Wikipedia talk:Manual of Style/Linking
From Wikipedia, the free encyclopedia
| This is the talk page for discussing improvements to the Manual of Style/Linking page. |
|
| Archives: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21Auto-archiving period: 6 months |
| The contentious topics procedure applies to this page. This page relates to article titles and capitalisation. Editors who repeatedly or seriously fail to adhere to the purpose of Wikipedia, any expected standards of behaviour, or any normal editorial process may be blocked or restricted by an administrator. |
| This project page does not require a rating on Wikipedia's content assessment scale. It is of interest to the following WikiProjects: | ||||||||||||||
| ||||||||||||||
RfC: Linking of three-part place names
There are two common ways to link to a place name with an "A, B, C" format where the article is titled [[A, B]]. Both can be read as fair interpretations of the guidance to "link only the first unit".
- Have the link span only the smallest unit, using piping if necessary
- Buffalo, New York, United States
- Have the displayed text match the title of the linked article
- Buffalo, New York, United States
Which style(s) is/are acceptable? If both, is one preferable to the other?
Note: See previous discussion above and above. This is not a question about whether "New York" should be linked to New York (state) in this example; basically everyone agrees that it should not be. 20:57, 17 October 2024 (UTC)
Discussion re RfC: Linking of three-part place names
- Both acceptable, approach 1 preferable. Approach 2 is, no doubt, more common, but both approaches are used in good and featured articles without issue. As a matter of MOS:RETAIN I'll stop short of saying approach 2 should be proscribed, but I think approach 1 is preferable for two reasons:
- Consistency: Having a prose guideline turn on the title of the article being linked to would be strange, given that the article title policy is informed by various considerations that do not apply to prose, such as disambiguation and the semi-arbitrary rule that is WP:USPLACE. To a reader seeing "Buffalo, New York, United States", next to "Boston, Massachusetts, United States", it is not at all obvious why the two are handled differently. It is cleaner and simpler to have the link span the exact place being referenced, not attached disambiguators like ", New York".
- Accessibility: The only difference between "Buffalo, New York" and "Buffalo, New York" is the color of the comma. For anyone who, like me, struggles to distinguish between blue and black in small quantities, it looks like clicking on "New York" in the first example will take you to New York (state).
- The main argument made in the opposite direction is simplicity of markup, but that's usually the lowest priority in MoS decisions, certainly lower than accessibility. We should not make our articles more confusing to readers just for the sake of slightly shorter source code. -- Tamzin[cetacean needed] (they|xe) 20:57, 17 October 2024 (UTC)
- None -- I struggle to understand why it wouldn't be necessary to link the first-level administrative division, in most cases outside the US (and even the US as well, but let's assume that Buffalo, New York is an accepted practice in that context). Who could be expected to consider Ialomița County or Simeulue Regency instantly recognizable terms across the vast expanse of the world? and if we're not linking unfamiliar terms, what is the point of having internal links at all? Seems like someone was peeved by having two links next to each other, and came up with this atrocious moratorium on having necessary links where they appear side by side (though neatly separated by a comma); this bewildering approach should not have been tried out at all, ever. Dahn (talk) 21:10, 17 October 2024 (UTC)
- The normal argument against linking the second-level entity is that it can be easily clicked from the first-level, if some wants. As discussed above, exceptions may apply when the first-level entity's article doesn't prominently discuss the second-level one, mostly in the case of former countries. -- Tamzin[cetacean needed] (they|xe) 21:17, 17 October 2024 (UTC)
- The exceptions are in fact the norm -- most subdivisions would be unfamiliar to anyone outside that country. Which is why "Buffalo, New York" is a misleading example, the sort of which has prompted some overzealous users to delete links to Olt County and Wallachia, thus leading to the absurd suggestion that Olt County has the same notoriety as New York, and Wallachia is a notion similar to the US. "It can be easily clicked from [somewhere else]" can be said about each and every bluelink out there, so I don't see why that was ever accepted as a valid argument in any debate. Dahn (talk) 21:32, 17 October 2024 (UTC)
- The normal argument against linking the second-level entity is that it can be easily clicked from the first-level, if some wants. As discussed above, exceptions may apply when the first-level entity's article doesn't prominently discuss the second-level one, mostly in the case of former countries. -- Tamzin[cetacean needed] (they|xe) 21:17, 17 October 2024 (UTC)
- Option 2 per MOS:SPECIFICLINK. It's also a normal unpiped link, without superfluous text: compare
[[Buffalo, New York]], United States(five words) with[[Buffalo, New York|Buffalo]], New York, United States(eight words). --Redrose64 🌹 (talk) 22:13, 17 October 2024 (UTC) - Comment — there are plenty of situations where linking place, subdivision and country is appropriate, and I think the guideline should do more to encourage that. Examples: 1) Bogdana, Tutova County, Moldavia; 2) Haraklány, Szilágy County, Kingdom of Hungary. It’s more than likely the average reader will have no idea where any of these places are/were, so why not link them all? — Biruitorul Talk 05:59, 18 October 2024 (UTC)
- Option 2 because it's shorter to write and leads to linked text and linked page title being in agreement. Later addition: Also per WP:NOPIPE, as pointed out below by Bagumba – don't use piped links when you don't have to, and here you very clearly don't have to. Gawaon (talk) 08:55, 18 October 2024 (UTC), edited 07:55, 19 October 2024 (UTC)
- Both acceptable, do not specify preference for either. I personally prefer Option 2, which cuts down on redundant text that looks extremely silly in the editor and in diffs. I suppose it also matches linktext with article titles, which I care less about. I don't think we should enshrine a preference for best practice here. Agree with others above that in many cases it may be helpful to link multiple administrative subdivisions: not long ago I had reason to mention
Yao Mangshan Ethnic Township (莽山瑶族乡), Yizhang County, Chenzhou, Hunan
. Leaving out the container state, that's still four subdivisions. I left Hunan unlinked since it appears in User:Ohconfucius/script/Common Terms, but there are probably editors who would argue for linking that as well. Folly Mox (talk) 09:52, 18 October 2024 (UTC) - Link only the most specific item—especially when the other two are so well known. Tony (talk) 10:25, 18 October 2024 (UTC)
- Option 2 Makes no sense to pipe and hide "New York", just to type it again and display it. Per WP:NOPIPE:
—Bagumba (talk) 10:54, 18 October 2024 (UTC)Unnecessary piping makes the wikitext harder to read.
