Template talk:Interlanguage link/Archive 4

No Dutch (Nederlands) version?
Although this template has versions in many (more than 80) languages, there does not appear to be a Dutch version (which would show up in the 'Languages' list as 'Nederlands'). Is there a special reason for this, or is it just an oversight? Hops Splurt (talk) 19:07, 30 November 2021 (UTC)
 * If there is a special reason for this, then what is the reason?
 * If it is an oversight, then what needs to be done to create a Dutch template? Suggested name:  or.
 * If it doesn't exist, it's because no one has created it yet. I've never translated a template so I couldn't rightly say. Primefac (talk) 19:21, 30 November 2021 (UTC)
 * If I understand correctly, you are asking about the absence of a particular template on the Netherlands WP site. That question is more appropriately asked at the Netherlands WP site, as the people running the English-language site have no say over that.  That said, you should keep in mind you can still use an interwiki link, e.g. NL:Hoofdpagina.
 * Having said that, there's some degree of controversy (at least on English WP) about what the best way is to provide interlanguage links. Part of this is about whether to trust that other WP sites uphold the same high standards, whether people can figure out how to get an article in another language translated to a language they understand, and whether Wikidata is a good or bad way to identify a list of other-language versions of a given WP page.  Fabrickator (talk) 19:38, 30 November 2021 (UTC)
 * Each Wikipedia language makes their own rules. Some languages don't allow interlanguage links in the body of articles. See e.g. Help desk/Archives/2018 July 10 for German. See nl:Help:Gebruik van links for Dutch. I don't know Dutch and a machine translation is unclear to me but I guess Hops Splurt can read it. PrimeHunter (talk) 20:00, 30 November 2021 (UTC)
 * {ping|PrimeHunter}} Ik versta ook geen Nederlands. Hoe goed vertaalt dit zich voor jou? Fabrickator (talk) 21:22, 30 November 2021 (UTC)
 * [fix ping]. Andy Mabbett ( Pigsonthewing ); Talk to Andy; Andy's edits 21:57, 30 November 2021 (UTC)
 * Google Translate on your post gives: 'I don't understand Dutch either. How well does this translate for you?' That sounds clear. From nl:Help:Gebruik van links I get: 'However, it is not the intention to do this in the "ongoing" article text! It is allowed, for example, with citations at the bottom of the page, and of course on talk pages'. Here 'not the intention' sounds unclear to me. I guess it indicates disallowed. PrimeHunter (talk) 22:13, 30 November 2021 (UTC)
 * For better or worse, this issue has been asked and answered at nl:Wikipedia:De kroeg/Archief/20200709. Of course, it can always be reopened. Fabrickator (talk) 22:45, 30 November 2021 (UTC)
 * Interesting. There was an edit war in that Dutch "Help" page, that started here, when a fact-tag was placed by a now-vanished editor. They pressed the point a bit here, and it's noteworthy to point out that the argument against the linking that the now-vanished user advocated dated from a poll done nine years ago. What's funny about the 2020 discussion, linked above (and participated in that), is that it seems that part of the argument is "well articles in other Wikipedias may not be good and thus escape editorial oversight"--which is funny given the average quality of Dutch articles, and given the many wholly unreferenced articles written by one of the editors who proposed that argument. Drmies (talk) 16:59, 2 December 2021 (UTC)
 * As I stated in the mentioned vote: In my opinion, the interwikilink in the body text is a sign of the author's laziness. If the author uses the article in another language, he may have taken note of the underlying sources and link to them directly. If the article does not exist, it would be better for him to write it (if necessary as a start) instead of assuming that the reader does master the language of the foreign article. That will still work with German and English, with French, Spanish and Italian a lot less and with the rest of the languages you simply cannot expect the reader to have enough knowledge of the language to understand the article.  I see a lot of interwikilinks go to Russian or Chinese articles. How many wiki-editors master those languages? The Banner  talk 18:13, 2 December 2021 (UTC)
 * I understand that we are talking about links to Wikipedia articles, which normally are merely links to topics mentioned in the article. Specficially, they are not citations.  Sources mentioned in these links are not really the concern of the editor, so I don't understand.  Also, there's no obligation or expectation that anyone is mastering the language of the article  (and btw, you shouldn't assume the machine translations of Russia or Chinese are incomprehensible).  If someone is interested in finding more about the linked topic, the editor has provided a link to use, and assuming the reader isn't especially fluent in the language or isn't fluent in the language at all, just tell it to translate.  Perhaps the translation is imperfect, though most times the translation seem to be really quite good, and this is not just for languages written in the Latin alphabet. Fabrickator (talk) 22:25, 2 December 2021 (UTC)

I've added the template to Dutch Wikipedia. The /doc page needs expansion, so please help with that. Also, there are no sandbox or test case pages yet, so it's not fully tested, but all the basic functionality is tested through ExpandTemplates and working, as shown by the live examples on the transcluded /doc page. None of the WD/Reasonator code is tested. The template required creation of new Dutch template Join, which also requires /doc, and maybe /sandbox and /testcases. Let's move further discussion of this template to the Talk page there. Thanks, Mathglot (talk) 01:37, 13 February 2022 (UTC)

Well, that seems to have been a giant waste of time. After creating the template and adding a few ills to some articles related to the French Revolution (see nl:Gebruiker:Mathglot (bijdragen) (overleg) ) they were all deleted within seconds. I guess I won't be spending any time "helping" Dutch Wikipedia with any missing templates in the future. Mathglot (talk) 08:16, 13 February 2022 (UTC)
 * Maybe you missed PrimeHunter's comment above? Each Wikipedia language makes their own rules. Some languages don't allow interlanguage links in the body of articles. See nl:Help:Gebruik van links for Dutch. I wouldn't interpret Dutch Wikipedia as being ungrateful or not wanting your help just because they revert you when you add stuff they don't allow. Regards CapnZapp (talk) 09:18, 13 February 2022 (UTC)
 * Saw it but didn't catch the import of it until I read the original plus a 2020 and a 2012 discussion on it over there. In the older discussion, they voted overwhelmingly to reject it, and their version of an Rfc is interesting; they just give their opinion under a "Support" or "Oppose" header (sort of like at an Rfa), or even just add their sig with no opinion and no appeal to policy or guidelines for support, and then that's the decision. To each their own, I guess. The newer discussion was very lightly attended, and reached no consensus that I could see, certainly not enough to overturn the 2012 one. The template may still have some very limited utility as it may still be used in bottom material. Mathglot (talk) 09:56, 13 February 2022 (UTC)
 * I wouldn't characterize the Dutch Wikipedia as "ungrateful"--I would characterize them as set in their ways and focusing on the wrong thing. Drmies (talk) 15:40, 13 February 2022 (UTC)
 * In general, NLWP users also don't really like to be 'surprised' by new methods and/or templates. A careful introduction is therefore recommended (in which I am willing to support by the way). At least the template now clarifies how it works, there was quite some confusion about that in the mentioned 2020 discussion above. Encycloon (talk) 16:13, 13 February 2022 (UTC)

Problem with Interlanguage link template
If the English article name is the same as an already existing one (in this case a disambiguation), the template doesn't work. Ex: I'm writing an article about an series, one of the actresses is Zhang Nan, there's other people on Wikipedia with that name so there's a disambiguation page (https://en.wikipedia.org/wiki/Zhang_Nan), I can't write Zhang Nan because then the template wouldn't work. Megutim (talk) 12:16, 20 February 2022 (UTC)
 * Simple: use the name as it would have to be for an article for that person: Zhang Nan (actress, born 1997). -- Michael Bednarek (talk) 13:15, 20 February 2022 (UTC)
 * You can write Zhang Nan (actress, born 1997) to produce Zhang Nan (actress, born 1997). —Kusma (talk) 14:13, 20 February 2022 (UTC)
 * @Megutim, have you seen the answer? —Kusma (talk) 18:06, 20 February 2022 (UTC)
 * Yes, I've seen it. But that's exactly what I didn't want to do. I guess that's my only option. Megutim (talk) 18:55, 20 February 2022 (UTC)
 * There's always the option to just write an article about her in English :) —Kusma (talk) 19:55, 20 February 2022 (UTC)
 * Sorry I didn't see your answer, I was talking about Michael Bednarek's answer. Your suggestion solves my problem. Megutim (talk) 10:36, 21 February 2022 (UTC)
 * Megutim: Not sure what you expected to be able to do? What is so bad about linking to an actually realistic target page - the one you have been given as a suggestion (Zhang Nan (actress, born 1997)) does follow our article naming conventions. You just added the 1997-born actress to the dab. The problem is, disambiguation pages are only meant for entries with links that send the reader onward, so I reverted that addition. CapnZapp (talk) 09:23, 21 February 2022 (UTC)
 * Oh okay, I didn't know that, I'm sorry. Megutim (talk) 10:35, 21 February 2022 (UTC)

