Wikipedia talk:Template index/User talk namespace

Besides the image in the article, File:Mediawiki-button-preview.png, corresponding to the source editor, we now have a new image, File:VisualEditor save dialog Review link-en.png (right), which can be used for those who use the Visual Editor (the default editor, and the one used by the majority of new users). I am thinking about how best to take advantage of this, and welcome input. Mathglot (talk) 09:43, 17 February 2026 (UTC)

Some obvious things come to mind, like add a new param to {{uw-preview}}, for example param |editor= with values VE or wikicode If we could interrogate the user profile and just use their preference, we could skip the param. Another idea I kinda like, is you know those radio buttons some city infoboxes have (like this one), where you can pull up either a locator map of the surrounding region, or the whole country? So, how 'bout we include both images in the template, show one by default (which?) and provide radio buttons labeled for each editor. What other ideas could we dream up? Mathglot (talk) 02:36, 18 February 2026 (UTC)

For the record: there is an issue involving the possible misidentification of substed warning templates. If you are not a template gnome, data geek, or a researcher investigating Wikipedia templating, you can stop reading now and go do something constructive, like weighing your belly button lint.^[a]

Warning templates generally include their own pagename within hidden text when substed, as an aid to users wondering where the message text source is located. They may be also be used to determine where (and how many times) a template is substituted. In some cases, warning templates have the wrong hidden id, whether intentionally to show a cloned template, or resulting from a forgotten copy/paste operation, or some other reason. In particular, at least three templates identified themselves with a Template:uw-vandalism1 hidden tag other than the template itself: {{Welcome-3rr}}, {{Huggle/warn-1}}, and {{Welcome-unconstructive}}. (There is also uw-cluebotwarning1 which uses two hidden tags, including uw-vandalism1; not sure about Twinkle.) The first three have been fixed, but searches such as this one (2.5M results; ~20 sec.; timeout likely) will return results from any templates that have the tag, and the tally returned is the sum of all of them. (Toldja not to read this far.) Thanks, Mathglot (talk) 01:25, 18 February 2026 (UTC)

I removed the extra tags from one series -- I have a guess as to what they were originally included for. Importantly: the extra tag in the welcome templates caused semi-automated counter vandalism tools like Huggle, AntiVandal, WikiShield, and Interceptor to count them as a level 1 warning, which was probably intentional and removing it will cause future uses of the template to not count as a warning in those tools. --tony 01:43, 18 February 2026 (UTC)

Thanks; that's what I just heard from Naomi at my Talk page. In that case, maybe we need a bigger discussion about it, as including wrong tags is a kind of quick fix that every programmer is familiar with, and the time savings may be worth it initially, but there is also a risk, and we just ran into one. It isn't that complicated to get tools to do the right thing, but then it ought to be based on a proper design that all tools follow, not a clever hack of using a wrong identifier that gets you by, until it doesn't. We are now at "doesn't". Mathglot (talk) 01:58, 18 February 2026 (UTC)

It depends on whether you see <!-- Template:uw-vandalism1 --> as a statement that this is literally {{uw-vandalism1}} or if it is a kind of level 1 vandalism user warning. It obviously started as the former, but these days is used as the latter. It works reasonably well, too. Otherwise the only real solution would be to add some sort of machine-readable categorization to the templates and update all of the tools to use that. Something like <!-- uw:1:vand --> or similar but then you'd also have to update the thousands of already subst'd warnings, or support both formats for a very long time in any tool. The only other way would be to have a comprehensive listing of all warning templates ever and hope that all the tools consume it, which seems even more difficult. — Naomi Amethyst 23:41, 18 February 2026 (UTC)

I was going to propose almost the same fix, but you beat me to it. This situation reminds me of old, legacy code like CONST: five = '7'; because the original guy who wrote five = '5' should've called the named identifier 'max_floors' and not 'five', because they couldn't conceive of the zoning height regulations ever changing. That is a failure of robust design on the part of the original designer, exacerbated by the underpaid, rushed contractor the city hired when the zoning regulation changed, who didn't give a bleep and anyway had another contract starting next Monday so they just took the easy way out rather than search the code for every occurrence of 'five' and change it to something reasonable.