- Option 2 don't find any specific reason to leave out the state from the muncipality, as it is kind of self explanatory. Also creates redundant piping per Bagumba. Takipoint123 (talk) 02:20, 19 October 2024 (UTC)
- Option 2 In this specific format, it seems more intuitive to match the title of the article. I will also add that including the non-linked country at the end may be somewhat out of place/redundant in either option. Symphony Regalia (talk) 14:38, 19 October 2024 (UTC)
- No preference. MOS should state this. I fully agree with Folly Mox here and would go one step further to say the style guide should be explicit in stating there is no dictated preference. It should list some things to consider, provide examples, and otherwise defer to editorial judgment. Things to consider might include MOS:NOPIPE and other rules or guidelines. A lot of this will come down to context-specific factors and personal judgment or consensus within an article. In nearly all cases it matters too little to mandate a single standard and doing so will likely result in more appeals for exceptions and workarounds. MYCETEAE 🍄🟫— talk 22:03, 19 October 2024 (UTC)
- Option 1, but both acceptable, per Tamzin and link intuitiveness. I don’t want people clicking “New York” and being confused at being sent to Buffalo. I also think all arguments based on what looks best in wikitext or is easier to type for the editor are wrong. Style decisions are not made for the wikitext editor’s benefit. — HTGS (talk) 00:50, 23 October 2024 (UTC)
I don’t want people clicking “New York” and being confused at being sent to Buffalo
: But that is exactly why we avoid consecutive links to begin with i.e. SOB. It is a single link to <city, state>, not consecutive links <city>, <state>. —Bagumba (talk) 04:37, 24 October 2024 (UTC)- And avoiding links that span two page-level topics is another step we can take towards making links clearer. — HTGS (talk) 02:06, 25 October 2024 (UTC)
- Right. Readers have no reason to expect that "New York" won't link to New York there: They don't know that MoS says it shouldn't, and in practice countless articles would link to New York there. Using a different state because the NYC/NYS ambiguity complicates things, there are 11,030 articles containing either
[[Boston]], [[Massachusetts]]or[[<someplace>, Massachusetts|<someplace>]], [[Massachusetts]]. These links are distinguished from e.g.[[Boston, Massachusetts]]by the color of a character that is less than a millimeter wide at standard zoom. -- Tamzin[cetacean needed] (they|xe) 00:20, 27 October 2024 (UTC)- Regardless of the outcome of this RfC, the standalone link to Massachusetts should be unlinked per MOS:GEOLINK in all these cases. MOS:GEOLINK is already very clear on this and it's not something that will change. Gawaon (talk) 08:39, 27 October 2024 (UTC)
- Right. Readers have no reason to expect that "New York" won't link to New York there: They don't know that MoS says it shouldn't, and in practice countless articles would link to New York there. Using a different state because the NYC/NYS ambiguity complicates things, there are 11,030 articles containing either
- And avoiding links that span two page-level topics is another step we can take towards making links clearer. — HTGS (talk) 02:06, 25 October 2024 (UTC)
- Single link In almost every case the purpose of the link is to take you to the article of a single, unambiguous, location. The link should be written in it's natural format (no piping). The larger regions are merely so that a printed form will lead you to the same place but we don't really expect the reader to want to go directly to the articles for the larger regions - ie, we are listing a city for a reason, the larger regions are just to make it unambiguous and are not a target in their own right. So, we give the link to the city in its natural format (without piping), and then add whatever else is needed in plain text. If it turns out that some cities in a list have the link encompass different portions of the hierarchy (eg Paris, France vs Paris, Ontario, Canada) then that is okay. Stepho talk 01:21, 23 October 2024 (UTC)
- Ooh, I really disagree with that last point. I’d rather a list be consistent regardless the choice between these two options. — HTGS (talk) 03:01, 23 October 2024 (UTC)
- How would you list those 2 cities? Stepho talk 03:10, 23 October 2024 (UTC)
- Assuming it’s a normal prose sentence, I would have something like:
“However in 1894, the government of Paris, France decided to implement the change, while the mayor of Paris, Ontario forced the city to withhold …”
But honestly I would still rather the opposite (… Paris, France decided to implement the change, while the mayor of Paris, Ontario did not…”
) to the split styling you suggested. — HTGS (talk) 21:40, 24 October 2024 (UTC) - And for most lists, the same (with disambiguation pages being the exception). — HTGS (talk) 21:42, 24 October 2024 (UTC)
- Assuming it’s a normal prose sentence, I would have something like:
- How would you list those 2 cities? Stepho talk 03:10, 23 October 2024 (UTC)
- I agree, but note that what you describe is in fact exactly Option 2 ("Have the displayed text match the title of the linked article"), so you're effectively voting for that. Gawaon (talk) 07:23, 23 October 2024 (UTC)
the larger regions are just to make it unambiguous and are not a target in their own right
: I'm not sure if "unambiguous" is the right word. For a large country, most people have never heard of most non-major cities, so a larger region is mentioned to provide context, whether or not the same city name exists in another region.—Bagumba (talk) 04:48, 24 October 2024 (UTC)
- Ooh, I really disagree with that last point. I’d rather a list be consistent regardless the choice between these two options. — HTGS (talk) 03:01, 23 October 2024 (UTC)
- Definitely Option 2. No pipe gymnastics needed, and the blue the reader sees tells him unambiguously where clicking will take him. EEng 00:33, 27 October 2024 (UTC)
- Link "Buffalo" alone. Tony (talk) 02:25, 1 November 2024 (UTC)
- As in
Buffalo, New York, United States
? -- Michael Bednarek (talk) 03:01, 1 November 2024 (UTC)
- As in
- Comment: Hi folks. I came here to raise a closely related point, then I saw this discussion and the previous ones. I think the examples should be changed to allow or encourage this type of thing:
Three very nice cities are:
- Madison, Wisconsin, US
- Terre Haute, Indiana, US
- Chicago, Illinois, US
- That is, in many cases it's preferable to be consistent with how the links are presented, and in my view it's *not* necessary to have the visible linked text exactly match the article titles. So in this example I've coded
[[Chicago|Chicago, Illinois]]to achieve that. Although coding[[Chicago, Illinois]]would achieve more or less the same thing, because "Chicago, Illinois" is a redirect to "Chicago". Edited to add: This suggestion does not match either option 1 or option 2. — Mudwater (Talk) 01:55, 26 November 2024 (UTC)
- At least correct the description of what is recommended: To me, it is false to say "Both can be read as fair interpretations of the guidance to 'link only the first unit'." In the string "Buffalo, New York, United States", there are clearly three units, and the first of those three units is "Buffalo". If we're going to say that "Buffalo, New York, United States" (
[[Buffalo, New York]], United States) is the preferred format, we need a different characterization than saying that for "a sequence of two or more territorial units, link only the first unit". For example, we could say to "link only as much of the name as is used in the corresponding article title" or "link only the initial parts of the name that form its conventional identification". (We might also need to refer the reader to MOS:USPLACE for the conventional form of US location descriptions). — BarrelProof (talk) 01:06, 3 January 2025 (UTC) - Comment: In most cases no one is ever going to link to "Terre Haute", as suggested above, just because the subject was born there. Who cares? (Unless that town had significant bearing on the notability of the subject.) So often we are faced with the logic of not linking because of insufficient relevance, or because the location is internationally known: the smaller and least consequential "village" ("Terre Haute") vs the too-well-known larger location ("Chicago"). In that case, nothing seems to need linking. Another example: "suburb of London, London, UK"—link nothing, unless the suburb has sufficient relevance to the subject (unlikely).
