Template talk:FAC-instructions

Requested move 27 May 2015

The following is a closed discussion of a requested move. Please do not modify it. Subsequent comments should be made in a new section on the talk page. Editors desiring to contest the closing decision should consider a move review. No further edits should be made to this section.

The result of the move request was: no consensus. Number 5 7 18:43, 15 June 2015 (UTC)[reply]

Template:FAC-instructions → Wikipedia:Featured article candidates/instructions – There's no reason for this to be in the template namespace. --Relisted. George Ho (talk) 19:44, 3 June 2015 (UTC) — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 23:09, 27 May 2015 (UTC)[reply]

Support, wiktionary:fac does not mention "Featured article candidates". Greg Kaye 18:39, 28 May 2015 (UTC)[reply]
Oppose - I see no valid reason. This template has served us well for years, "if it ain't broke don't try to fix it". Graham Beards (talk) 19:55, 28 May 2015 (UTC)[reply]
Comment I kind of fail to see the point. This is a long-standing page. What does it matter if it's in the template namespace or not? Ed ^{[talk] [majestic titan]} 02:24, 29 May 2015 (UTC)[reply]
- Oppose per GermanJoe. Ed ^{[talk] [majestic titan]} 08:21, 6 June 2015 (UTC)[reply]
Oppose - Unproductive use of times. This can be transcluded elsewhere, and has been, so it functions perfectly as a template. — Chris Woodrich (talk) 06:00, 31 May 2015 (UTC)[reply]
Response to all of the above: The purpose of the template namespace is templates, i.e. code that is reused. This is not a template, it's just a one-page transclusion of content to keep the top of that one page "cleaner". It's entirely routine, and sensible, for such things to be subpages of the pages they belong to. We move these things all the time. The above appears to be indistinguishable from WP:IDONTLIKEIT to me. It took more time to write the "unproductive use of time" post than it would have taken to move the page and change a few characters in the page transcluding the pseudo-template. If I'd just moved it manually without bothering with RM, I doubt anyone would have even noticed. These objections are of the I-don't-care-or-want-to-be-bothered character. But it won't be any bother. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 06:00, 1 June 2015 (UTC)[reply]
Comment - As with The ed17, I'm not sure I see the point. Help me understand why this matters? Is it just general curation of the template space or is there some technical reason why it needs to move? Looking in Category:Wikipedia featured content templates, I see that most of the pages supporting Featured Content processes are in the template space, even if they are used only once. I admit I'm not an expert on templates and transclusion, so I'd want to understand any implications of moving things out of template space. I do appreciate your opening an RM rather than moving it unilaterally. --Laser brain (talk) 11:29, 1 June 2015 (UTC)[reply]
- Just general curating, no technical reason. I'm not sure there could be a technical reason; you can transclude stuff from anywhere. As for the FC "templates", there's no reason for those to be in template namespace either. We've been doing maintenance like this since at least when the userboxes were moved out of template space years ago, and probably sooner (before my time). — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 08:35, 3 June 2015 (UTC)[reply]
Comment. SMcCandlish, thanks for the explanation. I'm not going to oppose because I can't see that any harm would result from the move, but I can't quite bring myself to support, because to be honest I don't see any benefit from it either. Some forms of housekeeping leave things in better shape for subsequent editors, but I don't see anything changing for good or ill with this move. I'll let others who feel more strongly about this support or oppose. I'm commenting, rather than just keeping quiet, really just to say that I'd rather not see more such proposed moves in situations like this -- we've spent a few minutes of various productive Wikipedians' time on this (and I'm including SMcCandlish in that), and I don't think the encyclopedia will be better off however this discussion is resolved. Mike Christie (talk - contribs - library) 11:09, 3 June 2015 (UTC)[reply]
- I would argue that having things in their correct place is a benefit to the encyclopedia. Beyond that, cleaning up template space may prevent future non-templates from being created there. I wouldn't worry about how others choose to spend their time, that should be of no concern to anyone else. PC78 (talk) 17:28, 4 June 2015 (UTC)[reply]
  - Right. This is part of many editors' general template namespace cleanup efforts here and there. I'm unaware of anyone spending any serious time on it, much less focusing on it. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 20:52, 5 June 2015 (UTC)[reply]
Support per nom; not a template, just like other transcluded subpages which are not in templatespace. -- 70.51.46.11 (talk) 06:40, 4 June 2015 (UTC)[reply]
Support per nom. No harm in a bit of housekeeping, don't really understand why some people are making a fuss. Don't forget others like Template:FLC-instructions. PC78 (talk) 06:46, 4 June 2015 (UTC)[reply]
Oppose - WP:FAC, WP:FAR and WP:FLC store their old review pages in a subpage-structure directly under the respective main page. Putting instructions together with review pages is ambiguous and could be confusing for a bot (a page named Wikipedia:Featured article candidates/instructions could also be an old FAC-nomination pre-2009). The current setup "as is" in a clearly defined and maintained area has worked for the last 11 years - without any tangible benefit of a move it shouldn't be changed. GermanJoe (talk) 04:05, 6 June 2015 (UTC)[reply]
- Is there any evidence that any bot would actually be confused? The tangible benefit of cleaning up the template space is to have templates in it, not random stuff from all over WP that aren't templates. We have distinct namespaces for a reason. — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼ 01:53, 10 June 2015 (UTC)[reply]
Support because the template namespace doesn't make sense for this and because we already do this on some pages. The example that came to my mind was Wikipedia:Redirects for discussion/Header, which contains the "instructions" for RFD, and I'm sure there are others. Tavix | Talk 16:23, 15 June 2015 (UTC)[reply]

