Wikipedia talk:Templates for discussion#TfD failure

I wish art had been kept. In whose contributions can I see the replacements? Nardog (talk) 12:46, 22 January 2025 (UTC)

:Mine. I used {{tl|IPA|und}} as the template says to do. Gonnym (talk) 12:55, 22 January 2025 (UTC)

::Thanks, added specific ISO 639 codes as far as I could. You also forgot {{para|label}}—see Bertilo Wennergren, Chicxulub crater. Nardog (talk) 13:48, 22 January 2025 (UTC)

{{Edit protected|level=none|Module:Tfd links|answered=yes}}

Please upgrade Module:Tfd links to add the extra "substed" link produced by {{tl|Tfd links2}}. E.g.,:

• {{tlg|Tfd links |Uw-tpv4|code=yes|_show_result=y}}
• {{tlg|Tfd links2|Uw-tpv4|code=yes|_show_result=y}}

Thanks, Mathglot (talk) 10:46, 14 March 2025 (UTC)

:Comment: The second template uses {{tl|Find substed}}, which is rated as "alpha", not ready for widespread deployment. – Jonesey95 (talk) 16:15, 14 March 2025 (UTC)

:: I also note that it assumes that the template includes a comment with the name of the template in the substed output. That's not guaranteed. Anomie⚔ 23:17, 14 March 2025 (UTC)

:: Yeah, I put the alpha rating there, because it involves some features of Cirrus regex that were new to me, but I doubt they will be new to Lua programmers. The point is, the concept is sound, you can test {{tl|Tfd links2}} (or {{tl|Find substed}}) with a ton of templates, and afaict it always brings up the correct accurate (even if not complete) results. Maybe a Lua programmer will find an even better, more robust regex than I did, but even if not, the new link is a huge boon to Tfd responders who up to now, simply cannot find real examples in the wild of templates that have been substed. {{ec}} Mathglot (talk) 23:25, 14 March 2025 (UTC)

:: Not guaranteed, but there for all user warning templates currently, and could be introduced gradually to others over time. We shouldn't let the perfect be the enemy of the good: this is a clear win for Tfd responders who currently do not have a clue who is using a substed template or why, and thus have little concrete data upon which to base a decision other than general principles. It could also be used to loop in actual, substed-template users into a discussion to get their perspective; this was done here, for example. If need be, just change the link text, to 'substed (labeled)' or something. At worst, the new link could be suppressed, for those always-subst templates that do not contain the template label in the source. Mathglot (talk) 23:51, 14 March 2025 (UTC)

::: Actually, not suppressed but unlinked (maybe even grayed out) would be ideal, because that would tell you at a glance that you were dealing with an untagged, always-subst template, so you are basically SOL trying to find uses of it, unless you know your regexes and the message has a unique expression you can search for. But not too many responders are going to do that, even if they know how. Mathglot (talk) 01:28, 15 March 2025 (UTC)

::::I support the addition of this link as quite useful, as long as we trust {{tl|Find substed}}. – Jonesey95 (talk) 13:10, 16 March 2025 (UTC)

:As I've raised here, if any new link should be added it's a general insource link. Gonnym (talk) 13:35, 16 March 2025 (UTC)

::That's what this is. If you have a different suggestion for different code, you can add it to the sandbox. – Jonesey95 (talk) 21:19, 17 March 2025 (UTC)

:::This isn't it. This searches for a hidden comment, which means that any actual usage of a template hidden in code isn't found. My oppose still stands, this code isn't good enough. Gonnym (talk) 10:42, 23 March 2025 (UTC)

:::: Can you elaborate or explain? I do not understand what "{{xt|any actual usage of a template hidden in code isn't found}}" means. Alternatively (or in addition), what code (or functional result of better code) would be good enough? Maybe it can be improved, if I can understand what the issue is. Mathglot (talk) 19:41, 23 March 2025 (UTC)

:::: Still struggling with understanding the objection. One interpretation I came up with for your remark, is this: "This template does not find substed templates that have no hidden identifier, so the result set will necessarily be incomplete." That is, of course, entirely true, and I know no way of doing so, as a substed template leaves no trace. That is the whole point of having the id, to leave bread crumbs, and what this template attempts to find. If this interpretation is correct, what if we simply change the name of the template to: Template:Find hidden template id? Then, the result set would be complete, there would be no further implication about finding substed templates, and the objection would go away in that case (iiuc). Would that work for you? Mathglot (talk) 19:51, 23 March 2025 (UTC)