- There are cases that could be linked as a matter of logic. Let's say the formative years were spent in the village of birth: Adalaj, Gujarat, India. Here, the article on Adalaj will reasonably contain a link to Gujarat, if the reader really wants to know more about the state. Remember that the one in 10,000 readers who really do want to know more, in situ, can spend 10 seconds typing a target into the search box. Otherwise we have systemic bunching, which MOSLINK discourages for good reason. Tony (talk) 23:42, 17 January 2025 (UTC)
- I'm fairly sure we have decided in the past to only visibly link to the smallest unit. The reason being that the difference between a link to the smallest unit and it's next superunit, and two links is merely the colour of the comma between the two place names. People will click on the superunit expecting to be taken there, and get a WP:Easter egg. I have done this myself. The priority needs to go to the reader, not the editor here. If you prefer less typing, go ahead, but don't oppose others improving it. All the best: Rich Farmbrough 20:34, 1 March 2025 (UTC).
Meta-discussion
LINKONCE?
There is an article on which someone posted on the talk page that a (brief) level 2 section is confusing because not only do many of the previously mentioned names have links to their articles, many leave out the given name as well, requiring a good deal of effort on the editor's part to figure out what is going on. This comment has been receiving a good deal of resistance on the talk page; I happen to agree with the OP and think "Link a term at most once per major section (Major sections are generally detailed sections with a level-2 heading...) , at first occurrence. " suggests that they are not incorrect, but I would like clarification here if any can be offered. Thanks... Tduk (talk) 18:42, 13 October 2025 (UTC)
- Opinions depend upon context. If you want another opinion, post the article name. Johnjbarton (talk) 21:43, 13 October 2025 (UTC)
- Sure, thanks.. Agent Carter (TV series), also the discussion is on the talk page there. Tduk (talk) 22:08, 13 October 2025 (UTC)
- The core issues on that page are unrelated to LINKONCE. For example, "Ryan" isn't even linked once elsewhere in the article because that person isn't even notable. Johnjbarton (talk) 16:50, 14 October 2025 (UTC)
- Sure, thanks.. Agent Carter (TV series), also the discussion is on the talk page there. Tduk (talk) 22:08, 13 October 2025 (UTC)
- Linking once in each level-2 section is explicitly allowed by our current policy even if the same article was already linked in earlier sections, so I don't understand what's the issue here? Gawaon (talk) 07:40, 14 October 2025 (UTC)
- The issue was that the section did NOT link once, and did not include the full names of the people involved, so it was very confusing if one is looking only to read that section. If this is the expected norm, where do I raise an issue with this - because not only is it difficult for those readers, it creates some accessibility issues, requiring the reader to do more work than I think is necessary. Tduk (talk) 13:19, 14 October 2025 (UTC)
- MOS:LINKONCE doesn't say that all terms should or must be linked in every major section in which they appear, nor is that our practice, nor do we privilege that hypothetical reader who dives into the middle of an article, expects everything to be made clear there and then, and is unwilling to scroll up even if fully understanding some later part of an aricle would be helped by reading an earlier part. NebY (talk) 14:00, 14 October 2025 (UTC)
- Are there use case studies to back up your implication that most people who read wikipedia read the entire article in one sitting? Setting aside the accessibility issues for now. Tduk (talk) 14:53, 14 October 2025 (UTC)
- That is your inference, not my implication. NebY (talk) 15:03, 14 October 2025 (UTC)
- What is your implication from "nor do we privilege that hypothetical reader who dives into the middle of an article, expects everything to be made clear there and then, and is unwilling to scroll up even if fully understanding some later part of an aricle would be helped by reading an earlier part."? Tduk (talk) 15:27, 14 October 2025 (UTC)
- That is your inference, not my implication. NebY (talk) 15:03, 14 October 2025 (UTC)
- @NebY: We literally have architecture to force a reader to dive into a particular part of a section and skip over the rest of the article: {{Anchor}}. The culture of linking to specific parts of a page has gotten so bad that Template:ATA shortcut notice had to be made for our own internal policies to remind people to read the start. If you want a case study, just look at how we can't even get editors who should know better to read the start of pages and somehow we're expecting this of our readers? Without even any tooltip that we've dumped them in the middle of an unfamiliar land? I think that's unfair. – Mullafacation『talk』 17:11, 15 January 2026 (UTC)
Comment: See also /Archive 8#Address hyperlink-era usage patterns in "Repeated links" section – Mullafacation『talk』 17:20, 15 January 2026 (UTC)
- Are there use case studies to back up your implication that most people who read wikipedia read the entire article in one sitting? Setting aside the accessibility issues for now. Tduk (talk) 14:53, 14 October 2025 (UTC)
- Since the MOS allows, but doesn't not require wikilinks to be repeated once per major section, this doesn't seem a MOS issue at all. Gawaon (talk) 15:40, 14 October 2025 (UTC)
- MOS:LINKONCE doesn't say that all terms should or must be linked in every major section in which they appear, nor is that our practice, nor do we privilege that hypothetical reader who dives into the middle of an article, expects everything to be made clear there and then, and is unwilling to scroll up even if fully understanding some later part of an aricle would be helped by reading an earlier part. NebY (talk) 14:00, 14 October 2025 (UTC)
- The issue was that the section did NOT link once, and did not include the full names of the people involved, so it was very confusing if one is looking only to read that section. If this is the expected norm, where do I raise an issue with this - because not only is it difficult for those readers, it creates some accessibility issues, requiring the reader to do more work than I think is necessary. Tduk (talk) 13:19, 14 October 2025 (UTC)
Always subst the anchor template within a section header?