Some degree of controversy
The following is taken from the previous section (named No Dutch (Nederlands) version?). I started a new section to not go off-topic there:

Having said that, there's some degree of controversy (at least on English WP) about what the best way is to provide interlanguage links. Part of this is about whether to trust that other WP sites uphold the same high standards, whether people can figure out how to get an article in another language translated to a language they understand, and whether Wikidata is a good or bad way to identify a list of other-language versions of a given WP page. Fabrickator (talk) 19:38, 30 November 2021 (UTC)


 * My thoughts:
 * trust that other WP sites: I don't think it is constructive to only offer the links if we first guarantee the validity of the other-language article. The usefulness of "actually, there is info, just written in another language" easily trumps the concern for that other site's quality. At most, make it more clear you are leaving English Wikipedia.
 * whether people can figure out: To me this is a quintessentially American reaction that should be ignored. Most visitors actually do know a second language, and a great portion have a different first language than English. Just trust the reader. Not exposing a reader who has only a single language (English) to "foreign script" he or she is unequipped to handle just shouldn't be our main concern.
 * whether Wikidata is a good or bad way I find Wikidata clunky and overused. A template like ill does the job perfectly fine (as long as we could get bots to stop wiping them out as soon as local articles are created, i.e. before they are established and no longer in immediate danger of being deleted.

"The problem occurs when there is no English WP artricle... and this is what is for. But the major issue is that the proper set of links to provide can change over time, and we should want to avoid having to update all articles which include such references when that happens.  Wikidata provides a proper solution to that problem.  The main objection given is that Wikidata is clumsy in how these links are displayed.  Thihs seems like a fixable problem, but there's not much motivation to get this fixed when we discourage or prohibit the use of Wikidata in this context.  If Wikidata were to be widely used for this purpose, the motivation would be there." I hope you will give these points due consideration. Fabrickator (talk) 16:44, 2 December 2021 (UTC)
 * Cheers CapnZapp (talk) 21:24, 30 November 2021 (UTC)
 * See nl:Wikipedia:De kroeg/Archief/20200709 for the perspective of a Netherlands editor advocating for Wikidata (same one as in the "No Dutch (Nederlands) version?" section). Fabrickator (talk) 22:45, 30 November 2021 (UTC)
 * If you want to advocate for using WikiData and/or explain how to use it smoothly, here on English Wikipedia, can I ask you to do it in English please? Otherwise your edit risks being ignored. Thank you CapnZapp (talk) 14:00, 1 December 2021 (UTC)
 * I'm not sure I understand your situation. As soon as I go to the page I linked, my browser just automatically asks me if I want to see it in English, and I think that's the way it works for the majority of people.  Anyway, if someone only understands English and doesn't have a browser which automatically translates non-English content, then it's understandable that they would be less likely to see the merits of linking to pages on non-English sites, so they aren't the real target for such links.  OTOH, if Wikipedia had more links to non-English pages, that would provide the motivation to arrange for their browser to do the automatic translation. Fabrickator (talk) 01:40, 2 December 2021 (UTC)
 * Oh, I don't have a technical problem - I am suggesting you will get better results when discussing on English Wikipedia if you communicate in English. I hope we agree anything you feel is important enough to be said is worth your while to translate into English yourself, and its corollary: anything you can't be bothered to express using the English language probably isn't worthwhile considering. This is why I would advise you to assume arguments written in other languages will be ignored.
 * tl;dr: Don't assume people will bother with links to discussions written in other languages - why not take the time to summarize the pertinent points here (in English)!? :-) Also see WP:ENGLISHPLEASE. Cheers CapnZapp (talk) 10:15, 2 December 2021 (UTC)
 * To clarify: I am not telling you what to do. I am merely attempting to explain why you might not get the responses you want. Have a nice day CapnZapp (talk) 10:18, 2 December 2021 (UTC)
 * Aside from my "brief" example Dutch text, my comments have been in English. I merely linked to Dutch-language content.  I assume good faith that people will be objective, and will give due consideration to comments made in a non-Englsih forum, assuming they can easily read those comments.  However, since you have expressed opposition to use of Wikidata in conjunction with, I will summarize my own position vis a vis the use of Wikidata in conjunction with links to non-English Wikipedia pages: "In an ideal world, links would be provided to a given WP article in all the languages in which that article exists. We already do this in those instances in which an English-language article exists, by virtue of including those links appearing (in the left margin) of the English WP page."
 * I assume good faith that people will be objective, and will give due consideration to comments made in a non-Englsih forum, assuming they can easily read those comments. Well, no - please don't construe ignoring non-English content as not being "objective" or "good faith". WP:ENGLISHPLEASE clearly states: "This is the English-language Wikipedia, so discussions should normally be conducted in English. If using another language is unavoidable, try to provide a translation, or get help at Wikipedia:Embassy." To be frank, this means your fellow editors are well within their rights to ignore arguments made by people that can't even be bothered to express themselves in English, without that being considered a slight on their part. Now then, you are of course free to ignore my advice if you want; just don't be surprised if you don't gain as much traction for any foreign-language arguments as you might hope. Thank you for summarizing in English. However, at this time I choose to limit myself to providing language advice, and not engaging in the WikiData discussion other than what wrote at the top of this section. Regards, CapnZapp (talk) 08:20, 3 December 2021 (UTC)
 * I infer that "good faith" implies that procedural objections are not raised unless they are in fact substantive. If there's some sort of non-trivial inconvenience resulting from not having some comments translated to English, that might be valid.  I presume that the WP:ENGLISHPLEASE guidance pre-dates the point at which machine-based language translation was integrated into the popular web browsers.  Fabrickator (talk) 09:02, 5 December 2021 (UTC)
 * Remember, I'm not here to police your behavior, Fabrickator. My entire aim here is to help you, a fellow editor, to understand why you might not get the results you want. If you wish to change current policy, I welcome your initiative. Until then, however, acting as if current policy is obsolete is simply unconstructive. I am merely advising you to discuss in English or risk having your arguments ignored, simple as that. (I should add that this thread is starting to feel exhausted. I really have nothing else to say) Best regards, CapnZapp (talk) 09:16, 5 December 2021 (UTC)
 * To User:Fabrickator and any other interested party: I made an edit to WP:ENGLISHPLEASE to clarify this stance to avoid future unclarity. I made this edit about two weeks ago, and the guideline page has been stable since. So I now feel reasonably confident it represents our community's consensus. The discussion is here: Wikipedia talk:Talk page guidelines. For future reference, a direct diff link to the addition: . Thank you CapnZapp (talk) 17:46, 28 February 2022 (UTC)

Back on topic

 * , thanks for raising this. To your original point, about whether to trust that other WP sites uphold the same high standards..., I don't think we have to worry about that question. First of all, by definition Wikipedia is not a reliable source, neither en-wiki, nor nl-wiki or any other. Put another way, you could consider your question in a much larger scope than that of the tiny universe of interlanguage links, namely, with respect to all wikilinks in the encyclopedia: who is to say whether we should trust the editors of any other article, so maybe we should have no wikilinks at all because other articles can't be trusted. The truth is, we don't know for any given article how well the guidelines have been applied. (Well, in theory, for articles with good or featured status I suppose we do have a clue, but for most articles we don't.) So even if we don't know whether the article we might link to at a foreign Wikipedia is "worthy" in some sense, that's also true with internal wikilinks to other English articles. And then, there's the time factor: maybe an article was high quality at the time it was linked, and then degraded to a much poorer quality over time; and once again, that's just as true of internal links to other articles at en-wiki as it is for interwiki links. So, the problem of trust exists, certainly, but it seems to me it's inherent in being part of the "encyclopedia that anyone can edit it", and there's not much we can do about it. To a certain extent the Tl;dr here is caveat emptor, which is also true when reading any given article at Wikipedia (in any language). Thanks, Mathglot (talk) 10:10, 13 February 2022 (UTC)
 * Can I just point out a possible misalignment here. WP:WINARS refers to using Wikipedia as a source for references in articles. It appears to me (but I could be mistaken here) you are instead talking about when our articles send readers elsewhere for more information User:Mathglot, or perhaps more generally are discussing the topic of "is Wikipedia reliable for use by readers?" Regards, CapnZapp (talk) 09:03, 14 February 2022 (UTC)
 * Yes, I think you've identified a bit of fuzziness in my logic, but I'm gonna blame it on the wine and the late hour, and think about whether I need to adjust it tomorrow... Mathglot (talk) 09:13, 14 February 2022 (UTC)

Proposal: reduce the number of foreign links
I propose that we limit the number of foreign language links to four. I can't personally see a reason to have more than that; certainly, it's absurd to have 25 positional parameters, even with the abstraction of twelve pairs of language links and pagenames. If nothing else, any attempt to add twelve links to an article should be resisted as a clear case of cite clutter. I wouldn't favor eliminating all the links in favor of just wd, as at a minimum that would involve an extra click to get where I want, and also eliminates information that is useful to me in deciding whether to click.

