve their sources as they write rather than have to go through archives for potentially years to find the source that originally said something if the source ended up being dynamic/changing. Regardless of that, I'm concerned that we're recommending preventing dead links here in this guideline and a page has taken it upon itself to wholly disallow this good and desirable behavior. I'll again point to Special:Permalink/1197984238#Reversion_of_archives, where an editor was basically hit with a hammer over this and their response was about as good as you'd expect (I am never touching this article again). Do we really want individual pages to unilaterally decide these guidelines are irrelevant and drive off productive editors doing what we're suggesting? —Locke Cole • t • c 18:44, 23 January 2024 (UTC)

"disallows including archive URLs for citations that aren't dead" – Good. It's not "good and desirable behavior". It's code bloat that we don't need, and additional cite-by-cite verbiage and link confusion that the reader doesn't need. Removing that cruft does nothing whatsoever to "stifle page development". It's entirely sufficient to have IA archive something while you cite it, and just not add to Wikipedia the archive-url that we do not presently need. If linkrot happens for a particular citation, the add it.  — SMcCandlish ☏ ¢ 😼  22:01, 2 February 2024 (UTC)

I think the theory is that if the archive links are added now, then they will less likely to be archive links to 404 pages (thus requiring manual intervention to find the correct one, rather than just using the most recent). WhatamIdoing (talk) 04:33, 3 February 2024 (UTC)

I don't understand how you think that changing Wikipedia to point to an archive link now, rather than merely telling the archive to make a copy but then only using that copy later when it is needed, would have any effect on what one finds at the archive link. If the archive link works, it works, and linking to it will not change that. If the archive link 404s, it 404s, and linking to it will not change that either. —David Eppstein (talk) 06:13, 3 February 2024 (UTC)

• If you make the archive link today, and you record the archive link today, then you know the content is good, and you know which archive link you need to use.
• If you make the archive link today, and sometime during the next several years, the page becomes a 404, then at some future, post-breakage date, you will have to go through multiple archived links, some of which have the desired content and some of which don't, to figure out which one actually verifies the contents (see "requiring manual intervention" in my comment above).

This is due to the structure and goals of the Internet Archive. They don't archive a URL just once. They make multiple copies at different points in time. WhatamIdoing (talk) 06:54, 15 February 2024 (UTC)

Does WP:DEADREF not reflect the current consensus here? I honestly don't care if people here want to shoot themselves in the foot anymore, so if the thought process from @David Eppstein and @SMcCandlish is that early archiving is code bloat that we don't need or it is entirely possible to "preventatively archive pages" without pushing the archive link into Wikipedia (sic, emphasis added) then perhaps it's time to strike DEADREF or shuffle it off to a different (non-guideline) page. Sources, especially online sources, can be brittle and subject to the whims of website designers and complete site overhauls where old links die completely (and current "archives" are just "not found" pages). I don't think "code bloat" should be a concern used to undermine preventative measures to preserve sources/citations. —Locke Cole • t • c 04:45, 3 February 2024 (UTC)

DEADREF is not broken in any way, and is quite clear: When permanent links [DOIs, etc.] aren't available, consider making an archived copy of the cited document when writing the article; on-demand web archiving services such as the Wayback Machine (https://web.archive.org/save) or archive.today (https://archive.today) are fairly easy to use (see pre-emptive archiving). That does not say "and put the archived copy into the article before it is actually needed". All of the other material in that section, as in every single word of it, is about repairing citations with dead links.

What is broken is WP:ARCHIVEEARLY (which is part of a supplementary how-to essay, not a guideline), which someone added as their opinion and which clearly does not represent an actual consensus. It says To ensure link accessibility and stability, please consider pre-emptively adding an archive URL from an archive source such as the Internet Archive or WebCite. This practice is actually and clearly disputed, and that material should be changed, unless/until there is a firm consensus that not only is it good advice but that we actually need it despite WP:CREEP. It should instead re-state in a how-to manner what is said about this at DEADREF: create the archive on-demand today, but do not put it into the article if it is not already needed.  — SMcCandlish ☏ ¢ 😼  05:07, 3 February 2024 (UTC)

create the archive on-demand today, but do not put it into the article if it is not already needed ^{[citation needed]} —Locke Cole • t • c 17:44, 3 February 2024 (UTC)