Our current situation is analogous to that, except that rather than underpaid we work for free, and it's human to take the easy way out. That doesn't change the fact that it was a hack from the very beginning, and now we are where we are.

It works reasonably well, too.

For what you want, maybe but it breaks everything in a domain that perhaps you don't find interesting or important. To me, it is a wrecking ball. I agree with half of your proposed solution:

the only real solution would be to add some sort of machine-readable categorization to the templates and update all of the tools to use that. Something like <!-- uw:1:vand --> or similar ...

Yep, that was what I was going to suggest, too. But I would not call it uw- anything, because that would imply only user warnings, whereas at least welcome templates would have to have it, too, and maybe other things as well. And I would not call it -vand- either, as it is not currently only used for vandalism, either. What the flag really wants to tell automated software, iiuc, is two things:

• "this is one level of a multi-level template sequence"; and
• "the last/current level is N" (e.g., where N = 1, 2, etc).

Did I miss anything? If not, I would propose an identifier label like |Last-multi-level-seq= for example.

... but then you'd also have to update the thousands of already subst'd warnings, or support both formats for a very long time in any tool.

Updating existing substed warnings – er, "warnings and welcomes [and other stuff?]" – is off the table for multiple reasons. I don't see why you have to support both for a long time, but so what if you did? Once the code is in place to recognize two tags, the template name or some new tag, I don't see it as a huge burden to leave it in forever. And it isn't the *only* solution. Date-conscious detection would be another approach, although I find it messy and wouldn't recommend it, but not my call. The bottom line is that the current situation is untenable and some solution needs to be found. I don't find it excessively unreasonable to require that a hidden tag that contains the string <-- Template:Sunny-vacation-spots --> means that this substed placement originated from Template:Sunny-vacation-spots, and not have some cryptic meaning completely unrelated to the template name in another domain. Mathglot (talk) 01:06, 19 February 2026 (UTC)

I would include something to indicate what "family" of warnings it is. Many tools care about what kind of template warning it is. ClueBot NG feeds that data into it's ANN, for example. — Naomi Amethyst 02:03, 19 February 2026 (UTC)

Not 100% sure I know what you mean by "family", but to the extent I understand, for some things there isn't always one. Are you saying there's a third bullet that ought to be included above, with a tag for "family"? I'd be okay with that, if you can define your terms and say exactly what you mean by it; I'm sure we could come up with something. Mathglot (talk) 05:56, 19 February 2026 (UTC)

I meant primarily that <!-- uw:{level}:{family} --> where, for example, the {family} would be vand for any of the vandalism warning templates ({{uw-vand}}, {{User:ClueBot NG/Warnings/Warning}}, {{Huggle/warn-1}}, etc), copyright for any of the copyright warnings, etc. I think it's important to distinguish between the different kinds/families of warnings rather than just the level of the warning, if we're going to rework them. — Naomi Amethyst 09:18, 19 February 2026 (UTC)

We might be talking slightly at cross-purposes here. Maybe we need some terminology so we know we are talking about the same thing. We could call hidden template ids in template code "matching tags" or "genuine tags" when they look like <!-- Template:TemplateName --> and match the template pagename, and "mislabeled" or "non-matching tags" when they are of that format and do not match the pagename. To be clear: I am not suggesting removing any of the genuine tags, so for example {{uw-vandalism1}} would still retain a genuine <!-- Template:uw-vandalism1 --> tag at the end of the template. It is only the mislabeled tags that cause a problem, afaic, but adding an *additional* tag of some new format t.b.d. which does not resemble the format of any hidden tag currently in use sitting next to a genuine tag is fine. What we have now, is templates with two tags, one of which is mislabeled (or only a mislabeled tag) and that is a problem. mislabeled tags turn up in Cirrus searches skewing the counts (and locations) of substed templates, making analysis difficult. mislabeled tags have a special meaning that is not identification of what template it came from; they should be converted to a new tag format that tells the tools what they need to know, without misidentifying the template. Hope this clarifies.