I was going to mention something about this in the edit request above, but it's really a separate topic, and deserves its own airing of views. I would just add that deciding exactly which links to add if many languages are available to choose from is a legitimate discussion to have as well, and is related to this one, but is not the same thing and imho is more about advice to the user on a /doc page rather than the usability and technical aspect of template usage that limiting the number of links would be.

It kind of feels like there are 24 positional params there for 12 languages "just because we can", that is, becauase it is easy to do coding-wise. By way of explanation for any non-template coders weighing in, imho it would be trivially easy to increase the number of languages in this template from twelve to twenty, or fifty, and easier to test, than for example, testing the change for the single new parameter lt-lang proposed in the section above (due to interactions between params, multiple test cases with LTR and RTL languages, and possible subst-ing issues). Of course, it would be nuts to increase the number of languages beyond twelve, but what we have now is nuts, too, and should be fixed. Mathglot (talk) 18:49, 13 February 2022 (UTC)


 * Oppose changing this in the template. I agree that more than three links are usually too much, and should be discouraged in mainspace. But that's not a good reason to cripple the template for use in red link lists in user or project space. —Kusma (talk) 18:53, 13 February 2022 (UTC)


 * This sounds like the exact sort of thing maintenance categories were created for. Oppose restricting template performance to enforce any such rule. Even if we had a strict policy about how many you could use, you would still want the current performance so that you could follow the links on preview/preliminary edit in order to choose the ones you really wanted to link to. Would definitely support a maintenance category for mainspace uses with five or more language links though. VanIsaac, MPLLcont WpWS 20:26, 13 February 2022 (UTC)
 * , that's a very good point. Coupled with 's objection bsed on "red link lists" (which I was not aware of) I think it's best to withdraw this proposal. However, the discussion has taken some interesting paths, and I don't at all want to quash discussion of the concomitant issues, so please do carry on, as long as it's fruitful. Mathglot (talk) 08:33, 14 February 2022 (UTC)


 * Whether this limit is coded in a template or established as a policy, how do you know which languages satisfactorily meets the needs of the reader? It's quite presumptuous for us to decide that we should give preference as to which languages are made more accessible to others based on which languages look more familiar to us.  As for the objection that a particular solution might require one or more additional clicks, that makes my head explode!  Frankly, though, I cannot get my head around the preference for any solution that isn't based on an external source (a la wikidata).  We can always design this so it limits the selection displayed based on a user-specified set of languages along with a default set that meets the needs of the bulk of the users.  We have the infrastructure to do this right, it's almost criminal not to do so. Fabrickator (talk) 23:51, 13 February 2022 (UTC)
 * Displaying a dozen iwlinks disrupts the reading flow for everyone, so we should not do that. Generally, it is helpful to link to the most well developed page(s) and to the language(s) with the closest connection to the topic (most of the time these criteria give the same answer). With better technology, other things could be possible, but we don't have a much better technology available right now. —Kusma (talk) 00:22, 14 February 2022 (UTC)
 * I didn't specify what the interface should look like, I just specified that it should be data-driven. As an editor adding the, I don't need to be distracted trying to evaluate the various versions of the article (which of course are subject to change).  Whether it displays the top n selections based on the user's preferences or a popup with those selections when I mouseover, that isn't what I'm arguing about.  Rather, improve scalability by minimizing the risk that some change on some other languge WP necessitates a change to en-wiki, while still leveraging the content of other language WPs. Fabrickator (talk) 01:19, 14 February 2022 (UTC)
 * we are on the same wavelength, or similar. I didn't want to burden the OP with further text, but in fact I was thinking of some of the possibilities you mentioned, either by aligning links with my user preferences, and also thinking about whether I could just add my userid as an entity at Wikidata (would that even be allowable?) and assign an array of language properties to it, possibly even with the Babel 1–5/N level as a property indicating my familiarity with it, and then match those up with the available links, and render the top three language links on pages that I view. (I'm imagining parser conditionals assigning classes with  for the languages outside my competence or too far down the list). Mathglot (talk) 08:43, 14 February 2022 (UTC)
 * This probably isn't possible from a performance point of view (too many different language combinations to cache). But having CSS classes and suppressing display of some language links via user CSS might be viable. (You'd have to ask at WP:VPT for options and ideas). —Kusma (talk) 08:50, 14 February 2022 (UTC)
 * I think I've got the CSS part under control, unless there's something I'm missing, as I've already done proof of concept in other situations, such as conditionally clearing parts of Template:Talk header based on classes in common.css (link). In theory, I could just display:none all the lang codes (read, "the top 20 or 30 most common codes") for languages I don't know, and all we would need are classes (or ids) on the language links. The downside here is it requires opt-in via common.css, and how many people are going to bother with that? (Okay, okay, everybody in this thread will, but I mean, who else? ) Mathglot (talk) 09:16, 14 February 2022 (UTC)
 * Oppose, but IMO the purpose of ill is not to link to all available articles in other languages, but to establish a connection to a resource of information for which we should probably have an article as well (in the judgement of the editor adding the link), or at least a footnote giving the necessary background info.
 * If there are multiple foreign-language articles we could link to, we should link only to those which have substantial content that we should add to our WP as well (regardless of the language-capabilities of us or our readers). So, if the Chinese article is substantial, and the German one is just a stub, the link should go to the Chinese article, not the German one (even if more people would be capable reading German than Chinese over here). Sometimes, there's different relevant content distributed over multiple foreign WPs, then we could link to several of them to cover the superset of relevant info. If several of them have the same content, then we could, of course, reduce the links to those providing it in languages potentially more familiar to our audience. So, the first priority should be the actual content, followed by the quality of the presentation, followed by the familiarity of the language. The project to create an encyclopedia covering the knowledge of the world past and present is international, language is just a vehicle to transport the message.
 * However, besides its purpose to provide information, I think, the absolute top priority and main purpose of ill is its use for identification purposes and to establish the initial connection to an already established piece of infrastructure elsewhere thereby helping to eliminate ambiguity (in i.e. names) and redundancy (caused by "parallel" partial networks of articles about the same or very similar topics without interconnections). In this case, the actual information provided in the other Wikipedia is mostly irrelevant, what counts is the link itself, the relation to that foreign node, because it is the starting point for anything else. The ideal link target for this later purpose would be Wikidata, but since Wikidata has a rather crude user interface, if at least one article exists in any of the other Wikipedias (regardless of language) I would rather link to that article (no matter how crude it is) than to Wikidata, knowing that this article will have a connection to Wikidata already (which is trivial to detect and resolve for bots once an article has been created in our WP as well).
 * I don't think we would normally need more than at most a handful of foreign language links (I never used more than one, and would not have used more than 3 if I had even known that it is possible to link to more than one ;-), but I also don't think that we should force our opinion on those who actually work on an article, and if someone feels like s/he has to link to 20 external sites, that should be their business, not our's, and in those rare cases where this might happen, if someone objects, they can discuss it on the relevant article talk page.
 * So, to sum it up, I'm against enforcing any artificial limits. Whatever we choose, 4, 5, 10, 20, 25, it will always be arbitrary. If I had coded this template, I would probably have stopped at 10, but since the template supports more already, someone felt like it should support more, and I respect that.
 * On a different note, in the list we should replace the "; " (semicolon followed by space) by ";&hairsp;" (semicolon followed by hairspace), this would considerably reduce the occupied space already.
 * --Matthiaspaul (talk) 17:44, 14 February 2022 (UTC)