I really wish the consensus at Donald Trump were exported site-wide. I had a discussion about a month ago on the same topic at Talk:Augustus. Basically, people are still wasting their time WP:MEATBOT-ing and the results of it are extremely disruptive to editors seeking to actually improve articles rather than "maintaining" them. Ifly6 (talk) 03:42, 18 March 2024 (UTC)

I'm struggling to see how having an archive of a citation used in our article is somehow a negative thing. I still haven't come across a convincing reason other than WP:IJUSTDONTLIKEIT. Which.. cool. I like an encyclopedia I can verify the information it contains through it sources, today and in the future. It kind of stuns me that anyone can defend not having archive links ready that capture sources in the state they were when they were used for a statement. —Locke Cole • t • c 04:41, 18 March 2024 (UTC)

Can you explain what these drive-by archivers are doing that isn't already done automatically? Ifly6 (talk) 05:12, 18 March 2024 (UTC)

automatically I'm assuming you mean bot that finds dead references and attempt to produce an archive after the link has died? That's easy, see WP:DEADREF, but basically it's better to create an archive before a page goes missing (or changes substantially) than to wait until the worst has happened. If an archiving system like archive.org hasn't produced a backup, then there's no getting that source back (because it's already gone). —Locke Cole • t • c 05:45, 18 March 2024 (UTC)

Bots create the archives automatically too... doing so around 24 hours after the site is added. And if you use |access-date= the bot will also choose the version closest to or before the access date if the link 404s. What is being done that isn't just drive-by archivers duplicating bot work? Ifly6 (talk) 09:45, 18 March 2024 (UTC)

What bot is doing this? —Locke Cole • t • c 15:10, 18 March 2024 (UTC)

It's all documented at Wikipedia:Link rot#Automatic archiving. There is a bot called No more 404 that archives added links. There is a bot, WP:IABOT, which monitors whether those links become dead and inserts |archive-url= when that occurs. Ifly6 (talk) 16:52, 18 March 2024 (UTC)

So... not a BOT in the WP:BOT sense but an opaque, off-wiki process that has no way of being verified? I'm still not entirely sure why people are so aggressively against pre-emptive archiving. Do you want your work to be unverifiable if a link goes stale, dead or changes? —Locke Cole • t • c 19:52, 18 March 2024 (UTC)

If the information being cited at the link source changes, then our articles need to reflect that change. Linking to “archived” (ie out of date) version of the source isn’t what we want. Indeed, an out of date source may be considered “no longer reliable.” Blueboar (talk) 20:04, 18 March 2024 (UTC)

That's wonderful. It sounds like something that should be addressed on an article talk page when a changed link occurs. It sounds secondary to wanting to preserve our sources so they can be verified even if they change or disappear. —Locke Cole • t • c 20:56, 18 March 2024 (UTC)

That is the opposite of how I see it. We cite a source to verify content in an article. If the information on a website changes, then it may no longer support that content. It is then necessary to either change what the article says to match the source, or find a new source to support what the article says. We need to be able to verify that the website in its previous state did indeed support the content in the article. If the original content of that website is no longer valid, then it doesn't matter whether the website is unchanged, has been updated, or is dead. We then need to assess available reliable sources to determine what the article should say. If we know that a website is likely to be updated, we should be citing an archived version of the website that supports the content of the article, rather than linking to something that is likely to stop supporting the contents. I think that in the overwhelming majority of cases, any changes to a website are likely to reduce its usefulness as a source for the contents of the article. Donald Albury 21:27, 18 March 2024 (UTC)

Exactly. Suppose we write that 25 people were killed in a deadly accident, based on a source that originally reports “25 people were killed”… ok, our content is verifiable. HOWEVER, let’s say that subsequently that source amends its reporting to say “25 people were seriously injured, and 3 died”… now our content is outdated, and is no-longer verified by the source. We need to update our content. If we prematurely archive the source, we might never catch that the source corrected its information and no longer supports the “25 dead” number. Blueboar (talk) 22:06, 18 March 2024 (UTC)

There are more kinds of articles than "current events"-type articles, you understand that right? There are other reasons to have archives prepared in advance as well, not least of which is being able to confirm if a statement was ever true (for behavioral issues where an editor makes a statement, provides a source, then claims it "changed"). —Locke Cole • t • c 03:21, 19 March 2024 (UTC)

Blueboar, in that unusual circumstance, both the article content and the archived link need to be updated.