Addressing what I think you were saying: if you wanted to, for example, change the last bit of {{uw-vandalism1}} to read: <!-- Template:uw-vandalism1 --><!-- family=vand; level=1 --> that works for me; no mislabeled tag, so no problem. Only one template could have the first tag (it would be globally unique, just like pagenames are) but any number of templates could share the second tag. Mathglot (talk) 10:27, 19 February 2026 (UTC) updated by Mathglot (talk) 11:11, 21 February 2026 (UTC)

I'm a little lost on part of this. As a coder I completely sympathize with wanting a more thought-out solution rather than something that was implemented organically over time, but is this being brought up in response to broken functionality, or is it more a reaction to what you're perceiving to be unclean coding practices? To put it another way, what is the impact of templates misidentified by their hidden tags? tony 06:11, 19 February 2026 (UTC)

Strictly broken stuff. I'm not going to waste words on an all-volunteer effort for stuff that is working, just because it isn't shiny. The impact I explained somewhere, maybe not here, but in a nutshell it blows tallies out of the water for things like the table at the top of WP:WT; the footnotes explain part of the story but basically tallies of how frequently certain templates are used can be used to compare or order them, and the plan was to hopefully track them over time to see usage and trends, perhaps in time-series graphs. This is still possible for most of them, but for a couple that have had their ids usurped (or the inverse: other non-matching ids inserted) it is a problem. Mathglot (talk) 06:30, 19 February 2026 (UTC)

Gotcha. Putting my antivandalism-tool-developer cap on, this can be done without breaking anything in at least CVPI, AV, and WS (I think Huggle as well) by just replacing the extra <!-- Template:uw-vandalism1 --> with <!-- Template:uw-level1 --> (or, really, <!-- Template:uw-anything# -->). The tools are just looking for that hidden template tag with a number on the end. This would also have the benefit of allowing the tables and statistics you're speaking of to count total aggregate "levels" in the same way they do actual templates.

I understand that sounds hacky (calling it a template in an HTML comment but not actually being a template) but think of it like this: We could, theoretically, create template:uw-level1 with the source <!-- Template:uw-level1 --> and transclude that new template into {{Welcome-unconstructive}} to indicate it is a level 1 warning. That would be more "true" (a template called "uw-level1" would be being used), but it would also probably be needlessly complicated when we could just include the <!-- Template:uw-level1 --> string directly into the template. tony 19:37, 20 February 2026 (UTC)

That's a clever solution, and I can support that. I would just like to request one slight improvement: let's guarantee we don't fall into the same situation as before, by locking out the string Template:uw-level1 from ever becoming a template about something else; i.e., either salting it (not in favor) or making it either a redirect, or a template that implements {{void}}, along with some explanatory comments at the doc page explaining the situation for the mystified. Then I'm fine with just having the hidden template id at the end of the template code without transcluding anything. Mathglot (talk) 20:51, 20 February 2026 (UTC)

On second thought, although that is enough for me to be happy, I'm not sure that's going to completely satisfy Naomi's needs, expressed previously in her comments about a 'family' indicator. I don't totally understand that concept yet or how it might solve the tools problem—maybe I don't need to—but I won't put words into her mouth, so we should wait to hear from her about your proposal.

Another example of this may have turned up fortuitously in the § next section about a different topic: as it happens, template {{uw-fv1}} uses the {{templatesnotice}} template with param |series=uw-unsourced, and I have no idea if param 'series' is like Naomi's conception of the 'family' concept, something related, or something completely different. (Also complicating the situation is {{uw-fv1}}'s use of a counterfeit id.) So maybe uw-fv1 is another thought experiment data point to help us think this through. Mathglot (talk) 01:04, 21 February 2026 (UTC)