WP:ANCHORSUBST explains that inside section headers, {{Anchor}} should always be substituted (and not a number of alternative approaches).
I asked at that page, and was directed here. Which is probably for the best since that page is an "ordinary" help page and can't prescribe policy anyway.
Could anyone point me to the discussion that is the basis for this recommendation? Thank you CapnZapp (talk) 21:19, 18 October 2025 (UTC)
- The rationale is already given in WP:ANCHORSUBST, which discusses each possible alternative in detail, explaining its problems. Gawaon (talk) 07:17, 19 October 2025 (UTC)
- Sure, but that's not what I'm asking for, Gawaon. I'm asking for the discussion or discussions where the current consensus was hashed out, where - I imagine - each approach's disadvantages (and advantages) were weighed against each other, before the current prescribed approach was selected. CapnZapp (talk) 09:23, 19 October 2025 (UTC)
- Template:Anchor/doc was amended after this discussion; I can't say if that's the only or most relevant one for your investigation. NebY (talk) 11:02, 19 October 2025 (UTC)
- The consensus in that discussion seems to be against substitution of the anchor template within the section heading. The discussion also doesn't include contemplation of putting the substituted span in the heading before the heading itself.
- I'm starting to wonder if we walked step by terrible step into an awful solution without ever having a discussion and arriving at a consensus. The current recommendation may work well technically and for readers, but is awful for editors:
=== <span class="anchor" id="Administrator"></span><span class="anchor" id="Admin"></span><span class="anchor" id="Sysop"></span><span class="anchor" id="sysop"></span><span class="anchor" id="admin"></span>Administrators and bureaucrats ===
- We should consider banning section anchors altogether and having a bot fix links to outdated section headings. --Srleffler (talk) 14:12, 19 October 2025 (UTC)
- Just to quickly say that at the very least, let us discuss this proposal separately. Personally I find the ability to safeguard incoming section links so they keep working even if well-meaning editors rename sections is crucial. But again, please let us have that discussion separately. Cheers CapnZapp (talk) 11:07, 20 October 2025 (UTC)
- That discussion was about editing the documentation of a template, which isn't a guideline, to align it with a guideline that already existed. The actual consensus was likely formed on the various policy pages which recommended the practice of subst'ing the template even before that discussion took place. These are the ones I could find:
- MOS:RENAMESECTION
- MOS:HEAD, specifically:
- WP:TARGET
- FaviFake (talk) 14:20, 19 October 2025 (UTC)
- The guidance to substitute the template if it was in the heading has been well established. What I haven't seen yet is the origin of the mandate that it must be substed in the heading before the text of the heading. In the discussion linked above, nobody was discussing that as an option, and there was clearly not agreement on mandating it having to be in the heading at all.
- Looking at those pages:
- MOS:RENAMESECTION was edited by you in August, referencing WP:TARGET
- WP:TARGET said to put the anchor after the section name until it was edited by you, again in August, as a "Bold edit inspired by User:Redrose64's comment on {{Anchor}}'s talk page"
- The relevant comment appears to be at the end of Template talk:Anchor, in the section "Are we sure WP:ANCHORSUBST is a good idea?" Other editors in that discussion favored options other than including the anchor in the heading; there was no consensus there for Redrose64's suggestion.
- You also edited MOS:SECTIONANCHOR that same day in August
- So, it appears that this change did not come from a consensus or from the MOS but rather was boldly inserted into both the MOS and WP:TARGET by you based on one editor's comment. I know you meant well here, but it's time to stop and look at where we ended up and decide whether we took a wrong turn. --Srleffler (talk) 02:44, 20 October 2025 (UTC)
- Yes, I did boldly edit the recommendation a while back, based on the recommendation in this comment.
A wider discussion on this may be needed but I think thatI've said it before, and I'll say it again: accessibility. Users of screen readers rely on anchors to be at the point from which reading should commence. Consider a section heading: if the anchor is after the heading, or even within the heading but after the heading text, the screen reader softare will not read out the heading text, but will begin with the text that follows the heading. Therefore, the anchor must be inside the heading and also before the heading text, so that the latter is read out.
banning section anchors altogether
is a tad bit too extreme. I would fully support a bot that would do this though! FaviFake (talk) 15:37, 20 October 2025 (UTC)
- Yes, I did boldly edit the recommendation a while back, based on the recommendation in this comment.
- Template:Anchor/doc was amended after this discussion; I can't say if that's the only or most relevant one for your investigation. NebY (talk) 11:02, 19 October 2025 (UTC)
- Sure, but that's not what I'm asking for, Gawaon. I'm asking for the discussion or discussions where the current consensus was hashed out, where - I imagine - each approach's disadvantages (and advantages) were weighed against each other, before the current prescribed approach was selected. CapnZapp (talk) 09:23, 19 October 2025 (UTC)
I'm starting to wonder if we walked step by terrible step into an awful solution without ever having a discussion and arriving at a consensus.