The above discussion is preserved as an archive of a requested move. Please do not modify it. Subsequent comments should be made in a new section on this talk page or in a move review. No further edits should be made to this section.

Update tool link

This edit request has been answered. Set the |answered= or |ans= parameter to no to reactivate your request.

Can the link to vs.aka-online.de/cgi-bin/wppagehiststat.pl (text "significant contributors to the article") be changed to xtools.wmflabs.org/articleinfo? It's more useful in seeing editor contributions to a specific page, among other reasons. Rhinopias (talk) 01:14, 7 March 2018 (UTC)[reply]

Done — Martin (MSGJ · talk) 08:24, 7 March 2018 (UTC)[reply]

Template-protected edit request on 7 March 2018

This edit request has been answered. Set the |answered= or |ans= parameter to no to reactivate your request.

At the top of the template, add {{Redirect|WP:FAC|featured article criteria|WP:FACR}} (or something to the effect). Smtchahal (talk) 10:58, 7 March 2018 (UTC)[reply]

Done... a separate documentation page (/doc) has been created (which can be edited by anyone) and the Redirect hatnote placed there so as not to conflict with the existing hatnote that is actually a part of this template. Thank you for your suggestion! Paine Ellsworth ^put'r there 00:45, 8 March 2018 (UTC)[reply]

Request wording change per discussion at WT:FAC

This edit request has been answered. Set the |answered= or |ans= parameter to no to reactivate your request.

Please change:

If a nominator feels that an Oppose has been addressed, they should say so after the reviewer's signature rather than striking out or splitting up the reviewer's text. Per talk page guidelines, nominators should not cap, alter, strike, break up, or add graphics to comments from other editors; replies are added below the signature on the reviewer's commentary. If a nominator finds that an opposing reviewer is not returning to the nomination page to revisit improvements, this should be noted on the nomination page, with a diff to the reviewer's talk page showing the request to reconsider.

to

If a nominator feels that an Oppose has been addressed, they should say so, either after the reviewer's signature, or by interspersing their responses in the list provided by the reviewer. Per talk page guidelines, nominators should not cap, alter, strike, or add graphics to comments from other editors. If a nominator finds that an opposing reviewer is not returning to the nomination page to revisit improvements, this should be noted on the nomination page, with a diff to the reviewer's talk page showing the request to reconsider.

per the discussion here. Not many people commented but there was no opposition and one explicit agreement, which I think suffices. Mike Christie (talk - contribs - library) 11:14, 3 October 2018 (UTC)[reply]

Reflects long-standing practice. Tks/cheers, Ian Rose (talk) 11:34, 3 October 2018 (UTC)[reply]

Marking as done. Looks like I kind of ec'd with Ian rose :) Galobtter (pingó mió) 11:35, 3 October 2018 (UTC)[reply]

Change wording of proscription of templates

This edit request has been answered. Set the |answered= or |ans= parameter to no to reactivate your request.