Identification
, Regarding the purpose of the ill, I think you have a somewhat different view of it than I do, and I wonder where others stand on this. You talk about the purpose being "to establish a connection to a resource of information for which we should probably have an article as well" (which sounds more like R with possibilities to me), and your third paragraph was somewhat opaque to me, seeming to imply that the link itself was somehow important rather than the information found once you get there; was that your intention?
 * Where we differ, I think, is who or what the focus of the links is; for me, it is the readers: I see the purpose of the links as delivering information to the reader *now* (well, one click away). Other purposes are secondary; there can certainly be other benefits, such as the "red link" function of pointing out missing content and providing an easy, one-click method for *editors* to start filling in the lacuna. The ill red link performs the same function as the local red link, but also gets you the information for *readers*, who far outnumber editors.
 * I get the impression that for you, the linkage that would be used by editors to expand en-wiki is primary. To the extent that I follow the second half of your comment (not very well) I would say that that kind of linkage is best provided by and more central to Wikidata-linked redirects, and not by ill. Since I discovered how to do these (it's pretty well hidden, and looking again just now I couldn't find it), I have been adding some: for example, Children of Izieu, Personenverbandsstaat, Ministério Público Federal. Imho this is the better way of establishing the kind of linkage that I think you might be referring to, and serves as a readily available launchpad for creating articles based on gaps in en-wiki that are present on other Wikipedias. The tracking category Category:Redirects connected to a Wikidata item provides the launchpad.
 * If this is something you'd like to discuss further (and I think it'd be worth it), it seems like it's o/t for this edit request and even from the scope of this template, and so should be taken up at a different venue; possibly at WT:WikiProject Intertranswiki or WT:TRANSLATION. Or perhaps I missed the meaning of your comment so badly, that all of this is irrelevant, in which case, my apologies. Oh, and "yes" to the hair space separator. Mathglot (talk) 01:08, 15 February 2022 (UTC)
 * I think the intent of Matthiaspaul's 3rd paragraph is indeed to establish "identification". I often found misspellings in red links that revealed themselves as such only after I found an interlanguage link. // The matter of "Redirects connected to a Wikidata item" is a bit different; for most names and terms no such redirect is an option (see the 1st example overleaf, Hanning Schröder), and more importantly, they don't provide a link to any further information for the interested reader of the redirected-to English article, because it's not there. -- Michael Bednarek (talk) 02:15, 15 February 2022 (UTC)
 * Huh? Of course the redirects with wikidata connections provide links to further information, that's half the reason they are there. In order, the three sample links above provide two, two, and six links to further information for the interested reader. Or am I misunderstanding something? Mathglot (talk) 09:12, 15 February 2022 (UTC)
 * I obviously didn't express myself clearly enough. Readers of the article Izieu or its section "Site of World War II Jewish orphanage" will not see the interlanguage links at Children of Izieu. -- Michael Bednarek (talk) 12:45, 15 February 2022 (UTC)
 * I see those interlanguage links as primarily serving the immediate information needs of readers, but the function of providing authority control of sorts (as described in that 3rd paragraph) is also relevant. This identifies the given entity among the other entities with the same name and it declares the preferred name for that entity here. – Uanfala (talk) 13:57, 17 February 2022 (UTC)
 * I obviously can't make you look at a template my way, but I want to emphatically state I don't see any of that in the Ill template. The only purpose of the template is to tell a reader "we don't have info on this stuff, but Uzbek or Norwegian Wiki does. Maybe you'll find it useful, maybe not." This "heads up" is given with zero quality promises - it is provided by an editor that probably does not speak Uzbek or Norwegian and likely haven't even LOOKED at the destination page, let alone vetted it. Please don't read more into the template than this. More importantly: don't assume editors use the template in any other way than this. CapnZapp (talk) 16:26, 17 February 2022 (UTC)
 * , that's a fair point, about providing some authority based on a "preferred" name (i.e, the content of the red link). I like the idea, but I don't think I agree with it under the current conception of ill, but I'd like to explore how we can improve that situation. The reason is, that the red link is really only the preferred name of the person who placed it, and we don't know what effort they took coming up with it. In the case of proper names, this rarely matters. However in naturally descriptive titles, it matters a great deal, and depending how knowledgeable they are about the topic, how skilled they are with wording, how good they are at search (to make sure the article doesn't already exist under some other name), and whether they bothered to check ngrams, books, and other reliable sources in English to see what the most common usage in English actually is (not always what a straightforward translation would suggest), the red link may have more or less authority. When I place an ill, I try to do the research first.
 * At Draft:French criminal law for example, there are a great deal of ill templates to concepts in French law which are poorly covered at en-wiki; I probably spent more time carefully researching these language links and coming up with the proper terms for article titles in English, than I did on writing the content of the article. Where is the evidence of this, and how do we know if the red links in draft are authoritative? Nowhere, and we don't.
 * I like your idea of having an authoritative name for the red link representing the forthcoming article at en-wiki, and I'd love to pursue some method of instituting that. Perhaps Wikidata would be the logical place to start? One can't link the Wikidata item for an existing concept in other languages not present in en-wiki directly in the "Wikipedia links" section of a Wikidata concept, but maybe we could add a new property where we could collect the "authoritative" or "preferred" name in English (and other languages, of course).
 * Consider the following example from the Draft article linked above:
 * ⟶ Public prosecutor (France)
 * I spent a fair bit of time researching this one link: do we use the French term in English sources or not; is there an overwhelming choice, or several (there are several possible translations, "public prosecutor's office" is one of them); which one is the most common for the canonical title of the missing article in en-wiki; does the context of the sentence call for a different form than the canonical one, so that the first param (title of the forthcoming article) should be one thing, and we should have an lt param pointing to one of the other formulations; and so on. Basically, all that work is lost, as soon as I place the link.
 * Instead, could we enlist Q16009380 at Wikidata in some way, to register the value "Public prosecutor (France)" as "preferred" or "authoritative" English label, perhaps reserving the notes about what research was done for another field, or for the edit summary? (Note that the "label" field at wikidata seems more of a description, so not sure that's really where we want to place this value; for one thing, the kind of WP:PARENDIS that en-wiki uses for an article title doesn't seem to map well to the label field. Then again, I don't know what guidelines they have for "label", and to some extent, WD seems like the Wild West to me, with anybody doing anything they please, there.) If not wikidata, then what? Anyhow, I'd be interested in your feedback in how we could so something to codify your idea of making the red link (actually, the canonical name&mdash;param 1&mdash;which is usually the red link value, unless an lt is present) authoritative. Thanks, Mathglot (talk) 18:55, 28 February 2022 (UTC)
 * I highly doubt that we can mandate people use a particular disambiguated term simply based on what an ill or WikiData says it should be, especially if there's no way of telling someone that this redlink exists. For example, if there's an ill call on Page A to Joe Bloggs (businessman), and a regular old redlink to Joe Bloggs (entrepreneur) on Page B, someone creating an article based on B's redlink will have no idea that the proper or preferred dab is "businessman" (at least until someone sees the article and moves it).
 * I'm not saying it's a bad idea, I just don't see how it would be useful. Primefac (talk) 20:12, 28 February 2022 (UTC)
 * I'd be horrified at the idea of a mandate, certainly; this is more in the nature of facilitating the finding of an authoritative link if one is available. If someone has the done the research, found the most likely canonical name, how can we store and access that information, and make it available for re-use? Everything is always subject to consensus, so even something designated as "authoritative" (perhaps not the right word, based on your "mandate" objection) can always change. Mathglot (talk) 20:48, 28 February 2022 (UTC)

General Purpose
The general purpose of Interlanguage link is to link our readers to relevant information only available on foreign language Wikipedias. Readers uncomfortable with other languages than English are free to simply ignore these links. I guess that to be precise I should add that the specific purpose of Interlanguage link is to "enhance" red links: yes the link is red, but while you wait for somebody to start the article, you could find out more on the corresponding page on German and/or Chinese Wikipedia.

More than one link is uncommon. The reason you would link to more than one foreign-language Wikipedia is because each contains additional information useful to our readers that is unique to that project. I see no harm in our template allowing 5 or 10 or 15 such links. The only reason I can see to remove this functionality would be for technical reasons - that it makes the template hard to maintain or update. Had we started from scratch I could see us discussing how many links we need to support, but now that the functionality is already there let us leave it intact.

The proper way to send the signal "don't use many links unless each add value" is instead to revert individual additions of such "spammy" Ill links, and possibly improve the documentation. In the unlikely case two editors agree the German and the Chinese articles contain about the same information, but can't agree which one to link, again: what is the harm in satisfying both editors by including both links? A third editor can always later boldly remove one of them as redundant.