Tony, in an effort to figure out the scope of the issue, I plan to create two tracking categories to see where we are with this. I've started with Category:Templates with matching hidden id, and it is populating now. (It only includes always-subst templates, so maybe the category name should change.) The other one will be for non-matching id, but I won't get to that right away, as I'm backlogged on several other things. At least this is a start. Thanks, Mathglot (talk) 11:31, 21 February 2026 (UTC) As of this comment, there were 1,199 templates in the cat. Now, there are 1,217. Mathglot (talk) 20:14, 21 February 2026 (UTC)

The label in column one for the uw-unsourcedN series in the table is this:

Addition of unsourced or improperly cited material

However, if you look at each of the four templates, only {{uw-unsourced3}} and {{uw-unsourced4}} say anything about poor sourcing—{{uw-unsourced1}} and {{uw-unsourced2}} do not. (I haven't done the homework to see if maybe they used to and then it changed over time.)

What got me wondering about this, is that recently I reverted a good-faith edit by a brand new editor who added good content accompanied by two reliable sources very closely related to the topic, but of six assertions in the brief added content, none are borne out by either source. My first thought was to go see if we had a {{uw-failed-verification1}} warning (corresponding to inline tag {{failed verification}}) but apparently not. Second stop was uw-unsourced1. But {{uw-unsourced1}} says,

I noticed that you added or changed content in an article, but you didn't provide a reliable source.

and there isn't anything in the rest of it that would make it suitable for an edit with reliable sources that had to be reverted for verification failure.

In the first instance, we need to decide if the series is supposed to cover only unsourced content, as level 1 and 2 say, or both unsourced and poorly sourced, as 3 and 4 say. Secondly, supposing we decide it covers both, is failed verification either of those? I'm not sure it is; sources are provided, but they are not 'poor' at all; they are totally useless. But they are not 'no sources' either. I think we probably need to have a {{uw-failed-verification1}}–4 series. What do you think? Mathglot (talk) 06:31, 20 February 2026 (UTC)

I've created a first draft of the 'missing' template here:

We already have an inline tag for {{failed verification}} but we don't really have a template destined for a user Talk page to draw a user's attention to such edits and explain what the problem is with their sourced edit, and the wording 'unsourced or poorly sourced' is not going to help them understand what is wrong with it. Hopefully something like this will. Your feedback would be welcome. Thanks, Mathglot (talk) 10:57, 20 February 2026 (UTC)

You may be looking for Template:uw-fv1, which is currently just a single (non-escalating) template. --tony 19:40, 20 February 2026 (UTC)

Thanks, TonySt, for linking that. Damn, and I had really looked *all over*, before giving up and writing the mock-up, both in the WARN list, as well as with several search-bar (Cirrus) searches. That template totally needs a couple of redirects so people can find it. As far as Cirrus search, neither failed, nor fail, nor verification nor verified nor verify will get you to it, and the same applies to browser on-page string searches at WP:WARN which don't find it, either, using any of those five strings. That's pretty annoying, and it's unreasonable to expect most users to think of the string "fv" when searching for it, even if that would be a reasonable redirect. At least the on-page searches should be easily fixable by adding some keywords to the "label" field in the table so no one else has to spin their wheels needlessly looking for it. Thanks again, Mathglot (talk) 20:33, 20 February 2026 (UTC)

Somewhat tangential for this section, but it needs to be mentioned somewhere that {{uw-fv1}} has a few problems regarding hidden tags and documentation:

1. {{uw-fv1}} contains non-matching id <!-- Template:uw-unsourced1 -->
2. it lacks <!-- Template:uw-fv1 -->
3. it documents its multilevel status using {{Templatesnotice|series = uw-unsourced|max = 4|...}} but neither of those params is correct; there is only the one template, and there is already a {{uw-unsourced1}}–4 series with that name.
4. if it is intended to be a single-level warning, the pagename should be {{uw-fv}}, not {{uw-fv1}} (although the latter can redirect)