Please change:

Please do not use graphics or templates on FAC nomination pages, including graphics such as

Done,

Not done and Example text: they slow down the page load time and lead to errors in the FAC archives.

to

Please do not use graphics or templates on FAC nomination pages. Graphics such as

Done and

Not done slow down the page load time, and complex templates can lead to errors in the FAC archives. The only templates that are acceptable are {{xt}}, !xt, and {{tq}}; templates such as {{green}} that apply colours to text and are used to highlight examples; and {{collapse top}} and {{collapse bottom}}, used to hide offtopic discussions.

per the discussion here. No exact text was agreed to there, but I think this is close to the spirit of the discussion. A named list of acceptable templates seems the easiest way to avoid the recursive expansion problem. Mike Christie (talk - contribs - library) 19:18, 11 November 2018 (UTC)[reply]

Done Laser brain is template protection necessary? Semi-protection would seem to stop the very occasional disruptive edit. Galobtter (pingó mió) 19:26, 11 November 2018 (UTC)[reply]

@Galobtter: I doubt it... reduced to semi. --Laser brain (talk) 15:11, 12 November 2018 (UTC)[reply]

Drop wording suggesting small graphics like "not done" should be avoided

This is kind of a follow-up on Mike's proposal above, but restricted to the issue of the advice about avoiding the use of graphics. Formally: please make the following change:

The reason in a nutshell, is that I doubt the advice about slow page load time is true, if it ever was, even when that advice was added in 2008. Detailed support for this can be found at WT:FAC#Fast-loading check marks and X marks, in particular, this comment, and I should perhaps import a summary of it here, but the gist of it is, that a hundred such check marks on the page would have no significant impact on load time. That's a testable assertion, and perhaps I'll mock something up to test it. Mathglot (talk) 22:08, 11 November 2023 (UTC)[reply]

Mathglot, it's not only loadtime. It chops the FAC archives when template transclusion limits are reached; we discovered this years ago, it's serious, and it's real, and it wipes FACs off the FAC archive page. Check the 2008 archives for when we first discovered it, and the explanation then. SandyGeorgia (Talk) 22:13, 11 November 2023 (UTC)[reply]

I'm sorry if that hasn't been explained adequately at WT:FAC-- institutional memory inactive. SandyGeorgia (Talk) 22:15, 11 November 2023 (UTC)[reply]

Graphics that are not templates only contribute their own size to the PEIS limit, doubled because they are transcluded from the FAC. I don't imagine even the longest Unicode character is more than a few bytes, so these would not cause the problem that things like {{tq}} cause. {{done}} was excluded because it's a template. I think Mathglot is right to say as they do at WT:FAC that the original "page loading time" problem is not going to be affected by using these characters. Personally I'm used to seeing the FAC page without graphic checkmarks and Xes and like it that way, but I'm not sure there's a usability reason to deny them any more. Mike Christie (talk - contribs - library) 22:24, 11 November 2023 (UTC)[reply]

And to clarify for Mathglot: Sandy's talking about pages like the ones linked here, which can contain more FACs than the live FAC page, so the PEIS for FAC has to be well under the PEIS limit or those archive pages will fail to load properly. Mike Christie (talk - contribs - library) 22:28, 11 November 2023 (UTC)[reply]

So it's not that mathglot misunderstood the template limits problem in archives, rather because I'm not reading FAC I missed a piece ? If that's the case ... the instructions are now hidden at FAC (for reasons that make no sense, except that they aren't being followed anyway) and are likely causing new participants to miss them, so making more prominent the templates that shouldn't be used, rather than less, would help per the kind of confusion here. The load time problems are more likely related to the page being overburdened by segmentation, subheads, and length, but the template transclusion problem is separate and real. SandyGeorgia (Talk) 22:36, 11 November 2023 (UTC)[reply]

Yes, they're both real problems, and the template issue is not something we can ever get around except by reducing template use. The page load time isn't really related to the template limits -- if you search WT:FAC for a post of mine to Victoria (search for "PEIS") you'll see a link to an example page that shows that. Page load time is just the raw size of the FACs -- I know you would like to see FACs be shorter and quicker, with material moved to the talk page, and although we disagree on aspects of that there's no question but that that would reduce page load time. Mike Christie (talk - contribs - library) 22:43, 11 November 2023 (UTC)[reply]

OK, just making sure Mathglot had the full picture (who knew I still had this page watchlisted). Not ready yet to go read FAC, and trust your representation. What's not being heard is that some editors would return to review and writing if the page clog/stall would be dealt with, and some editors don't want and never will want to process FACs via nomination viewers or any such script or page that sorts them. They want to review the whole page for the reasons I put on your talk. Bst, SandyGeorgia (Talk) 22:54, 11 November 2023 (UTC)[reply]