CapnZapp (talk) 10:27, 15 February 2022 (UTC)
 * I don’t know if i missed it in the tome above, but do we actually have hard numbers on how common this issue is? I somehow missed any realworld examples mentioned.  Is there a PetScan or something?  Then the next step I would think is to make a tracking category.  Then if it is a really big issue, we can put limits on the template.  If not, the tracking cat can just be monitored, especially if there are many exceptions and edge cases as I have gleaned from above.  --awkwafaba (📥) 01:38, 16 February 2022 (UTC)
 * Not above, but in the archives and at CFD. Primefac (talk) 08:17, 16 February 2022 (UTC)
 * The limit was increased from five to 12 in 2017, however these additional parameters are rarely used. The TemplateData report indicates that there are less than 100 uses where six or more languages are linked and just one with ten language links. EdwardUK (talk) 22:02, 16 February 2022 (UTC)
 * Then I (too) oppose a change of the template, but I encourage the nominator to visit those relatively few articles to, if possible, reach a consensus around reducing the number of languages linked by each such page. Regards, CapnZapp (talk) 06:54, 17 February 2022 (UTC)

New documentation
I have significantly trimmed/rewritten the documentation for this template, mainly to remove the verbosity and put more emphasis on the basics of "do this to get that". It is currently located at Template:Interlanguage link/doc/sandbox. Note that for simplicity I left out the header/footer templates and the TemplateData, since there wasn't much to change there. Feedback is appreciated so that there aren't any huge issues when it goes live. Primefac (talk) 14:21, 28 February 2022 (UTC)
 * While I appreciate the effort, I wonder why you based your work on an outdated version of the documentation thereby trashing a week's worth of enhancements. The documentation was already much better than this, more consistent (but still a long way to go before reaching "perfection"), and explaining a lot more stuff in details (but still in the "gathering missing information phase" rather then "nice prose phase" ). I do not consider your proposal to be suitable to go live, because it lacks a lot of important information.
 * What I do agree with is that the current documentation (that is, after restoring the stuff temporarily deleted today) needs some refactoring and that we should have a section with easy usage examples early on (after giving an intro on the templates use cases), so that users don't waste their time reading through the full feature set when they only want a quick answer for a trivial use case. However, further down the documentation should be comprehensive up into the minute details, have a detailed discussion of the various features, describe parameters and their uses in detail, and point out more advanced use cases and usage tricks. Some stuff could also be moved into footnotes, an effort I started as well. In general, the documentation needs to be considerably expanded (but in a good structured way), not trimmed.
 * --Matthiaspaul (talk) 22:24, 28 February 2022 (UTC)
 * I copied the sandbox directly from the main /doc... and then I removed the reams of prose. As I said, it's a first draft but the intention is to give more examples and less verbosity. Primefac (talk) 10:03, 1 March 2022 (UTC)
 * Generally in favor of this philosophy, although suggest it should be put on hold pending resolution of proposed template changes above.
 * There is, however, a way that further changes to the /doc via sandbox(es) could be very useful, and that is in order to define what it is we want the template to do. Normally, our templates do not approach the complexity (or multiple separate proposals) that would require the creation of a functional specification as is the case in the standard model of software development, however, this might be one of those cases. In fact, a well-written /doc page is an excellent proxy or stand-in for a functional specification, that is, showing in detail what the template should do, before consideration of how to do it. With respect to differing visions of how to proceed with the addition of new parameters or not, I propose that each person having an idea of the best way forward for the template prepare a sandbox (sandbox2, sandbox3, ...) which we can then compare as a set of alternative specs, and also invite other, non-template writer users to comment on, or even !vote for in an Rfc, if it comes to that.
 * I'll take up my own suggestion, and I'll come up with a /doc/sandboxN in the coming days, and I hope that other interested parties do, too. It looks like has already reserved /doc/sandbox, so if it's all right, I'll take /doc/sandbox2. Thanks, Mathglot (talk) 23:42, 1 March 2022 (UTC)  wording tweak; by Mathglot (talk) 07:04, 2 March 2022 (UTC)
 * Mathglot, is there a reason why we cannot collaborate on the sandbox page? I'm not sure how much "simpler" my version can get, but if you've got suggestions I'm not opposed to tweaking my version to incorporate the suggestions (assuming they're reasonable). No point in reinventing the wheel etc. Primefac (talk) 10:30, 2 March 2022 (UTC)
 * Sure we could, especially if we talked it through here, but I had an idea for a reorg and didn't want to step on anyone's toes, as it might depart from current practice. By talking through first, it could be done without examples, it's just that having something concrete to look at while discussing, can be helpful, in the manner of "a picture is worth a thousands words". Either is fine with me. Mathglot (talk) 00:18, 4 March 2022 (UTC)
 * (edit-conflict) I interpret the middle part as in line with something I had in mind as well, to have an "Intent" or "Purpose" chapter as the first chapter after the lede, perhaps combined with "Usage" - or change the "Usage" chapter into an "Easy & common usage examples" chapter. The latter suggestion to start multiple sandboxes, however, is a bad idea, wasting your own and other people's time. The template is not that important that we need three (or more) different sandboxed/forked versions to decide on, and in particular not now when all that is actually needed to be added soon is the documentation for the new features and use cases. Everything else is "nice to have" sugar on the cake, but it can also happen in a month, in a year, or whenever someone has time for it. (It is also okay to have it now, but not when it is destroying other efforts, as it did now. The way it is approached right now is sand in the gear.) After all, the documentation has basically been untouched for years and nobody bothered to improve it until I took the stance and started to document the recent enhancements and fill in areas where important information was lacking and to make it look at least a bit more consistent (although, as I said, there's still a long way to go) - and until these efforts were now basically all trashed because some editors suddenly decided that they must delete stuff rather than flesh out the missing bits. No good.
 * Three grown-up and good-willing editors should be able to coordinate their efforts while working non-destructively on the main document. So far, I only added to the document and started to slightly restructure it (moving some stuff into footnotes, using the same parameters throughout the document), so my work was completely non-destructive. (I was wondering about IMO completely unimportant contents like multiple interwiki-links or Reasonator stuff being discussed in significant detail in prominent places, which IMO would only belong into their dedicated chapters far down in the documentation, but I respected that some other editors obviously found this more important than me, so I did not delete it - I wished other editors would show the same respect to other people's contributions, because this is a collaborative effort). However, now that a lot of information was deleted, the documentation is worst than it was before I started to work on it.
 * Unless some editor tries to enforce his style and idea it should be possible to efficiently move forward with minimum communication overhead and still only very little back&forth. The approach must be to improve on the contributions of the other editors, instead of trashing them. It will have to be hundreds of little improvements over months or years, not one bold switch to something new and with less information than before.
 * My typical approach to improving an existing document is to first add missing information (where available) to the apparently most suitable places (at least at that moment) and, once more contents have accumulated in the document, think about combining the redundant pieces and restructuring the document so that users seeking for easy examples will find the answers to their questions early on in the document (or at least pointers where their answers can be found), and that terms/parameters will be introduced before they are used further down in order to also allow for linear reading top to bottom from easy to complex.
 * This is also our normal collaborative approach in article work, therefore I suggest it as a working model to move forward here as well (instead of the recent forking and deletion of contents). Moving information to better places is fine, deleting it is not (unless it would not be relevant in the documentation - but the discussion of the ranges of parameter arguments is relevant in a template documentation).
 * --Matthiaspaul (talk) 12:05, 2 March 2022 (UTC)
 * I think brevity and hitting the main points first that most users come here looking for is key, so that we don't scare off new users and frustrate occasional users just looking up how to do a simple, one- or two-language ill. The template is already quite rich in possibilities for experts, including wikidata and Reasonator options which I'll wager get used rarely but are nevertheless powerful, but experts are okay with going to the bottom of the documentation page (or even to page 27 of Appendix D of the documentation) to get the information they need, but the same cannot be said for the occasional user. I could even see splitting the doc into major sections, "Basic usage", "Additional options", "Expert usage" or some such, if that's what you were suggesting in your "easy to complex" comment, which I think is a good idea.
 * As far as this is concerned:
 * As for the second sentence, I agree, and had you made little improvements to the existing document, that would have been great. But instead, rather than doing that, your first three edits altered the TemplateData to correspond to the dozen or two new parameters that were added to the template recently, and since rolled back. The next four edits added inter-wiki prefix in a number of places in the doc. Later edits made some incremental improvements to prior existing doc, but by that point it was already on the back of doc changes that introduced the new parameters, and it would've been difficult to undo only the earlier changes while keeping the beneficial ones that came later.
 * As for the first sentence, if anything was trashed, it was the idea of consensual change, by doubling the number of parameters in an already-complex template with minimal discussion, accompanied by concomitant changes to the doc. Removing content lacking consensus following a bold change is not "trashing" it, it is the  R emoval part of the WP:BRD cycle, which I grant is perhaps intended more for article space than here, but I think the same principle generally holds.
 * The proposal above in section is still open. In my opinion, we should deal with that first, both in the template (where it was never separately implemented, so no rollback point to go back to), and then in the doc; that's assuming we get consensus for it, which I think is likely. After that, we could start a new section proposing whatever it is we want to propose regarding named parameters, and take it from there. It's also my opinion that we ought not to have the hubris that three template editors of whatever level of good will are necessarily sufficient to green-light a major change to a complex template such as adding a dozen or more parameters, and I'd prefer a larger base of consensus to feel comfortable with it. Mathglot (talk) 01:08, 4 March 2022 (UTC)
 * Let's take this stepwise - is there any opposition to me replacing sections 2 and 3 ("one language" and "multiple languages") with the contents of Template:Interlanguage_link/doc/sandbox? Are there any major improvements that could be made before doing that? My primary motivation for combining these is that it makes it too wordy and it doesn't really need to be in two separate sections. Primefac (talk) 11:10, 4 March 2022 (UTC)
 * I think it could be further improved on several fronts, but generally no objection, except:
 * The casa example need to be replaced, because that's a disambig page. (Or at least, moved down to a section that talks about less-frequent or edge cases.)
 * I'd like to see the embedded "(if different)" inside-the-example explanations removed and replaced with an explanation after the fact. I did this in this example.
 * To the second point: This could either follow up right after, or even be in an explanatory note, which could be useful in this template, perhaps. As an alternate, we could use a stylized tip format to explain it.
 * In addition, I wonder which is the more frequent case: page name the same, or not the same? If there's a clear winner, it should probably be placed first, and if it's not clear, probably the different-page-name case should be placed first. Also, for the doc, we should probably only use languages with Latin or Latin-extended alphabets for the canonical examples, and move non-Latin examples to the "Examples" section. The template /doc is complex enough, without having to deal with scripts you can't decipher.
 * I still think it's easier to show what I mean, so please see /sandbox2 (rev 1075248360 in case this goes further). The illustrative examples used in the "Parameters" section weren't ideal, so I added a couple. The "casa" example is there, but hidden in html comments for now. There were other changes that might or might not help by way of explanation: you'll see some small-font explanations tacked on to the examples; on the one hand, this makes it wordier, which I'm not crazy about, but on the other, it seems much clearer. Not sure what to do with them. Maybe place them in explanatory notes instead. If they never see the explanation at all, it doesn't really matter, because there's no downside to actually repeating the identical page name multiple times, so maybe that does argue in favor of expl notes&mdash;keeps the main narrative flow of the doc brief, but still provides the longer explanation where needed. Mathglot (talk) 18:31, 4 March 2022 (UTC)
 * Have been mulling over how to make the pairing of lang-code and page-name clearer. To the extent that the concept of paired params can be made clear to the reader, this abstraction cuts the number of effective param [pairs] they have to think about in half. I tried one way to do this, in rev. 1075264704. The dropping of final vertical bar was never adequately explained, and seemed to go against what the doc was saying, so added that as an expl note; any other method seems to interrupt the flow way too much. Mathglot (talk) 20:04, 4 March 2022 (UTC)
 * Have been mulling over how to make the pairing of lang-code and page-name clearer. To the extent that the concept of paired params can be made clear to the reader, this abstraction cuts the number of effective param [pairs] they have to think about in half. I tried one way to do this, in rev. 1075264704. The dropping of final vertical bar was never adequately explained, and seemed to go against what the doc was saying, so added that as an expl note; any other method seems to interrupt the flow way too much. Mathglot (talk) 20:04, 4 March 2022 (UTC)