The far more common circumstance is: the article gets cited, the bot adds an archive link, the original site (or at least that article) dies, and we can still see what the original article said when it existed.

On a side note, I wonder if people are really understanding each other. We're talking about the difference between these two versions:

• Regina Milanov. "Istorija ribarskog gazdinstva Ečka". Retrieved 2018-07-31.
• Regina Milanov. "Istorija ribarskog gazdinstva Ečka". Archived from the original on 2019-10-04. Retrieved 2018-07-31.

If you've got the first, and the website dies (this particular website now throws a HTTP 403 error), then you can't tell whether the website used to say something relevant without someone digging through the Internet Archive to see whether they happened to archive that page before it died. WhatamIdoing (talk) 05:35, 26 March 2024 (UTC)

That article will have a thousand citations by the election. Will 1000 extra parameters and links slow the page loading? Rjjⁱⁱⁱ (talk) 05:18, 18 March 2024 (UTC)

Will 1000 extra parameters and links slow the page loading? Even if it does, it shouldn't be the basis for how we edit the project. See WP:AUM for a time when page loading was used as an excuse to try and prevent editors from creating a better encyclopedia. It's on the devs to look at things that are causing site problems and address them using technical means. —Locke Cole • t • c 05:48, 18 March 2024 (UTC)

I can't see anything in policy that dictates the point either way, so editors seem to be allowed to make article by article decisions on the matter. Personally I would be pro-inclusion for the reason outlined by WhatamIdoing above, but I don't see anything that says it must be done one way or the other. -- LCU ActivelyDisinterested «@» °∆t° 10:12, 18 March 2024 (UTC)

Perhaps I am missing something, but I don’t really understand why anyone would want to archive a citation “preemptively”. Could someone who supports doing so enlighten me? Blueboar (talk) 16:51, 18 March 2024 (UTC)

Is there something at WP:DEADREF and WP:ARCHIVEEARLY you don't understand? —Locke Cole • t • c 19:45, 18 March 2024 (UTC)

Yes… I understand using archives for dead links… but when we expect a webpage to change its content (say because it is out-of-date or incorrect) why would we want to cite an archived version? I would think we would want to cite the most up-to date version (and if necessary change OUR article content to match the up-to-date, corrected website). Blueboar (talk) 23:01, 18 March 2024 (UTC)

There's one case I know of where it's worth doing: Galactic Central hosts bibliographic details such as this which are autogenerated from a database that is updated once a quarter. When the quarterly update happens, all the URLs change, so if you were citing that page to show that Keith Laumer's The Planet Wreckers appeared in the February 1967 issue of Worlds of Tomorrow, the page will no longer contain that information. When I cite this website I usually preemptively archive it so that I don't have to go hunting for the right archive page a year later. Mike Christie (talk - contribs - library) 23:11, 18 March 2024 (UTC)

Blueboar, I see on a very regular basis articles with dead references that were never archived. Would it be nice if those references had been archived shortly after they were originally added to the article? Yes. Do I think it must occur? Not really. So I guess I'm not really in either of the camps discussing the issue in the main thread, but I guess from a maintenance standpoint I am pro-archive. Primefac (talk) 16:49, 19 March 2024 (UTC)

That is a good argument for triggering off-site archival of pages that you use as references. It is not an argument for using the archived copy to replace the source on Wikipedia. —David Eppstein (talk) 17:51, 19 March 2024 (UTC)

Who's talking about "replacing"? The idea, I gather, is to have both live URL and archive URL listed before it's too late. Which sounds reasonable enough. Frankly, I can understand if people say "I'm too lazy for that", but I don't have the slightest idea why anyone would want to prevent others from doing it. Gawaon (talk) 18:04, 19 March 2024 (UTC)

@David Eppstein, @Gawaon Yeah, is this the reason for the attitude thus far? At no point has anyone suggested we replace functional URLs with their archive's. This is why ((cite web)) has both a |url= and |archive-url=, and only when |url-status=dead (or if |url-status= is not set) does the |archive-url= get used if it is present. If |url= is still live, simply using |url-status=live will keep |archive-url= from being shown. This is all explained in the docs at ((cite web)). The only argument at Talk:Donald Trump appears to center around "bloat" of the page, which again, is not a reason to avoid good maintenance of one of our more popular articles. —Locke Cole • t • c 22:32, 19 March 2024 (UTC)