This is exactly why I asked for previous discussions as my first step...! :-) CapnZapp (talk) 10:48, 20 October 2025 (UTC)
I fully agree that the current recommendation/prescription is awful for readers. I fully understand how the current recommendations could have come to be; editors active in these discussions definitely lean towards the tech savvy (and tech tolerant) side of things, I know, because I am one myself. However, what cannot be taken for granted is technical experts' ability to accurately judge what regular people will accept when it comes to jargon or code. It is a complete disservice to readers, simply put. My first instinct, before realizing how we might have gotten here, was to simply (and boldly) replace the current recommendation, recommending transclusion instead of substitution because the only reason given against it, "it messes up edit summaries", to me appears much much more tolerable to the average reader (but crucially looks ugly to tech-savvy editors). After all, regular web users are able to replace the prefilled contents of the edit summary text box if they don't like it. For instance, the prefilled contents of the edit summary for THIS post is long: "/* Consensus for the current recommendation to always subst the anchor template within a section header */" I do not believe it is nearly as confusing for a reader to just mark that text before writing their edit summary - replacing the existing text altogether - to avoid it being so long. (If you look at the edit summary of this post it will read simply "my own view" just to demonstrate). In contrast, asking average readers to tolerate HTML comments and to know enough to disregard them without worrying you've done something wrong, is just not a reasonable ask of the average Wikipedia reader. Regards CapnZapp (talk) 11:03, 20 October 2025 (UTC)
- Just for the record, it seems like bad advice to me. We should avoid having any more html or css in articles than absolutely necessary. All the best: Rich Farmbrough 11:39, 20 October 2025 (UTC).
- I think we can all agree the current method isn't great. The question is, what should it be changed to? Or rather, which of tradeoffs listed at WP:ANCHORSUBST should we accept just to have better wikicode readability? FaviFake (talk) 16:45, 20 October 2025 (UTC)
- Not commenting on the rest, but the problem listed at WP:ANCHORSUBST isn't that "it looks ugly". It's more functional:
FaviFake (talk) 16:43, 20 October 2025 (UTC)If it is not manually removed every time, the browser may not return the user to the section and the section link of that edit in the history page will be broken.
- I don't actually understand why I should care about that. Being automatically returned to the section you just edited and having a link to the section in the history seem like really trivial conveniences to me; hardly worth worrying about. Having headers that start with a line or two of HTML gobbledigook seems like a much bigger problem. --Srleffler (talk) 04:49, 21 October 2025 (UTC)
- Maybe you don't care, but I for one would sure hate to lose that convenience, trivial or not. Gawaon (talk) 08:14, 21 October 2025 (UTC)
- "having a link to the section in the history" Is not just about convenience, it's also about telling folks who view the history what part of the article was changed (without them having to open up the edit to see). That is a non-trivial time-saver for history-viewing editors. - Butwhatdoiknow (talk) 12:58, 21 October 2025 (UTC)
- I don't think that's relevant. Even if the anchor is transcluded, the edit summary will still accurately describe what section was edited. It just won't provide a functional link to it. A minor inconvenience, but not a loss of information about what part of the article was changed.--Srleffler (talk) 04:00, 22 October 2025 (UTC)
- If someone edits multiple single sections of an article, it would be more annoying if the software sent them to the top of the page every time they clicked publish rather than if they took 1 more second to read the name of the section. In my opinion, at least. FaviFake (talk) 15:54, 21 October 2025 (UTC)
- I don't actually understand why I should care about that. Being automatically returned to the section you just edited and having a link to the section in the history seem like really trivial conveniences to me; hardly worth worrying about. Having headers that start with a line or two of HTML gobbledigook seems like a much bigger problem. --Srleffler (talk) 04:49, 21 October 2025 (UTC)
- In my experience, using section links from my watchlist or a history page is something that happens multiple times every day. I would prefer to keep that functional. I very rarely edit section headings—maybe once a month or so—and I've probably only edited a heading that includes the gobbledygook a dozen or so times ever. We generally prefer to simplify wikitext where possible, but functionality is more important. I hadn't thought before about whether anchors should come before or after the current heading, but it seems like after would be more editor friendly. I'm sorry if I've missed the reasons for preferring it the other way around. Firefangledfeathers (talk / contribs) 15:04, 21 October 2025 (UTC)
- Sounds like the summary regarding location is:
- Before the heading: Pros: Links to a good place. Mostly sensible wikitext. Cons: Doesn't open the collapsed section on mobile. Not visible in section editing.
- Inside the heading, before the text: Pro: Most functional for screen readers and mobile. Con: Ugly wikitext. If not substed, breaks auto-summary links.
- Inside the heading, after the text: Pro: Somewhat less ugly wikitext. Con: Screen readers don't read the heading text. If not substed, breaks auto-summary links.
- After the heading: Pro: Sensible wikitext. Cons: Link puts the heading off the top of the screen. Screen readers don't read the heading.
- Thanks for the summary. With prompting, I do remember the accessibility point. I favor the status quo here. Let's maximize functionality. Firefangledfeathers (talk / contribs) 15:23, 21 October 2025 (UTC)
- I don't think this discussion relates to location. Rather, is is about whether anchors placed inside the heading should be transcluded or substituted. To learn about the difference, see the first two paragraphs of Wikipedia:Substitution. - Butwhatdoiknow (talk) 20:53, 21 October 2025 (UTC)
- Seems to me it's about both "always subst" and "within the section header". Anomie⚔ 21:58, 21 October 2025 (UTC)
- It's more than both those things. It's about "always subst" and "within the section heading" and "at the start of the header". What triggered this discussion was a change made without consensus from "always subst within the heading, at the end of the heading" to "always subst within the heading, at the start of the heading". This change has negative consequences and was pushed out to multiple MOS and guidance pages by one editor in one day, with no discussion. So, location is very definitely a key part of this discussion. --Srleffler (talk) 04:11, 22 October 2025 (UTC)
- Yes, those are issues raised by the original edit, but are they issues pertinent to this particular discussion?
- This section begins with the following sentence: "WP:ANCHORSUBST explains that inside section headers, {{Anchor}} should always be substituted (and not a number of alternative approaches)." The discussion that follows talks about transclusion vs. substitution when Anchor is placed inside a heading.