Template rendering as [] in text
In several articles such as Gerasimov doctrine I've seen interlanguage links work fine one day, and then start to render as " []". When I go through revision histories of articles they appear that way even from the beginning, so I believe it is the product of some kind of strange rendering error, or an update to the template that broke most links. Here is an example of the code that makes this show up, which I believe was generated by the visual editor or something like that because it normally looks like when I type them manually. MaitreyaVaruna (talk) 01:46, 3 March 2022 (UTC)


 * It seems to be somehow related to the link= in the template. renders as garbage and  works fine. Why was link= being added to it anyways since I never added any of these weird new fields and relied on the old fashioned form MaitreyaVaruna (talk) 01:52, 3 March 2022 (UTC)
 * A number of changes were made to this template recently and then were rolled back, so there's the possibility that some of the newer parameters were used and are now invalid. Primefac (talk) 09:00, 3 March 2022 (UTC)
 * Thank you, that all makes sense. I was getting worried that dozens of articles might have been losing content MaitreyaVaruna (talk) 14:10, 3 March 2022 (UTC)
 * I also observed the problem since two days: not only the field link= was involved, but also the field prefix=. The field prefix= also contributes to the problem as the foreign language is not properly displayed after only removing link=. When I removed both fields from the edits I made last weekend, it works again. I left the field foreign= but I see that all these field names are no longer present in the main page. Thank you all for your useful feedbacks and explanations. Shinkolobwe (talk) 14:26, 3 March 2022 (UTC)~
 * I had also to remove the field "foreign=" in the interlanguage link undefined templates that were also not properly working. All the fields must be unnamed for working normally because of these recent changes made to the undefined template. So, the correct syntax without explicitly naming the fields is the following: Charles Darwin (botanist) . The first link is in English (the main red link for the non-existing page in English), followed by the language code and then by the name in the corresponding target foreign language. One can add multiple languages successively to link to more than one single language as also done in the example given hereabove. Shinkolobwe (talk) 18:24, 3 March 2022 (UTC)
 * You have the syntax right, but I find the "Charles Darwin (botanist) example problematic as a usage model, as it is misleading: it links the non-existent botanist (not only a non-existent article, but a non-existent person in history) to the articles about the famous evolutionary biologist in French, German, and Spanish, which would be an error if the goal is link articles about the same topic. A better example is this one, if the titles are identical in other languages:
 * which results in: Tobias Hedström
 * If the names are not identical in the other language, then you would have to fill them in:
 * ⟶ Olena Chaplynska
 * Hope this helps. Mathglot (talk) 00:13, 4 March 2022 (UTC)
 * Thanks for your answer and your relevant comment. I agree with you. I directly borrowed the example from the documentation of the ill template and probably did not use it as it is exactly presented, but the idea of Charles Darwin (botanist) directly comes from the ill template documentation which is also sometimes not so clear. Have a look at the excerpt from the main page hereafter: Modifying the display Displaying different text To create a piped link (text displayed that is different from the title of the page to which the text links), use the |lt= parameter (which stands for "link text"). This is useful if disambiguation is necessary, for example to hide the (botanist) in Charles Darwin (botanist):... So, I would propose to modify the example in the documentation of the ill template to avoid to be misleading as you observed. Shinkolobwe (talk) 00:39, 4 March 2022 (UTC)
 * , thanks very much for your comment. I'm very interested in what users think about the documentation, especially if they haven't been here before, or are not a "regular" with this template. I'm probably *too* familiar with it in some ways, to see as clearly as I'd like where the difficulties lie, and so your insights into the doc are highly valuable, imho.
 * The documentation page (here) is open to editing by any editor. If you feel comfortable doing so, I encourage you to edit the doc page directly, and improve the example in a way that you think would make it clearer. Feel free to throw out that example entirely if you think a different one would be better. Don't worry about getting it "perfect" or "exactly right" if it's an improvement, that would be wonderful. If you're not comfortable editing the /doc page directly, can you explain below your idea, or write out the documentation text that you would like to see go into the page, and add it below? Thanks! Mathglot (talk) 01:19, 4 March 2022 (UTC)
 * Thank you very much for your answer and the link to the documentation page. I also discovered it yesterday when viewing the source code of the ill template. I will made some comments, or at least leave unclear templates, where I think it could improve the understanding. However, today and in the next days, I have also other tasks to complete, so that I am lacking of time. I think that the documentation of the ill template is now clearer than one or two years ago when I was sometimes quite puzzled. It is an important template to indicate to the reader that information is available in other languages and to foster the translation of pages. Best regards, Shinkolobwe (talk) 10:56, 4 March 2022 (UTC)