• Ok… I can see that there might be situations where prermptively archiving a source is helpful (thanks)… I hope people can understand why I had concerns. Perhaps we need to work up more guidance on when to do so and (perhaps more importantly) when not to do so. Blueboar (talk) 18:34, 19 March 2024 (UTC)

  To be a little more precise, it's not pre-emptive archiving that is questioned; after all, you can't make a copy in the archive once the site/page is already gone. Archiving must be done in advance, or it can't be done at all. The complaint is that the Wikipedia article is storing a link to the archive copy. I believe there are two complaints about this:

  • Depending on the parameters chosen, the existence of the archive might be shown in the (visible) references list. It won't be linked as the regular/main link, but readers will see that it exists, and some editors think that's ugly.
  • Even when it's not visible to readers, the extra URL is visible to editors in the wikitext, and some editors think that this "unnecessary" (so far) information is very inconvenient for them to work around.

  WhatamIdoing (talk) 05:45, 26 March 2024 (UTC)

  Those seem very weak reasons. Gawaon (talk) 07:24, 26 March 2024 (UTC)

  I'm sure that all reasons held by any given individual seem strong to that particular individual. WhatamIdoing (talk) 17:35, 26 March 2024 (UTC)

  • Beyond the fact that you have to archive preemptively, as described, another element is that you don't necessarily know a page will get properly archived before its content completely changes to something like a redirect (as is the case with many long-running sites, at some point they change their structure and a bunch of old content essentially gets black-holed, and it's not going to be very visible to editors who added that link originally that the content is effectively gone; at least with the archive we're giving readers a fair shot of finding it without having to check for archives themselves.) I appreciate the people who hate the density of the wiki text, but there are ways around that (putting refs in the ref section at the end, for instance, rather than inline) and WP:V is a much more important principle than "it looks nicer to me without the extra text in references". Der Wohltemperierte Fuchs ^talk 17:39, 26 March 2024 (UTC)

• The wording is strange and I think it should be improved before we discuss this proposal much longer. WP:CONLEVEL, as I understand it, already says that a local consensus, say on a talk page, cannot override guidelines such as this one. The problem with the mentioned section, however, seems to be that it mentions "making an archived copy of the cited document" but doesn't say anything about adding a link to the reference using the |archive-url= mechanism or so. Surely it was the intent that one should do that too – after all, what would be the point of an archived copy if nobody knows where to find it? So I think a simpler fix, and more in line with the usual wording of guidelines, would be to add something like "and adding it to the relevant reference, for example using an |archive-url= parameter" after "consider making an archived copy of the cited document when writing the article". Once that's there, CONLEVEL should handle the rest and no new sentence is needed. Gawaon (talk) 19:02, 27 March 2024 (UTC)

  I would support this and agree it's a cleaner approach. I would not object to closing this early and proposing your change instead. =) —Locke Cole • t • c 01:55, 28 March 2024 (UTC)

  You could just try adding it to the page first and see if anyone reverts it. Maybe it'll be fine even without requiring further discussions? Gawaon (talk) 07:13, 28 March 2024 (UTC)

  Done, with minor addition. —Locke Cole • t • c 15:45, 28 March 2024 (UTC)

  Well, it was challenged, so I guess further discussion will be needed. Gawaon (talk) 13:19, 29 March 2024 (UTC)

  @Nikkimaria: You don't get to revert "per talk" and not actually comment on the talk page. Why are you reverting this? —Locke Cole • t • c 16:24, 29 March 2024 (UTC)

  Multiple editors raised concerns with this practice above. While I agree that your proposed addition was poorly phrased, I don't agree with the advice to just implement your new change given the previous discussion. Nikkimaria (talk) 16:28, 29 March 2024 (UTC)

  Multiple editors were confused about what was being discussed apparently, believing working (non-dead) URLs would be replaced with links to Archive.org/WaybackMachine. Since that correction there has not seemingly been any push back on it. —Locke Cole • t • c 16:30, 29 March 2024 (UTC)

  I don't see multiple editors conflating those situations, but do see specific objections to adding, eg the concerns about code bloat and additional parameters. I see that you disagree with those objections, but that doesn't mean they don't exist. Nikkimaria (talk) 16:40, 29 March 2024 (UTC)

  Do you have an actual objection to this or are you simply objecting because of your interpretation of the conversation above? —Locke Cole • t • c 19:21, 29 March 2024 (UTC)