Fixing 1–3 should be uncontroversial, I believe. Not so sure about #4, because I believe this template should be multi-level (as the doc and the pagename already imply) but that should probably be discussed. I may mock up levels 2, 3, and 4 so there will at least be something to look at in discussion. Mathglot (talk) 22:15, 20 February 2026 (UTC)

I am just not happy with the wording at {{uw-fv1}} as I don't think it will quite do the trick for the use cases I have in mind. For example, I have substed User:Mathglot/sandbox/Templates/Uw-failed-verification onto a User talk page (diff) which I think does a better job of it. I particularly like the use of the expression, directly supports twice in the template, once in boldface, echoing the six repetitions of the term at WP:Verifiability. I think the {{uw-fv1}} template should be updated to use this wording.

Btw, regarding findability: looks like others have experienced just as much trouble as I did trying to find it, or at least, there must be some reason why there are only 131 uses in over two years (compare 342 uses of {{uw-unsourced4}} in 7 weeks; 5,352 in 2 years). I have now added 3 redirects which will hopefully make it easier to find. Fixing the hidden id will make it easier for those using the search bar. Mathglot (talk) 01:37, 21 February 2026 (UTC)

Fixed points 1 & 2 above. Mathglot (talk) 01:45, 21 February 2026 (UTC)

I'm a mentor in the Wikipedia WP:Mentorship program, and since I first wrote, I have had two cases where I saw the need for the {{uw-fv1}} template, and can easily see the need for making it multi-level, as the document already claims it is, and the level-1 title implies. Maybe that was the original intent and they ran out of steam, or who knows, but I think we should complete the series through level 4. Mathglot (talk) 03:39, 28 February 2026 (UTC)

I went ahead and created a 2-level (i.e., {{uw-fv2}} ). Levels 3 and 4 escalate to {{uw-unsourced3}} and 4. Mathglot (talk) 06:01, 28 February 2026 (UTC)

I'm doing some work on astronomy units formatting templates. I would need to use {{So}} as a shorthand for {{Solar constant}}, which according to the IAU standard should be written as S_☉, similar to how we already have {{Lo}} as a shortcut for {{Solar luminosity}} L_☉. {{so}} currently redirects here, but it's not the only shorthand available {{soa}}, {{soab}} are also available. @Primefac notified me that since it's subst-only it's impossible to tell how many are using this shorthand, and I want to check if it's safe for me to hijack {{so}}. Slovborg (talk) 12:56, 22 February 2026 (UTC)

Just noting that the target of the redirect is {{Uw-soablock}} (re: "redirects here" since this is a meta talk page). Primefac (talk) 13:15, 22 February 2026 (UTC)

@Slovborg: Template:soTemplate:so has been a redirect since 2008, so you will need to go through WP:RFD if you want to reporpose it. --Redrose64 🌹 (talk) 18:40, 23 February 2026 (UTC)

Fwiw, {{Uw-soablock}} has had 32 substed placements in 2026, and 42,546 total. That doesn't imply anything about the redirect, of course, but I think it's possible to get that info from the logs upon request.

In the meantime, while you're waiting for that one, I've created redirect {{S0}} (ess, zero) for you, and you can use that for the time being. Or forever, if you want. Cheers, Mathglot (talk) 06:48, 24 February 2026 (UTC)

Add ping: User:Slovborg. Mathglot (talk) 09:44, 26 February 2026 (UTC)

Thanks for the heads up. Is there anywhere else it would be good to notify before I try with WP:RFD? Slovborg (talk) 21:57, 26 February 2026 (UTC)

Slovborg, well, I see at least six editors using soablock in 2026, so I suppose you could poll them by pinging them here to get a sense what kind of reaction you might get in an WP:RFD before making a formal request. Otoh, that might duplicate effort unnecessarily, and the audience for an Rfd might be different.