- To avoid this already long discussion from dealing with too many issues at once, please raise other issues separately, perhaps by starting subsections from Transclusion vs. substitution. - Butwhatdoiknow (talk) 06:12, 22 October 2025 (UTC)
- The title of this section says (emphasis added) "for the current recommendation to always subst the anchor template within a section header". The "(and not a number of alternative approaches)" you quote encompasses the other positionings. Attempting to artificially limit the discussion isn't going to help. Anomie⚔ 12:59, 22 October 2025 (UTC)
- All bold edits are
made without consensus
. That's the point of WP:BOLD... FaviFake (talk) 15:50, 22 October 2025 (UTC)- Yes, and note WP:BRD. Bold edits are often summarily reverted. I have not done that, preferring to bring this issue to a discussion first. I have emphasized that this was a bold edit, made without consensus, because I want it to be clear that the mandate to put the substed anchor at the start of the heading is new, and not something that has been discussed before. --Srleffler (talk) 06:09, 24 October 2025 (UTC)
- My understanding is that it was the result of a previous discussion (if maybe not a very long one). Gawaon (talk) 07:20, 24 October 2025 (UTC)
- Yes, in a way. The change was supported by at least one other editor at the time of my bold edit. FaviFake (talk) 14:20, 24 October 2025 (UTC)
- My understanding is that it was the result of a previous discussion (if maybe not a very long one). Gawaon (talk) 07:20, 24 October 2025 (UTC)
- Yes, and note WP:BRD. Bold edits are often summarily reverted. I have not done that, preferring to bring this issue to a discussion first. I have emphasized that this was a bold edit, made without consensus, because I want it to be clear that the mandate to put the substed anchor at the start of the heading is new, and not something that has been discussed before. --Srleffler (talk) 06:09, 24 October 2025 (UTC)
- It's more than both those things. It's about "always subst" and "within the section heading" and "at the start of the header". What triggered this discussion was a change made without consensus from "always subst within the heading, at the end of the heading" to "always subst within the heading, at the start of the heading". This change has negative consequences and was pushed out to multiple MOS and guidance pages by one editor in one day, with no discussion. So, location is very definitely a key part of this discussion. --Srleffler (talk) 04:11, 22 October 2025 (UTC)
- Seems to me it's about both "always subst" and "within the section header". Anomie⚔ 21:58, 21 October 2025 (UTC)
- Pinging @Redrose64 in case they wish to expand their previous comment around the accessibility of subst'ed anchors. Or partecipate in the discussion in general. FaviFake (talk) 15:48, 21 October 2025 (UTC)
- Sounds like the summary regarding location is:
Comment: While I prefer another recommendation, I consider FaviFake's change to be a perfectly reasonable bold edit. He hardly edited "without consensus." For whatever this is worth, I consider us to be in the BRD cycle, except without the R step (now that we're discussing). CapnZapp (talk) 14:47, 22 October 2025 (UTC)
Transclusion v. Substitution for anchors placed in headings
Please edit this post freely and discuss below.
Edit: The location of the anchor gets a separate discussion; below. CapnZapp (talk) 14:43, 22 October 2025 (UTC)
============
- Advantages of Transclusion:
- Cleaner section headings when editing.
- Disadvantages of Transclusion:
- Adds anchor template(s) text to edit summary link.
- Breaks edit summary link to edited section.
- Advantages of Substitution:
- Preserves edit summary link.
- Disadvantages of Substitution:
- Adds ""HTML gobbledigook" to section headings when editing.
============ - Butwhatdoiknow (talk) 14:46, 21 October 2025 (UTC)
- That doesn't really sum up the issue. Anomie had it better, above:
- Before the heading
- Pros: Links to a good place. Mostly sensible wikitext. Cons: Doesn't open the collapsed section on mobile. Not visible in section editing.
- Inside the heading, before the text
- Pro: Most functional for screen readers and mobile. Con: Ugly wikitext. If not substed, breaks auto-summary links.
- Inside the heading, after the text
- Pro: Somewhat less ugly wikitext. Con: Screen readers don't read the heading text. If not substed, breaks auto-summary links.
- After the heading
- Pro: Sensible wikitext. Cons: Link puts the heading off the top of the screen. Screen readers don't read the heading.
--Srleffler (talk) 04:17, 22 October 2025 (UTC)
- Anomie's summary covers multiple issues. I have amended the heading of this subsection to clarify that it only deals with Anomie's "If not substed, breaks auto-summary links" issue (which, as my summary shows, is not a complete statement of that issue). - Butwhatdoiknow (talk) 06:18, 22 October 2025 (UTC)
- Thanks for this summary of the various options. Based on it, it seems clear to me that the recommended status quo (substed at start of heading) is indeed the best solution. It has one downside (ugly and hard-to-understand wikitext) but that one's less severe than the downsides of the alternatives, which would all entail loss of important functionality for either editors or readers, especially readers relying on screen readers. Gawaon (talk) 06:39, 22 October 2025 (UTC)
- Hold on, is that really a genuine representation of the situation. You are, if I'm understanding you correctly, making the following claims. Let's discuss each one separately:
- Location inside and at the start of heading. I have no issues with this. It is the best (or least worst) location option.
- The downside isn't merely "ugly and hard-to-understand wikitext". For starters, it isn't regular wikitext (H:MARKUP), it's pure CSS code (part of H:HTML)! Presenting editors (including newbies) with straight up programming code is in my view not an acceptable downside. It is the reason I'm starting this entire discussion. Experienced editors and coders (like the majority of you participating in this discussion?) very likely severely underestimates how user-hostile this really is.
- You aren't actually enumerating the downsides of the alternatives, which I really wished you would. If there's one I am forgetting about, you know why.
- As for screen readers, the downside there appears to be related to the location of the anchor, not whether it's substed or not. Please don't argue for the status quo based on conflated arguments.
- The downside that you're not taken directly to a section from page history or your watchlist (breaking edit summary links): While I don't use these links myself, I do acknowledge the usefulness of such a function, but I think sacrificing newbies for the convenience of experienced editors is the wrong way to go. (As an aside, edit summary links aren't even colored blue in the watchlist, meaning there is no visual indicator the edit summary contains a link unless you hover over it. This makes me suspect it isn't meant to be core functionality but more of an editor perk. We can live without editor perks, at least when it significantly decreases user hostility.)