You're right, I don't yet have the full picture, and am doing the best I can, and learning as I go. Mathglot (talk) 23:59, 11 November 2023 (UTC)[reply]

Mathglot, I've figured out I need to write a guide; I'm not getting any younger, and institutional memory is being lost. I recall reading recently about "onboarding new Coords" and a "new FAC Coordination page", and wondered what became of the detailed instructions and Coord page I moved to a different Coord a decade ago! And on the ArchiveN thread, I'm fairly sure I'm the only who could have filled you in, given the background, walked you through it, as I helped Gimmetrow build all that. SandyGeorgia (Talk) 01:12, 12 November 2023 (UTC)[reply]

Yeah, the test is not ideal and should be improved, because instead of varying only variable (namely size: X-mark image, vs. X-mark character) it also uses the {{not done}} template in the table, so it's varying no-template vs. yes-template, which could cause issues for reasons Mike has explained. Redoing the 100-row table to include the X-mark image file directly, and not via the templat {{not done}}, would be a better A/B test. That said, I don't see a problem from that one test, so far. Mathglot (talk) 23:50, 11 November 2023 (UTC)[reply]

I have no idea what those tests are doing, but can you test whether the number of section headings have an effect? And, when you run these tests, are you accounting for the effect of transclusions within a tranclsusion? SandyGeorgia (Talk) 23:44, 11 November 2023 (UTC)[reply]

I'm trying to run a test to determine whether adding one hundred "not done" X marks to the latest FAC page slows it down, any. I've only run one test, and it doesn't slow it down. The test could be improved, as noted above, and one test doesn't prove anything. Ideally, one would have to run it numerous times in various conditions. It might still be true that overloading the page with check marks slows it down, but it's not a "slam-dunk" or we would have seen that already, even from one test. As Mike points out, this test doesn't say anything one way or the other about the effects that templates have on the page load time, or whether it cuts off templates lower on the page; that would be a different kind of test. Mathglot (talk) 23:57, 11 November 2023 (UTC)[reply]

Hi Mathglot. A couple of years ago Mike looked at a number of templates and determined that quotation ("tq") templates are the most expensive. Is it possible to run a test with some rows (not necessarily 100) of tq templates containing at least 25 characters? Thanks. Victoria (tk) 00:33, 12 November 2023 (UTC)[reply]

Hi, Victoria. I updated the 100-row table to add a third column, containing 100 {{tq}} templates (rev. 1184687933). Transcluding this in rev. 1184674114 of the FAC test page took 1.582/2.007 seconds, or about 0.1 to 0.3 seconds longer than before, of which 121 milliseconds were due to the 100 added tq templates. Once again, one test is not a proof, and rerunning the same test may well have longer (or shorter) times. It is shown as Test 3 below. Mathglot (talk) 01:04, 12 November 2023 (UTC)[reply]

Yea, I don't know if it will make a difference, but current FACs include WAY longer text strings than that ... and Mike explained years ago how those got multiplied ... test with strings of two long sentences in a tq template? SandyGeorgia (Talk) 01:19, 12 November 2023 (UTC)[reply]

Mathglot ... But it's important to transclude them within a transcluded FAC. When you first added the new blurb to the top of the Philosophy FAC, you got it directly on the FAC page-- not within the transcluded Philosophy FAC (note the edit when I moved it in). The transclusion within a transclusion matters. SandyGeorgia (Talk) 01:15, 12 November 2023 (UTC)[reply]

Sandy, see Test 4 below, with longer {{tq}} remarks in the table. The total time of 2.2 seconds includes 0.28 seconds dealing with the {{tq}} templates. Mathglot (talk) 01:38, 12 November 2023 (UTC)[reply]

Mathglot, this interesting. Thanks for running the tests. It looks to me like you have plenty of lipspans but I think that they (& perhaps the rows of graphics) have to added to the actual text of the review that's transcluded from the article talk page to the FAC page. My understanding from years ago was that the transclusions caused some multiples of expansions, which then causes problems. It's also possible, based on your tests, that these problems no longer exist. If you'd like a live sample, I'm willing to volunteer St John Altarpiece (Memling) to be reviewed - it can have some sort of a "test" added to the title - then we can add as many rows of graphics & of lipspans as needed. And, since it's not perfect people can review to add to the test. We've often found that a single long review can cause problems for the page as a whole. Of course the coords would have to agree. But if you think that would be a better test & can explain to them why it would be needed, then let me know & I'll submit the article for review. Victoria (tk) 03:42, 12 November 2023 (UTC)[reply]