 * Just noted the issue, any chance to have it fixed without having to manually edit each link? It affects dozens if not hundreds of articles I created (Ivan Matušík is a recent example). Cavarrone 11:17, 6 March 2022 (UTC)
 * I checked your last 100 page creations (back to 23:19, 10 February 2022, and the creation of Alberto Cavaciocchi) and the only problems I found with the use of interlanguage link were two occurrences in the article Ivan Matušík, which I've fixed. If there are other examples that I missed, please provide the dates or list the articles, and I'll take care of it. Mathglot (talk) 08:56, 8 March 2022 (UTC)
 * , thank you very much for reviewing the pages and for fixing the links (side note, let's not exaggerate my productivity, currently Cavaciocchi is just my 36th most recent article!) It's my fault, I had no time to check and as long as I do a large use of interlanguage links I expected a massive disruption, now I'm taking a cursory look and even going before Cavaciocchi (randomly checking Heo Cham, Stanislav Rudolf, Alexander Timofeevskiy, Cinico Angelini) all the interlanguage links are apparently working. Many thanks again. Cavarrone  19:53, 8 March 2022 (UTC)

How to display target language link as-is?
How do I do this using the "ill" template? Pasco Shikishima Corporation (敷島製パン) Currently coded like so: Using the "ill" template, I would like to display the English title for the English article, and the original title of the linked article in the other language. --Kai Carver (talk) 17:09, 20 March 2022 (UTC)


 * Using ill, you could write Pasco Shikishima Corporation (敷島製パン) to hide the jawiki link once an English article is created but still display the Japanese name. —Kusma (talk) 20:11, 20 March 2022 (UTC)


 * thanks... I guess that's a solution.
 * Too bad it's ugly with the repeated language label and repetitive to write... Maybe the inter language template will get better... —Kai Carver (talk) 03:53, 22 March 2022 (UTC)
 * As I understand your question, you are dissatisfied with two things: a) that the template does not spell out the foreign-language link (it just gives &#91;ja&#93; ), and b) that the offered solution is cludgey to write? If so, the proper discussion regarding a) to have first is: why is it important to you to spell out "敷島製パン" - why can't you settle for merely the link? Once we know your reasoning perhaps we can offer a better solution! Remember, it is generally a good thing that the foreign-language links of ill remain as unobtrusive as possible. As for b) sorry but editor convenience is not and should not be our first priority. Ponder how seldom this will be used. Maybe the slight inconvenience is acceptable... and that the template doesn't really need getting "better"? Regards CapnZapp (talk) 10:56, 22 March 2022 (UTC)
 * Hi, and thanks for your answer @CapnZapp, and sorry, I did not mean to be over critical.
 * I came across this case for Chinese, here: March 18 Massacre
 * This:
 * Liu Hezhen (刘和珍)
 * seems to me like a good presentation to link to an article which has no English version (and may never have one).
 * But I coded it like so:
 * and I'd rather use ill, and couldn't figure out a way to have the same presentation.
 * a) I think it's important to also show the Chinese characters in the English text because it's more precise, and clicking through to the Chinese link may be confusing for non-Chinese speakers. Both points may well be arguable.
 * b) I'm happy to use whatever inconvenient method is best, I just couldn't figure out what the right solution was to both make the presentation best and use ill. Kai Carver (talk) 08:55, 23 March 2022 (UTC)
 * I would recommend you to first consider this usage: Liu Hezhen that so far is deemed a good presentation of an article which has no English version (but could have). If you decide there likely will never be an English language article, don't create a red link at all. In those cases, don't use Ill - instead link to the foreign-language content by other means: Help:Interwiki linking is likely a good start. If you do feel an English article is plausible (and a red link is justifiable), and deem the current way Ill presents the foreign-language link insufficient, you can start a discussion whether it really is less precise or confusing. I really don't think it is a good use of the project's resources to offer what you want prior to forming a consensus it is actually needed. Regards, CapnZapp (talk) 09:38, 23 March 2022 (UTC)
 * > I really don't think it is a good use of the project's resources to offer what you want prior to forming a consensus it is actually needed.
 * huh? OK, my bad, whatever. Sorry to put such a strain on "the project's resources" with my question, jeez. Kai Carver (talk) 10:36, 25 March 2022 (UTC)
 * I agree on the importance of showing the Chinese characters, especially if we don't have a local article. This introduces precision and avoids at least the homophone-induced disambiguation headaches. The main encyclopaedic issue here is the complete lack of sourcing on both March 18 Massacre and zh:劉和珍; once there are some sources, it may be easier to decide whether Liu Hezhen should be a red link, a redirect to the massacre, or a standalone article. frwiki and jawiki have some references that one could check out. —Kusma (talk) 10:03, 23 March 2022 (UTC)
 * huh? OK, my bad, whatever. Sorry to put such a strain on "the project's resources" with my question, jeez. Kai Carver (talk) 10:36, 25 March 2022 (UTC)
 * I agree on the importance of showing the Chinese characters, especially if we don't have a local article. This introduces precision and avoids at least the homophone-induced disambiguation headaches. The main encyclopaedic issue here is the complete lack of sourcing on both March 18 Massacre and zh:劉和珍; once there are some sources, it may be easier to decide whether Liu Hezhen should be a red link, a redirect to the massacre, or a standalone article. frwiki and jawiki have some references that one could check out. —Kusma (talk) 10:03, 23 March 2022 (UTC)

Option to remove enwiki link
Maybe there should be an option to remove the en wikilink in case, for example, such a link would violate WP:SELFRED. I can think of many examples of such self-redirection (one such example on the Bitbird article) that would benefit from such a wikilink removal. Jalen Folf  (talk)  23:58, 11 March 2022 (UTC)
 * That would be problematic. I assume you refer to Marcioz in that article. If the English link,, would be suppressed and that article is created in the future, there would be no link. But that is exactly the purpose of . I suppose this scenario could be avoided if the the template code would observe the parameter to suppress the English link for redirects. What name do you suggest for such a parameter? -- Michael Bednarek (talk) 03:10, 12 March 2022 (UTC)
 * I think the greater issue here is the to my mind inexplicable drive to remove the foreign-language link when English content is written. That is, how a bot will remove the portuguise link once Marcioz is expanded into an article. Anyway - while a bot can devote the computational power to ascertain whether a link is "only" a redirect or a "proper" article, we should not ask a template to do this job (this template is too demanding as is). So my answer would be maybe not. Maybe better to amend SELFRED to not worry about special cases, which I think Ill qualifies as? CapnZapp (talk) 12:37, 12 March 2022 (UTC)
 * a) No bot is required to remove interlanguage links once an English article exists – that's what this template does automatically (although there is a periodic clean up by Cewbot). b) I agree that WP:SELFRED is a very minor guideline and I wouldn't worry about it. -- Michael Bednarek (talk) 13:44, 12 March 2022 (UTC)
 * These redirects will likely all be R with possibilities and not violate the spirit of SELFRED. I don't think action at this template is needed or helpful. —Kusma (talk) 15:32, 12 March 2022 (UTC)