- All in all, I must confess I find your summary to have an unfortunate (and hopefully unintended) appearance of being geared towards making readers inclined to agree with it, rather than examining every aspect equally and neutrally. With respect, CapnZapp (talk) 15:18, 22 October 2025 (UTC)
- Are you addressing me or Srleffler? I was merely drawing my own conclusion from Srleffler's summary, which I'd consider fair and adequate. If one can get others to agree, that is, of course, always a welcome side-effect of expressing one's opinion (so certainly not "unfortunate" nor "unintended"). Gawaon (talk) 15:29, 22 October 2025 (UTC)
- (edit conflict) Allow me to repost Srleffler's example from above. I wholeheartedly agree the following is awful for (new) editors, and I sincerely hope we can prioritize the new editor experience over some convenience shortcuts:
- Hold on, is that really a genuine representation of the situation. You are, if I'm understanding you correctly, making the following claims. Let's discuss each one separately:
=== <span class="anchor" id="Administrator"></span><span class="anchor" id="Admin"></span><span class="anchor" id="Sysop"></span><span class="anchor" id="sysop"></span><span class="anchor" id="admin"></span>Administrators and bureaucrats ===- I genuinely ask everybody to consider the choice we're about to make here not as a coder or experienced wikipedian, but as way to warmly greet new arrivals to Wikipedia. Best, CapnZapp (talk) 15:31, 22 October 2025 (UTC)
- This is indeed unpleasant to read, but having five anchors for the same section will in any case be very rare. Also I imagine that newbies using the source-code editor will fairly quickly learn to ignore the HTML stuff and focus on wiki-formatted content instead. There are other situations, such as tables with CSS styling, where the situation is the same. Gawaon (talk) 17:24, 22 October 2025 (UTC)
- Having 2+ anchors in the same section doesn't seem that rare to me though, and results in ugly, verbose, and unclear wikitext. Example. {{Anchor}} is very clear.
<span class="anchor" id="Reviewer"></span>is unclear to most people that this is there to be an anchor, unless you're a web programmer or very familiar with how anchors work. –Novem Linguae (talk) 17:40, 22 October 2025 (UTC) - Yes, tables are hard, but headers usually aren't. Bizarre code in them is perturbing even to moderately experienced editors like me; we might be deterred from rewording a section header - if we can even find it in all that - or we might even assume that we're looking at some weird glitch, clean away the entire mess and return the header to normal. NebY (talk) 18:12, 22 October 2025 (UTC)
- Yeah. In a way it also helps newer editors, as the VE hides old anchor HTML tags but displays templates. I think the status quo is the best option we have, but I agree it's not great. FaviFake (talk) 19:01, 22 October 2025 (UTC)
- I tried using VE to reword a header that had (invisibly to VE) an HTML anchor embedded between "==" and the wording. Doing that deleted the anchor. I tried it twice, once with anchor HTML code added directly and once with substititution, but I might have done something wrong; perhaps someone could check if that's repeatable. NebY (talk) 19:51, 22 October 2025 (UTC)
- Another reply to User:Gawaon: "Because other things exist" is not a compelling argument. You focusing on the example being extreme is also not compelling; even a single substed anchor is one too many. CapnZapp (talk) 09:55, 23 October 2025 (UTC)
- User:FaviFake: If substed anchors only appeared in project space you might have a point, and I would indeed be willing to let the issue slide, since editing guidelines or help pages (etc) isn't critical to the new user experience. But it isn't - I'm here because our current guideline tells people to subst anchors everywhere (you would know, because you made it say that). This definitely includes article space, where new editors will trip up on it. CapnZapp (talk) 09:56, 23 October 2025 (UTC)
- That remark is hardly fair, since the guidelines already said the same thing before FaviFake's edits (which were only about the exact placement of the subst'ed anchors). Gawaon (talk) 10:17, 23 October 2025 (UTC)
- @CapnZapp I don't what you're replying to. I said
[a] large number of anchors is only common on projectspace
and you aren't talking about large numbers of anchors being common outside projectspace. Of course anchors in general are used in the main space, that's why we have guidelines about using anchors.And, as Gawaon said, the assertion that Imade [the current guideline] [...] [tell] people to subst to subst anchors everywhere
is simply untrue. (emphasis in original) FaviFake (talk) 14:18, 23 October 2025 (UTC)
- CapnZapp: "even a single substed anchor is one too many" is your opinion, but an opinion isn't an argument either. I'd agree that it would in principle be better to avoid subst'ed anchors if there was a good way of doing so, but considering the downsides of the possible alternatives, it seems to be the least bad alternative. Not ideal, but sometimes that's all one will get. Gawaon (talk) 10:20, 23 October 2025 (UTC)
- I wasn't trying to pass off my opinion as argument. I was pointing out that your reply to me repeating Srleffer's example focused solely on how it substed more than one template, not the basic fact substing leaves programming code in the edit window of regular editors. It doesn't need to be said, but you are obviously entitled to have and share your opinion. However, can I ask what you mean by "downsides"? That is, in plural? To the best of my knowledge, the sole argument so far for writing the guideline to exclude any alternative to substing is that section links (from watchlists and page histories) aren't broken. CapnZapp (talk) 10:47, 23 October 2025 (UTC)
can I ask what you mean by "downsides"? That is, in plural?
There are multiple, I've ce'd the section in WP:ANCHORSUBST to distinguish them:
— Template:Anchor/doc § Not transcluded FaviFake (talk) 14:32, 23 October 2025 (UTC)- [This is visual clutter. It annoys people who are browsing the edit history] The template call will appear in the edit summary each time the section is edited, like this:
/* {{anchor|Transcluded anchor}} Basic format */. - [This is what you're saying] In watchlists and history pages, the clickable link to the section will fail to bring the user to the correct section, unless the template call is manually removed before the edit is published.
- [This also affects functionality] After editors save an edit to a section, they are taken to the top of the page instead of to the section they just edited.