  I share the concerns expressed in the conversation above. Nikkimaria (talk) 19:24, 29 March 2024 (UTC)

  Which one? —Locke Cole • t • c 04:48, 30 March 2024 (UTC)

  That it is appropriate to simply ensure that an archive exists if it is needed, that pre-emptive additions result in a lot of bloat, that this is a matter for local editorial consensus, and that a dead link is indicative of the need for editorial reconsideration rather than simple technical substitution. Plus, peripherally, the meatbotting is annoying. Nikkimaria (talk) 05:04, 30 March 2024 (UTC)

  it is appropriate to simply ensure that an archive exists if it is needed This does nothing for future editors or readers in determining the validity of a statement if an existing source becomes dead (either temporarily or permanently). pre-emptive additions result in a lot of bloat This is irrelevant, and can be mitigated by placing citations at the end of the article and referencing them earlier (see Ridgeline High School (Washington) for an example). that this is a matter for local editorial consensus Our articles being verifiable with citations that are able to withstand sources changing or disappearing is not something up for local editorial consensus. a dead link is indicative of the need for editorial reconsideration rather than simple technical substitution And an archive can aid in finding a live version if it is a copy of a widely published paper, but without a copy of the original source to refer to for quotations, such a search becomes more problematic depending on the citation/source used and the statement needing to be cited.

  meatbotting is annoying ??? I assume you mean people doing the good work of providing archives for references or are you engaging in personal attacks?

  Is there anything else? —Locke Cole • t • c 17:03, 30 March 2024 (UTC)

  There are benefits to creating archives pre-emptively, which is why no one AFAIK is proposing banning the practice. But there are also drawbacks, and requiring rather than simply permitting adding archivelinks pre-emptively requires engaging with those drawbacks in good faith. Instead dismissing concerns as "irrelevant" and "personal attacks" (??) weakens rather than strengthens your case. Nikkimaria (talk) 17:51, 30 March 2024 (UTC)

  It hasn't been suggested that they be required, only that people doing the work not have their efforts reverted or omitted simply because of a WP:LOCALCON. If an article starts without archives, and six months, a year, or years later, someone adds them, it should not be permissible for that to be a point of editorial discussion on just that page. It's disruptive to the maintenance of a project like Wikipedia. —Locke Cole • t • c 17:58, 30 March 2024 (UTC)

  If you are arguing that editors cannot choose to exclude something, you are arguing that that something is required. Nikkimaria (talk) 18:01, 30 March 2024 (UTC)

  The two statements are not mutually exclusive. So, no, not required. Required would mean they needed to be added with any new or changed citation. Nobody here has proposed that. Simply acknowledging that a local consensus cannot unilaterally remove archives (or in this case, ban them outright) does not mean editors must add them. Only that they cannot remove them once added without a good reason (wrong archive, source changed, etc). —Locke Cole • t • c 18:28, 30 March 2024 (UTC)

  The people who formed a local consensus to exclude archivelinks felt they had good reason to do so. If you somehow manage to ban local consensuses on this issue without making archivelinks required, they can just revert citing those reasons instead. How is that an improvement on the current situation? Nikkimaria (talk) 19:03, 30 March 2024 (UTC)

  The people who formed a local consensus to exclude archivelinks felt they had good reason to do so. Well, you can certainly go see the good reason for yourself. I hesitate to call the reasons provided good however:

  • They appear to not understand how archive links work (some proponents of omitting them appear to believe simply having an archive at all is good, even though it may mean wading through hundreds of revisions to find the correct one (as discussed above))
  • They appear to have concern over "bloat" of the page wikitext (which can be resolved completely by formatting it to have references all at the bottom, this also makes reference/citation maintenance easier)
  • Belief that it will somehow make the page larger for downloading (hint, MediaWiki (the software that runs Wikipedia) uses gzip to compress page results on modern browsers, so while the HTML response for Donald Trump is 1.79MB, the compressed size is 317KB; Archive.org/WaybackMachine links are typically an archive.org link, followed by web and a ISO formatted date/time, then the original URI, so the major part, the original URL of the source, will be duplicated, which should compress very well; example: https://www.newsweek.com/fact-check-us-rounding-turn-covid-trump-claims-1542145 vs https://web.archive.org/web/20230510211847/https://www.newsweek.com/fact-check-us-rounding-turn-covid-trump-claims-1542145)