::::: We could upgrade it to check whether the template contains the identifier, and by default, return empty string if it doesn't have it, which would signal Tfd links not to emit the link (unless {{para|force|1}}, then it could emit a link that returns zero results). Mathglot (talk) 20:51, 23 March 2025 (UTC)

:::: Pinging {{u|Primefac}}, who took part in the archived discussion you linked, who maybe can shed some light on this. Mathglot (talk) 20:18, 23 March 2025 (UTC)

:::::I've seen this, but I also just moved so I have a bit of a backlog of things to look at. I'll remove this comment and add something more substantial when I can. Primefac (talk) 08:43, 25 March 2025 (UTC)

:::::: Didn't know this yesterday, but by pure serendipity, today I ended up at {{tl|Uw-vaublock}} making a small wording change to satisfy an edit request – and then I noticed this edit of yours, in order to "{{xt|rmv tracking template following TFD (can be tracked via commented code)}}", which seems eerily connected. Indeed it can be tracked that way, and this proposal hopes to democratize that and put it in the hands of everyone. Mathglot (talk) 00:32, 27 March 2025 (UTC)

:::::::Just noting that the tracking Z-number templates were removed because they were getting ridiculously large (I think at one point we had something like 300 of them) and also being re-used. Since most subst-only templates have a hidden comment giving the template name (i.e. the provenance of the message) TFD determined that the Z-numbers were unnecessary. If a subst-only template does not have a hidden comment, then it should probably have one. I'm not sure what you mean by democratisation, but anyone is welcome to edit a template to add a hidden comment (or put in a TPER if it's protected). Primefac (talk) 11:09, 5 April 2025 (UTC)

:::::::: Sorry I wasn't clear; what I meant was that folks in this discussion are likely to know how to toss out a Cirrus search like User talk: insource:"Uw-tpv4" insource:/\/ without much trouble but most users won't (even Tfd users; even a simpler one without the hidden bit). The whole point of the template is "democratization"—i.e., to bring this squirrely search ability to everyone. Mathglot (talk) 16:35, 5 April 2025 (UTC)

• Listed at: WT:WikiProject Templates and WT:Template index. Mathglot (talk) 16:41, 25 March 2025 (UTC)
• How big are the template limit concerns these days? I removed some links a few years ago when we were hitting them like every week. Trialpears (talk) 10:41, 27 March 2025 (UTC)
• :They are of much less concern today. Back in 2020 and 2021, we worked on a big project to clean out a thousands of unused template-space pages, so a perhaps over-eager group of us were crowding TFD with hundreds of nominations per week. That caused a bunch of problems with limits. That project worked through its backlog a couple of years ago, and now we are at a steady state, just picking off a few at a time. The current TFD page has just 87 nominations, which is about average and doesn't come close to hitting the limits. – Jonesey95 (talk) 23:04, 28 March 2025 (UTC)
• :: Deactivating edit request as there are plenty of template editors discussing this so you don't need an uninvolved template editor to implement. * Pppery * _{it has begun...} 20:31, 30 March 2025 (UTC)
• In general I am not opposed to adding an extra search function for finding subst-only templates, but I also agree with Gonnym that the proposed implementation is a bit unnecessary; we don't need to be looking for an entire commented-out template name, just the template name itself. Yes, there might be some false positives (e.g. someone just typing "uw-tpv4" like we are doing here) but the point of the tfdlinks is to show general/broad usage, not be super-specific. For example, swapping namespaces once the search is pulled up is pretty trivial, so I don't really see the need for extra params. I guess my point is that it's reduced overhead for maintenance and updating if we just add a single line of a simple URL to the main template, rather than relying on a secondary template to do so. Primefac (talk) 11:05, 5 April 2025 (UTC)
• : Not opposed to adding a link directly to Module:Tfd links rather than employing a separate template, and I am okay just searching for the template name (as opposed to the longer string) if that doesn't include too many false positives in the result set (and even if so, it's probably manageable). As far as what is or isn't trivial for users to do in search, this links up with the earlier comment about "democratization": I wonder how many users even know about the namespace checkboxes, and that if they start off with a regular search for {{nowrap|"uw-tpv4"}} in the search box and then click the advanced link, that they can refine the resulting zero-result search page by switching namespaces via a checkbox; my guess is pretty few. Perhaps something to look into, or maybe a user plugged into wmf research like {{u|WhatamIdoing}} might have some data on that. One purpose of the template is to finesse that, and give them results immediately; a module upgrade could do that as well.