 * Just a thought, but might there be an advantage to having a self-redirect parameter that could invoke a different color and action? I'm thinking the ideal would be having it detect if it's a redirect and show the link in red and be . You'd definitely need to detect the redirect first, because you would want to link once even a stub article was created, but that could be complicated - maybe even requiring a change to site-wide CSS. But another possibility that would be easy to implement would be to just link as   without changing the link color at all. VanIsaac, MPLLcont WpWS 21:04, 12 March 2022 (UTC)
 * I just ran into this situation at Eric X. Li, where an interwiki link to zh:观察者网 is justified, but using ill creates self-redirects to Guancha.cn. Self-redirects are confusing for readers, so I don't like the idea of just letting this happen. I could just do it manually, but then that wouldn't go away if the article is created; I could put a custom #ifexist there, but directly using parsers in articles is rarely desirable, plus it wouldn't be cleaned up by Cewbot. I don't like the idea of adding a no-link parameter to this, but what we could do is change  to   (or probably something a bit more optimzed than that, to avoid calling Redirect twice... might make more sense to Luafy this after all these years). The downside here is that, yes, no link to click on to expand that r with possibilities into an article, but SELFREDIR represents a consensus that such links should not exist.On that note, Vanisaac, I think your suggestion is interesting, but it would only make sense if there were a way to do that in non- situations too, something like a self-redirect with possibilities, which would produce something like , either with that link the standard light blue or custom-styled yellow or orange or something, and with the same bot cleanup that this template gets. If there were consensus for such a template, then I would support adding the same functionality to this template.  --  Tamzin  [ cetacean needed ] (she/they) 02:22, 20 April 2022 (UTC)
 * I don't fully understand what you are proposing (can you give a mockup?) As for self-redirects, I think those that go through ill are generally less problematic and more useful than others (and while I don't particularly like redirects to sections, is the one in your example really so bad?) Is there any systematic SELFRED cleanup going on anywhere? —Kusma (talk) 15:31, 20 April 2022 (UTC)
 * Mockup: Sure!, when used on this page, to which Template talk:Ill redirects, outputs:
 * I'd actually forgotten that I'd refined that redirect to a section, which I do agree makes that particular case less problematic. Still, there are many cases where refining to section isn't an option.
 * I'm not aware of a systemic cleanup, although XFDCloser automatically removes them when closing as delete, and I think many editors (myself included) remove them when we come across them. Looks like there are about a quarter million self-redirects, but most are alternate names that are linked in navboxen or infoboxen—something to be avoided still, but less of an issue.
 * The more I think about this, though, the more I wonder if this is a more general question about links to rs with possibilities. They fall in an awkward middle-ground between SELFREDIRECT and REDYES. Maybe some sort of "orange-link" template really is the way to go. --  Tamzin  [ cetacean needed ] (she/they) 18:20, 20 April 2022 (UTC)
 * I'm not entirely sure what you are discussing but I do know more colors in links is not the way to go. The simplicity of red and blue links must not be destroyed by extending it. Of course, what you do (with skins or whatnot) is your business; talking the general vanilla Wiki experience here. Sincerely CapnZapp (talk) 19:47, 20 April 2022 (UTC)
 * Thanks, now I understand. But I'm not convinced this is an improvement. (Readers aren't used to black text with little light blue decorations about foreign languages). —Kusma (talk) 23:12, 20 April 2022 (UTC)
 * Let me here add that I genuinely think this might be overthinking it (or maybe let's call it misguided concern for the reader). I do not believe the reader is to any significant degree used more to red links with "little light blue decorations" than to black text with the same. I think it is only us experienced Wikipedians who think Subject &#91; fr &#93; is odd while Subject &#91; fr &#93; is not. CapnZapp (talk) 09:21, 21 April 2022 (UTC)
 * In my initial comment, when I was talking about controlling the color, it was specifically in the context of coloring self-redirects as a red link when the software would normally color them blue. It was not about making a whole rainbow of link colors, even though I run an extension which colors disambiguation pages orange. VanIsaac, MPLLcont WpWS 23:17, 20 April 2022 (UTC)
 * In my initial comment, when I was talking about controlling the color, it was specifically in the context of coloring self-redirects as a red link when the software would normally color them blue. It was not about making a whole rainbow of link colors, even though I run an extension which colors disambiguation pages orange. VanIsaac, MPLLcont WpWS 23:17, 20 April 2022 (UTC)

Links appearing as red when they should be blue
Recently I have noticed numerous cases where this template displays links as red when they should be blue. For example here Secular Shrine Theory#Internal Shinto controversy with this link why is this happening? Was something changed?

Archive link where it is happening in case it fixes itself https://web.archive.org/web/20220426205616/https://en.wikipedia.org/wiki/Secular_Shrine_Theory

I also notice links that go to redirects display as blue but also have the foreign link still present, is that an intentional feature (so that you can get more detail on the foreign page when english wikipedia doesn't have a dedicated page), or an oversight?

Immanuelle 💗  (please tag me)  20:58, 26 April 2022 (UTC)
 * The example you gave seems to have suffered from a cache-induced delay; it's showing blue now. // The behaviour for redirects is intended. -- Michael Bednarek (talk) 01:53, 27 April 2022 (UTC)
 * To clarify for newer Wikipedians. The reason this discussion was had was because the link Proclamation of the Great Religion was newly created (its only days old). The coloring of the links happens on the server side; if the server's cache doesn't yet mark this page as created, we get a red link even though the destination page exists. If you visit the supplied web archive you'll notice it's not just the color that is different with red links. (In fact the color is supplied by CSS and thus the client. To do that something needs to be different about what the server sends our browsers) As explained, this is a short-term issue only, and indeed was resolved in the time between question and answer. The reason redirects behave this way is because it is not uncommon for a redirect to be less useful than a "proper" article. When a proper English-language article is created, the ill template switches off its behavior (and the template itself is eventually removed by a bot), since presumably foreign-language articles are no longer needed (and can be reached through the sidebar to the left, under "Languages). But when a redirect is created, this behavior often turned out to be less than desirable, since it could redirect to a page with less info specific to the subject sought out than present on the foreign-language article linked through ill. In an effort to remain helpful, the ill template keeps displaying both (and isn't removed by the bot). CapnZapp (talk) 07:43, 27 April 2022 (UTC)

Use in hatnotes
The documentation currently has the following prominently displayed sentence (added in 2019 after a suggestion in a discussion):

This is justified by a reference to WP:REDHAT. But that sort of misses the point of the guideline. Red links are disallowed in hatnotes not because there's something wrong with the colour red, but because hatnotes are for navigation and a red link doesn't navigate to anywhere. But a link produced with this template will have at least one navigable blue link, so this isn't really applicable; interlanguage links are, in principle, accepted in at least one other area where navigation is everything: disambiguation pages.

Of course, given the prominent position of hatnotes in articles and the usually small target audience of interlanguage links, their use should probably be avoided in many, if not most, circumstances. However, such avoidance is a natural consequence of the sort of thought processes that normally go into editing hatnotes, and exceptions may be numerous. I don't think we should be pre-empting such processes by an explicit rule that rules out common-sense exceptions, gets editors into unnecessary quandaries, and – given how rare it is for an interlanguage link to be added to a hatnote in the first place – stands as an instance of instruction creep that distracts from the more widely relevant passages in the documentation.

Thoughts anyone? – Uanfala (talk) 21:09, 13 March 2022 (UTC)
 * If there's significant use case potential, something like an interlanguage version of template might be justified. VanIsaac, MPLLcont WpWS 21:48, 13 March 2022 (UTC)
 * I don't think there's a need for a separate template: a templated link can be inserted in the same way as any other custom text (e.g.  or   (though there is the recently created Template:Further ill as well). – Uanfala (talk) 23:48, 14 March 2022 (UTC)
 * Here's a use case. I've written a GA with ill in a distinguish hatnote: Hermann Boeschenstein, who is not Hermann Böschenstein, although both can be found using either spelling, both were Swiss, both lived during most of the 20th century, both were widely published in Swiss newspapers. I almost added a picture of the wrong Hermann Boeschenstein to the article before I found out. To prevent others from getting confused (apparently also happens to librarians, see de:Diskussion:Hermann Böschenstein) I added the hatnote. —Kusma (talk) 23:09, 13 March 2022 (UTC)
 * Thank you for that, User:Kusma. CapnZapp (talk) 07:02, 14 March 2022 (UTC)
 * It appears you have identified an acceptable exception to a rule, User:Uanfala. That does not necessarily mean the rule is wrong however. I would at most insert a "normally" to the cited documentation sentence, that is, to add "normally" into "This template should not normally be used in hatnotes as red links should be avoided in hatnotes" (here I also replacing "should not be present" with the better "should be avoided" but this is just a suggestion). In other words, avoiding red links in hat notes except where justified, is the way to go, and we normally don't need to qualify all our rules with "but please break them when you feel you have to": [[WP:BRAR. Regards, CapnZapp (talk) 07:01, 14 March 2022 (UTC)
 * That doesn't really address my concerns above. To be clear, I think we shouldn't have explicit rules about the use of the template in hatnotes: the situation occurs only rarely and when it does it can be handled by the more or less straightforward application of common sense. If we're going to have something explicit here, then that, in my opinion, should be something along the lines of This template may be used in hatnotes, with a pointer to a section to be added at WP:HATNOTE, which will then 1) mention that interlanguage links in hatnotes can be formatter either directly or with the template; 2) list the advantages and drawbacks of each method, and 3) outline the factors to consider when deciding whether to include such a link in the first place. – Uanfala (talk) 23:48, 14 March 2022 (UTC)
 * Then you should go ahead and make an edit, which will serve as your suggestion! (So far nobody seems to be in direct objection) CapnZapp (talk) 08:05, 15 March 2022 (UTC)
 * I come in here a little late and perhaps I miss the point, but my thinking is that the statement in the documentation not to use in hatnotes should be replaced by a statement providing explicit permission to use it in a hatnote, inasmuch as  actually provides a working link.  This avoids the risk of somebody re-inserting the current statement to avoid using it. Fabrickator (talk) 22:22, 20 March 2022 (UTC)
 * I see you went ahead and boldly amended the instruction. Has anyone told "the hatnote people" about this? What I am asking is: could it be that the discouragement about using ill in hatnotes is related to other things than the redness of its (main) link. Perhaps there once formed an consensus to avoid "weird" links (such as foreign-language links created by ill) in hatnotes, perhaps in the olden days when hatnotes themselves were new and slightly weird...? (Just an observation; not opposing anything here) CapnZapp (talk) 07:49, 27 April 2022 (UTC)