  The consensus version of Wikipedia:Citing sources encourages pre-emptively archiving. Per WP:CONLEVELS, Consensus among a limited group of editors, at one place and time, cannot override community consensus on a wider scale. I'm very sorry that the editors at Talk:Donald Trump don't understand why we pre-emptively archive, but editors here have already made it a site-wide recommendation. The edit we were making was attempting to clarify this already existing consensus. —Locke Cole • t • c 04:20, 31 March 2024 (UTC)

• The consensus version asks editors to consider creating an archived copy, nothing more. I'm very sorry that you'd like it to be much more than it is, but it's not, plus even the version I reverted wouldn't be enough to achieve what you seem to be desperate for. And given the above, I don't think further discussion is likely to shift the needle. Nikkimaria (talk) 04:54, 31 March 2024 (UTC)

  I mean if you want to be obstinate, I can't stop you. Go with God. But if you think there's anything less than a consensus that pre-emptively archiving is a desired practice sitewide, then you're going to be sorely disappointed. Good day. —Locke Cole • t • c 05:15, 31 March 2024 (UTC)

  I also find it troubling that, even faced with the "reasons" for omitting archive links being refuted, you simply ignore that entire part of my response to focus on my reading of the text as it stands. It's almost like you don't really care if there's a reason for excluding archive links, you WP:JUSTDONTLIKEIT. —Locke Cole • t • c 05:35, 31 March 2024 (UTC)

  They appear to have concern over "bloat" of the page wikitext (which can be resolved completely by formatting it to have references all at the bottom, this also makes reference/citation maintenance easier) Re that last, a subject dear to my heart. See the case study at Killing of Michael Brown. Yes, it eliminated a lot of clutter in the prose, which is why I did it. I soon discovered its downsides, including the fact that nobody else wanted to change the way they had always done citations. I ended up spending tons of time converting their work to conform. Every day I would have a number of new citations to convert. Editors could see me doing that, and still they didn't help out. No thanks. Have a look at the wikitext to see how well the convention held up after I left (spoiler: it didn't). And there are other more obscure downsides that I could get into but won't.

  I'm very sorry that the editors at Talk:Donald Trump don't understand why we pre-emptively archive - Do you mean "pre-emptively add the archive parameters"? Those aren't the same thing. Let me reassure you that the article's editors do understand the link rot issue. You'd be surprised how much we understand. ―Mandruss ☎ 10:44, 31 March 2024 (UTC)

  I was discussing it in the sense of how Wikipedia:Citing sources talks about it. They are the same thing, on this page, because that's literally what the page is recommending be done to avoid linkrot. —Locke Cole • t • c 17:21, 1 April 2024 (UTC)

  Quite remarkable that this discussion has been open for over two months and the subject article was not notified until now.[1] Heaven forbid one should facilitate opposing opinions and additional insights.

  It's quite simple. Donald Trump has historically had problems with exceeding the Post-expand include size limit, effectively breaking the article. We tried various solutions over the years, some of which helped for some period of time until the article grew more. Ultimately the article's editors decided to omit the archive parameters for sources that are not dead, and the problem has not recurred since then. The PEIS limit is 2,097,152 bytes and the most recent attempt to add the archive parameters (to 585 citations) increased the article's PEIS to 1,980,594 (94.4% of the limit). In other words, just a little more article growth, which is likely to happen soon given the current election situation, would break the article again. Things like this are precisely why guidelines are only guidelines.

  Persuade the powers-that-be to increase the PEIS limit substantially (if they have the power to do that), and I'm sure we would be happy to revisit our consensus. ―Mandruss ☎ 06:01, 31 March 2024 (UTC)

  Tracked in Phabricator
  Task T189108
  

  I looked over the "consensus" reasons provided and PEIS was never mentioned. Looking at that further, I think it's an issue, but not as significant as it's being made out to be by you. In a revision where archives were added (Special:Permalink/1186826799) the PEIS was 1980826/2097152 bytes (94.45%). Once it was reverted (very next revision) ((Special:Permalink/1186827656), it was 1869188/2097152 bytes (89.13%). As PEIS is a technical restriction, one which hasn't changed in decades, the correct solution is to get that setting changed, not to make our articles worse. —Locke Cole • t • c 17:26, 1 April 2024 (UTC)