:: Also, I had in mind to save Tfd users a wasted click by not rendering the link (or rendering it grayed out, unlinked) if there are no results but I haven't sandboxed that yet, so if we do implement the functionality in the module without a template, I would like to see that included as well, if possible; however, if that complicates implementation too much, just having a basic link there to find id-bearing, substed templates (or just any mentions, even at some cost in false +) would be a boon to Tfd users trying to make a more evidence-based decision. Mathglot (talk) 16:43, 5 April 2025 (UTC)

:::I have seen no research on that specifically. However, in general, if it's behind a collapsed box (and this sort of is), then most people won't know about it. That feature should be known to the typical power user but not to ordinary users (e.g., readers). WhatamIdoing (talk) 21:17, 5 April 2025 (UTC)

There is a discussion at WT:CRIC about squad navboxes that may interest some of the editiors here. It is an attempt to gain consensus on which squads in international tournaments should have a navbox created. It does not list a specific navbox, so not sure if I should add a link to it on one of the daily pages, or if this is the correct place. It pings a number of WP:CRIC members, but seemingly no TfD regulars. Spike 'em (talk) 10:36, 25 March 2025 (UTC)

There are many navboxes that are absurdly long or dense to where it would be better served as a few different navboxes, and I'd like to get some of them split up into multiple smaller navboxes, but it's unclear how this should be discussed since they affect so many articles. Thebiguglyalien (talk) 🛸 17:02, 11 April 2025 (UTC)

:I believe there is a specific place to discuss merging templates, but I could not find anything when it comes to splitting them. I simply sent messages to those I believed to be most familiar with the topic. Hotdog with ketchup (talk) 21:16, 11 April 2025 (UTC)

::The first place to discuss a split is on the template's talk page. If there is little/no discussion, and/or no consensus can be reached, then TFD would be an appropriate venue for wider discussion. Primefac (talk) 00:12, 12 April 2025 (UTC)

: For starters, imho it always helps a discussion if you link a few examples at the outset, especially when you OP starts out, "{{xt|There are many <Wikipedia thingies> that are <undesirable attribute>...}}", as it's hard to know if you have something specific in mind or it's just a general gripe, and also it helps tune responders tune into the same wavelength you are on.

: Additionally, you pose one problematic attribute (long/dense) and one solution (split them) whereas in reality, there may be many other solutions you have not considered (better group/label hierarchy, child navboxes, selective collapse, compound navboxes, create and/or link to list page or outline page) and so on. You increase the likelihood of getting a satisfactory response for your situation if you describe the underlying issue that prompted your question, rather than your suggested fix (split), because that tends to limit responses to "How do I split this" rather than whatever the underlying issue is.

: Is it that it is hard to find a link in the box? Is the organization or order of the many links non-intuitive or arbitrary? Does it take too much vertical space and makes scrolling past it annoying? Does it take too long to load the page? Do you prefer succinct lists as one tends to see in a sidebar on the right side of the page rather than in a navbox? Do you hate navboxes and wish they would all be suppressed when you look at a page? Are you trying to track down where a particular source string resides using Advanced search, and the voluminous navboxes give you too many search results to pages that only transclude the string but do not contain it? Is the font too small and your eyes glaze over when there are more than a few links?

: All of these are legitimate issues and might all be [overly] summarized as, "navbox too long", but they all have different focus and a different range of solutions that may apply. Describing your pain points regarding navboxes might help you get to a solution easier and faster. Mathglot (talk) 00:23, 13 April 2025 (UTC)

::I want to split them. I want this because they are too long. Examples: {{t|United States topics}}, {{t|Fascism}}, {{t|Russian invasion of Ukraine}}. Thebiguglyalien (talk) 🛸 00:35, 13 April 2025 (UTC)

::: At the risk of repetition, what does "too long" feel like to you, or where does it impede usage of the encyclopedia, or make some article worse by including the navbox, or cause some other negative effect? Personally, I do not think they are "too long" but I suspect we mean different things by that expression. As one approach, I am the kind of person that occasionally likes to browse or read a printed dictionary (or an online one) stopping at some of the unfamiliar terms to read them in detail. A bunch of editors spent considerable thought and effort into designing {{t|Russian invasion of Ukraine}}, and it's a great one-stop shop of topics with comprehensive coverage of the issue; I don't know where else one could find a list like that to quickly access almost any aspect of the invasion; a tremendous time-saver for someone interested in a deep dive. Finding a navbox such as that one or the others you linked feels to me like hitting the jackpot, with all those yummy unfamiliar topics to learn about, and maybe improve, as well. So to me, they are a boon. It takes a little longer to scroll past, but that is a minor inconvenience to me. Knowing where your discomfort lies with the length of {{t|United States topics}} might lead to a solution you didn't expect. Mathglot (talk) 00:48, 13 April 2025 (UTC)