The talk page for {{uw-soablock}} is 'Wikipedia talk:Template index/User talk namespace', i.e., this page (as it is for all user warning templates), so this is clearly a place you'd want to place a feedback notice if you do start an Rfd. The creator of the Template:SoTemplate:So redirect, User:MCB, stopped editing in August 2025, but probably they should be notified too, just in case. Other than that, nothing comes to mind. It's up to you how to proceed, but if I were you, I'd probably just head over to Rfd at this point, and add appropriate notification pointing to it at other relevant locations. Mathglot (talk) 22:34, 26 February 2026 (UTC)

Hi there,

A new template should be created that encourages users to seek consensus on a talk page when their changes have been reverted for this reason. References should be made to WP:CONSENSUS in this template to tell users why this type of collaboration is so important. This seems like a very obvious template to create and could be very useful. Currently the ways we do this are to notify users via an edit summary or by sending them a message on their talk page, but these methods are not standardised or graphical unlike warning templates, standardised language could make Wikipedians' jobs a lot easier when warning users this way by allowing experienced Wikipedians to notify a user who has not gained consensus with a few clicks rather than by typing out a whole message many times.

Thanks for your consideration,

Qwerty123M (talk) 01:17, 27 February 2026 (UTC)

It's unclear to me under what circumstances you'd intend for an editor to use this template. Is it simply, "your edits were reverted, please seek consensus before repeating them"? If so, I usually have more specific reasons for reverting someone's edits (e.g. unsourced, edit-warring, WP:FILMPLOT, WP:IPCV), and usually there's a template that covers my reasoning or it's specific enough that no template would cover it. If you could write up sample text, it might help other editors to understand what your intended use case for this proposed template is. DonIago (talk) 02:42, 27 February 2026 (UTC)

Qwerty123M, I agree with Doniago, here. Imho, all warning templates should be tied to a violation of a specific Wikipedia policy or guideline. If they are not, that opens the door to abuse: one editor decides to place a warning template on the Talk page of another editor, not tied to any P&G violation, but just tied to—what exactly? If you are talking about WP:Edit warring, we already have {{Uw-ew}} for that. If it's not about edit-warring, then what?

We generally don't have templates about the *remedy* (i.e., reverting an edit, deleting a page, etc.) but about the *violation*. This is a good thing, imho, because it is only fair to point out to a user exactly what they did wrong, along with a link to the governing policy or guideline, so they can read it and know how to avoid it in the future. How does your proposed template fit in with this; is it related to the violation of some rule? — Preceding unsigned comment added by Mathglot (talk • contribs)

I feel like this text in this is too long; people with short attention spans will not read it properly. This has bugged me for a while; I post these notices and the receivers clearly don't read it grapesurgeon (talk) 16:59, 3 March 2026 (UTC)

grapesurgeon, It is a strong, policy-based warning template, and imho it needs to lay out exactly what the issue is, and exactly what the user has to do to remain in good standing, as violations will likely get them blocked or indeffed. Do you have any evidence that they are not being read? I don't think the template is too long, but if you would like to suggest what text to remove, or entirely new wording for the template, this is the place. Thanks, Mathglot (talk) 02:23, 14 March 2026 (UTC)

This edit request to Template:Uw-uhblock-double has been answered. Set the |answered= parameter to no to reactivate your request.

Change the unblock template used on this page to "unblock-un" to match with other username block templates as the user is expected to change their username. --Prothe1st (leave me a message)-- 19:21, 3 March 2026 (UTC)

 Done * Pppery * _{it has begun...} 03:07, 5 March 2026 (UTC)

Looking at e.g. {{uw-attempt4im}}, it seems to imply that if you get around the edit filter you don't get blocked. I think it should besimiplified to just "if you attempt to disrupt Wikipedia again", losing the "and trigger the edit filter" clause. Other wording suggestions are welcome. lp0 on fire () 19:42, 8 March 2026 (UTC)

This edit request to Template:Uw-vandalism1 has been answered. Set the |answered= parameter to no to reactivate your request.

Add this right before the end of the message in the case the user who made the edit wants to discuss it. Gommeh ^{(talk! sign!)} 19:59, 11 March 2026 (UTC)

 Not done for now: please establish a consensus for this alteration before posting an edit request. —⁠andrybak (talk) 20:52, 11 March 2026 (UTC)