  I looked over the "consensus" reasons provided and PEIS was never mentioned. Exactly what I previously said below. the correct solution is to get that setting changed, not to make our articles worse. Self-quote from above: Persuade the powers-that-be to increase the PEIS limit substantially (if they have the power to do that), and I'm sure we would be happy to revisit our consensus. Key word "revisit": upping the PEIS limit wouldn't guarantee a consensus change, since editors may still prioritize other factors over the avoidance of limited link rot. You and others are welcome to participate in a revisitation of the consensus at the article's talk page ("the article's editors" don't "own" the article by any means), but the issue is still subject to local consensus. All this talk about CONLEVEL appears to assume that there is a community consensus to use the archive parameters regardless of any other factors or considerations, and I'm not aware of any such community consensus (feel free to correct me, but WP rarely imposes such bright lines for anything). As I've said, guidelines are only guidelines. ―Mandruss ☎ 23:12, 1 April 2024 (UTC)

  WP:CONLEVELS is that way, you're welcome to take this up with the arbitration committee. —Locke Cole • t • c 01:39, 2 April 2024 (UTC)

  Belatedly reviewing the 2017 and 2018 discussions, I don't see any talk about PEIS. Editors were concerned about raw file size, download time, and code clutter. The article currently sports 837 citations, virtually all CS1. I'm apparently conflating the two issues, but the PEIS limit has in fact been a serious problem and PEIS is in fact impacted by the archive parameters. So my above argument stands. ―Mandruss ☎ 07:33, 31 March 2024 (UTC)

• Back to the beginning of this discussion: The fine folks over at Talk:Donald Trump currently have a "Current consensus" item on their talk page that disallows including archive URLs for citations that aren't dead. None of the current 837 cites at Donald Trump are dead, as proven by the last time someone "rescued 291 sources and tagged 0 as dead" on March 12, 2024, and added 57,600 bytes to the page’s 430,000 bytes. The page mostly relies on "news reporting from well-established news outlets" (WP:NEWSORG), i.e., news articles that are permanent links and routinely and repeatedly archived on the Wayback Archive. I’m fine with a bot tagging an allegedly dead link or three and adding archive-urls, although they usually turn out not to be dead and easily found under their new url. Space4Time3Continuum2x🖖 11:50, 31 March 2024 (UTC)

  Please read the above discussion. —Locke Cole • t • c 17:14, 1 April 2024 (UTC)

  If you wanted the article's editors involved from the beginning, you should have notified us from the beginning, instead of relying on Nikkimaria to do that more than two months later.[2] Per community norms and basic ethics. Clearly, you didn't want the article's editors interfering with your agenda. Better yet, you could have raised this at the article's talk page (where the article's local consensuses are discussed) and posted a note here to bring in outside voices. ―Mandruss ☎ 00:59, 2 April 2024 (UTC)

  I didn't want you involved, actually. My question was less about this specific article and more about "should articles be deciding on their own" when WP:CONLEVELS is a thing. Nothing you've said here changes that. —Locke Cole • t • c 01:34, 2 April 2024 (UTC)

  I didn't want you involved, actually. Lol. Yeah, I got that. Nothing you've said here changes that. Your opinion. I think plenty of what I've said here changes that. You are free to assert your opinions in a local consensus discussion and, if you can persuade the majority, So Be It. That's how it works around here. ―Mandruss ☎ 03:02, 2 April 2024 (UTC)

  That's how it works around here. Somebody better tell WP:ARBCOM that their thoughts at WP:CONLEVELS aren't relevant to the editors at Talk:Donald Trump. After all, their opinion (Where there is a global consensus to edit in a certain way, it should be respected and cannot be overruled by a local consensus) is only eleven years old and been used repeatedly in ArbCom decisions... —Locke Cole • t • c 03:09, 2 April 2024 (UTC)

  I think you misunderstand the function of guidelines. As I've indicated, they are only guidelines and do not represent community consensuses that things must be done a certain way. They are one of the things to be considered in consensus discussions, not the only thing. As I've indicated, there is no community consensus that the archive parameters must be used for all live sources; that would have to be separate from the guidelines, as an RfC or something. Feel free to show me where ArbCom meant "global consensus" to be interpreted in that way.

  Again, you're free to assert your opinions in a local consensus discussion. Even if you're correct as to CONLEVEL, ArbCom, PEIS, or anything else, you still need to persuade the majority of the merits of your arguments. There are very few trump cards in Wikipedia editing. If you think other editors are too ignorant to be trusted with these decisions, you need an attitude adjustment.