- [This is visual clutter. It annoys people who are browsing the edit history] The template call will appear in the edit summary each time the section is edited, like this:
- I wasn't trying to pass off my opinion as argument. I was pointing out that your reply to me repeating Srleffer's example focused solely on how it substed more than one template, not the basic fact substing leaves programming code in the edit window of regular editors. It doesn't need to be said, but you are obviously entitled to have and share your opinion. However, can I ask what you mean by "downsides"? That is, in plural? To the best of my knowledge, the sole argument so far for writing the guideline to exclude any alternative to substing is that section links (from watchlists and page histories) aren't broken. CapnZapp (talk) 10:47, 23 October 2025 (UTC)
- User:FaviFake: If substed anchors only appeared in project space you might have a point, and I would indeed be willing to let the issue slide, since editing guidelines or help pages (etc) isn't critical to the new user experience. But it isn't - I'm here because our current guideline tells people to subst anchors everywhere (you would know, because you made it say that). This definitely includes article space, where new editors will trip up on it. CapnZapp (talk) 09:56, 23 October 2025 (UTC)
- Having 2+ anchors in the same section doesn't seem that rare to me though, and results in ugly, verbose, and unclear wikitext. Example. {{Anchor}} is very clear.
- This is indeed unpleasant to read, but having five anchors for the same section will in any case be very rare. Also I imagine that newbies using the source-code editor will fairly quickly learn to ignore the HTML stuff and focus on wiki-formatted content instead. There are other situations, such as tables with CSS styling, where the situation is the same. Gawaon (talk) 17:24, 22 October 2025 (UTC)
- I genuinely ask everybody to consider the choice we're about to make here not as a coder or experienced wikipedian, but as way to warmly greet new arrivals to Wikipedia. Best, CapnZapp (talk) 15:31, 22 October 2025 (UTC)
Location of anchors placed in or near headings
Please edit this post freely and discuss below.
============
- Before the heading
- Pros: Links to a good place. Mostly sensible wikitext. Cons: Doesn't open the collapsed section on mobile. Not visible in section editing.
- Inside the heading, before the text
- Pro: Most functional for screen readers and mobile. Con: Ugly wikitext.
- Inside the heading, after the text
- Pro: Somewhat less ugly wikitext. Con: Screen readers don't read the heading text.
- After the heading
- Pro: Sensible wikitext. Cons: Link puts the heading off the top of the screen. Screen readers don't read the heading.
============ - CapnZapp (talk) 14:43, 22 October 2025 (UTC)
- Why you do repeat what was already said by Srleffler above? That just pointlessly forks the discussion. Gawaon (talk) 15:16, 22 October 2025 (UTC)
- The forking of the discussion was made by Butwhatdoiknow when he started the subsection to discuss "subst or transclude?" specifically. To me, having also a separate place to discuss the other aspect (excluded from that section) makes things more clear, not less. Regards, CapnZapp (talk) 15:22, 22 October 2025 (UTC)
- I do not understand the point of yet another subsection? FaviFake (talk) 15:51, 22 October 2025 (UTC)
- There are two separate issues: (1) Where to place the anchor. (2) Whether, when an anchor is placed within a heading, it should be transcluded or substituted. Hence, there are two separate subsections. - Butwhatdoiknow (talk) 05:26, 23 October 2025 (UTC)
- As it would make no sense to substitute if the anchor is placed outside the header, I don't think there's a point in discussing those separately. Discussing whether or not to subst only makes sense if one already has decided that placement within the header is best. But realistically that decision won't be made without taking the question of substitution vs. transclusion into account. Gawaon (talk) 07:19, 23 October 2025 (UTC)
- There are two separate issues: (1) Where to place the anchor. (2) Whether, when an anchor is placed within a heading, it should be transcluded or substituted. Hence, there are two separate subsections. - Butwhatdoiknow (talk) 05:26, 23 October 2025 (UTC)
Wikipedia:Village pump (proposals)#RfC: Should links to the redirect Search Engine Land be restored?
I created an RfC at Wikipedia:Village pump (proposals)#RfC: Should links to the redirect Search Engine Land be restored? where this guideline is mentioned. Cunard (talk) 09:30, 20 October 2025 (UTC)
Linking to new names
There is a program that was originally written in conjunction with CP-67 and was called Cambridge Monitor System (CMS). In VM/370, the System/370 replacement for CP-67, IBM renamed CMS to Conversational Monitor System, retaining the initialism. There is a redirect from the old name to the new name.
The article CMS EXEC links to the new name in the first sentence and mentions both the old name and the new name in the second sentence. Should the reference to the old name be wikilinked even though it points to the same page as the new name? -- Shmuel (Seymour J.) Metz Username:Chatul (talk) 13:11, 22 October 2025 (UTC)
- If one of the two links doesn't redirect to a section, then only one link should be kept. FaviFake (talk) 19:07, 22 October 2025 (UTC)
- I disagree. The purpose of links is to explain unfamiliar concepts to the reader. Both Conversational Monitor System and Cambridge Monitor System are terms that may be unfamiliar to the reader, and it is not obvious that they refer to the same thing. It's not even obvious that the two terms will always link to the same article, although they do at present. Both terms should be linked in this case, unless the article text makes it clear that they are two terms for the same thing.--Srleffler (talk) 23:39, 26 October 2025 (UTC)
- I'd rather suggest to tweak the text in such a way that it's obvious that both names refer to the same thing, removing any need for a second link. Gawaon (talk) 08:36, 27 October 2025 (UTC)
- I agree insomuch that if you can reword the text to remove the need for two links (that lead to the same place), that is indeed preferable. However, if you can't or won't do that, linking both terms is much better than linking only once (relying on the reader understanding that both terms are explained wherever the single link leads to). That is, don't remove links just because our guidelines might say so - making thinks easy to understand for readers is and should be our top priority, even if at the expense of a couple of technically superfluous links. CapnZapp (talk) 10:32, 27 October 2025 (UTC)
- Yup, this seems th ebetter solution. Two links in two sentences usually means two different things. FaviFake (talk) 16:05, 27 October 2025 (UTC)
- I'd rather suggest to tweak the text in such a way that it's obvious that both names refer to the same thing, removing any need for a second link. Gawaon (talk) 08:36, 27 October 2025 (UTC)
- I disagree. The purpose of links is to explain unfamiliar concepts to the reader. Both Conversational Monitor System and Cambridge Monitor System are terms that may be unfamiliar to the reader, and it is not obvious that they refer to the same thing. It's not even obvious that the two terms will always link to the same article, although they do at present. Both terms should be linked in this case, unless the article text makes it clear that they are two terms for the same thing.--Srleffler (talk) 23:39, 26 October 2025 (UTC)