  Me, I've been on the losing side of many debates where I "knew" we were in the right and the opposing side was "patently" wrong. It's a tough pill to swallow, but it gets easier with practice. I had to learn to let go and not care so damn much (and that benefited me in real life, too, so WP editing has been therapeutic for me). Welcome to Wikipedia. ―Mandruss ☎ 05:19, 2 April 2024 (UTC)

  @Locke Cole: - Thank you for raising the PEIS limit issue at VPT, at WP:Village pump (technical)#WP:PEIS. But reading it suggests that you expect that the archive parameters will be added if the limit is increased, no questions asked. As I said previously, Key word "revisit": upping the PEIS limit wouldn't guarantee a consensus change, since editors may still prioritize other factors over the avoidance of limited link rot. You would still need to get a new local consensus after the increase. I don't wish to be accused of moving goalposts. ―Mandruss ☎ 00:03, 3 April 2024 (UTC)

  WP:CONLEVELS. As we're starting to deviate into behavioral issues, I'll just open an WP:AN/I and go from there. Thank you for making it clear it's not about WP:PEIS though. —Locke Cole • t • c 05:19, 3 April 2024 (UTC)

  @Locke Cole: Further up the page, multiple editors expressed disagreement with your interpretation of the guideline at WP:DEADREF. You commented, "There appears to have been some misunderstandings above about what archiving a citation means [...]". Are you discounting the lack of consensus for your interpretation based on the belief that dissenting editors don't understand your position? I don't think that opening a discussion at ANI about Mandruss will have any positive effect in this discussion, Rjjⁱⁱⁱ (talk) 06:42, 3 April 2024 (UTC)

  Thank you for making it clear it's not about WP:PEIS though. That is not what I said. I said it's not only about PEIS. ―Mandruss ☎ 06:52, 3 April 2024 (UTC)

  No, it's not about PEIS at all, as evidenced here. When presented with a solution to a problem, your reaction is to revert it because... you want to maintain the status quo (which runs counter to WP:CONLEVELS). Then there's the WP:BATTLEGROUND behavior exhibited directly above with this quote: Me, I've been on the losing side of many debates where I "knew" we were in the right and the opposing side was "patently" wrong. It's a tough pill to swallow, but it gets easier with practice. I had to learn to let go and not care so damn much (and that benefited me in real life, too, so WP editing has been therapeutic for me). Welcome to Wikipedia. You're not here to improve the encyclopedia. You're here to win a battle and apparently welcome people to a project that have been here longer than you. I'm not seeing how that's civil at all. Like I said, there's a reason I didn't go out of my way to solicit comments from editors at Talk:Donald Trump here. You're proving my point in spades though. —Locke Cole • t • c 06:58, 3 April 2024 (UTC)

  It seems strange to accuse someone of wanting to win battles when they actually just mentioned they're used to losing debates (not battles). This is not constructive nor fair, and I suggest you drop it. Gawaon (talk) 07:14, 3 April 2024 (UTC)

  What precisely is constructive about welcoming an editor who has been here longer? Or bemoaning "lost battles" with the clear insinuation being that somehow I've lost? No, I won't be dropping it. —Locke Cole • t • c 20:21, 5 April 2024 (UTC)

Hi, I have a general question. Why is it required by the MOS to always put a citation immediately after the final character, instead of leaving a space in some cases? For the body of an article, I understand leaving no space. But for some areas, like an infobox, my humble opinion is that a space looks far better. Please see the infobox on this page. By the time you click the link, hopefully nobody has edited it, but currently some of the lines have spaces before citations and some don't. I may be in the minority, but I think when there is no space it looks dreadful, cluttered, and sometimes difficult to read if the word ends with a certain character, such as lowercase "i". If there's plenty of room for a space without messing up text or formatting, is there any flexibility for using spaces? Sorry, but this is just my pet peeve. I hate seeing those citations slammed up against the words when there is apparently no practical reason for it, other than adhering to a rigid policy. Wafflewombat (talk) 16:48, 20 April 2024 (UTC)

There's a note[3] in WP:Manual of Style#Punctuation and footnotes that suggests using a hair space for this purpose. -- LCU ActivelyDisinterested «@» °∆t° 20:29, 20 April 2024 (UTC)

Thanks for the tip! Very helpful to know. Wafflewombat (talk) 22:38, 20 April 2024 (UTC)