Mike, have a question about the validity of my test 4 for Sandy's test. I used {{lipspan|2}} to create the "long tq remarks" in col. 3 of the transcluded 100-row test table where I'm trying to burden the transcluding mock FAC page with lots of content to see if it slows down page load. Should I subst the lipspans instead, so the table actually gets huge, or is the transclusion of lipspan good enough? Mathglot (talk) 01:50, 12 November 2023 (UTC)[reply]

Tests

I haven't done comparisons yet, but the test page is here, and the data is below:

Test 1. NewPP limit report for the test page, including the 100-row test table:

NewPP limit report
Parsed by mw2421
Cached time: 20231111224702
Cache expiry: 1645981
Reduced expiry: true
Complications: [vary‐revision‐sha1, vary‐revision‐exists, show‐toc]
CPU time usage: 1.245 seconds
Real time usage: 1.610 seconds
Preprocessor visited node count: 8372/1000000
Post‐expand include size: 1086921/2097152 bytes
Template argument size: 48870/2097152 bytes
Highest expansion depth: 32/100
Expensive parser function count: 4/500
Unstrip recursion depth: 0/20
Unstrip post‐expand size: 25400/5000000 bytes
Lua time usage: 0.247/10.000 seconds
Lua memory usage: 20046326/52428800 bytes
Number of Wikibase entities loaded: 0/400
-->
<!--
Transclusion expansion time report (%,ms,calls,template)
100.00%  677.907      1 -total
 17.97%  121.800      1 Wikipedia:Featured_article_candidates/Philosophy/archive1
 16.65%  112.873      3 Template:Blockquote
 14.58%   98.827      1 Template:FAC-instructions
 12.79%   86.707      1 Wikipedia:Featured_article_candidates/Benty_Grange_hanging_bowl/archive1
 11.15%   75.556      1 Template:Lang
  9.53%   64.605      3 Template:Cite_book
  6.52%   44.201      1 Wikipedia:Featured_article_candidates/Eye_(Alexander_McQueen_collection)/archive1
  6.09%   41.273      1 Template:FApages
  5.62%   38.090      2 Template:GBPConvert

Test 2. NewPP limit report for the test page, without the 100-row test table:

NewPP limit report
Parsed by mw2377
Cached time: 20231111232858
Cache expiry: 1643464
Reduced expiry: true
Complications: [vary‐revision‐sha1, vary‐revision‐exists, show‐toc]
CPU time usage: 1.414 seconds
Real time usage: 1.830 seconds
Preprocessor visited node count: 8142/1000000
Post‐expand include size: 1065893/2097152 bytes
Template argument size: 48592/2097152 bytes
Highest expansion depth: 32/100
Expensive parser function count: 4/500
Unstrip recursion depth: 0/20
Unstrip post‐expand size: 25074/5000000 bytes
Lua time usage: 0.308/10.000 seconds
Lua memory usage: 18771751/52428800 bytes
Number of Wikibase entities loaded: 0/400
-->
<!--
Transclusion expansion time report (%,ms,calls,template)
100.00%  803.758      1 -total
 18.59%  149.432      1 Wikipedia:Featured_article_candidates/Philosophy/archive1
 17.26%  138.705      3 Template:Blockquote
 15.36%  123.438      1 Template:FAC-instructions
 12.40%   99.703      1 Wikipedia:Featured_article_candidates/Benty_Grange_hanging_bowl/archive1
 10.68%   85.862      1 Template:Lang
  9.63%   77.402      3 Template:Cite_book
  6.00%   48.198      1 Template:FApages
  5.79%   46.531      1 Wikipedia:Featured_article_candidates/Eye_(Alexander_McQueen_collection)/archive1
  5.30%   42.617      1 Template:Navbar
-->

Test 3. NewPP limit report for the test page, with the 100-row table, and a {{tq}} template in each row:

<!-- 
NewPP limit report
Parsed by mw2408
Cached time: 20231112004705
Cache expiry: 1638778
Reduced expiry: true
Complications: [vary‐revision‐sha1, vary‐revision‐exists, show‐toc]
CPU time usage: 1.582 seconds
Real time usage: 2.007 seconds
Preprocessor visited node count: 13351/1000000
Post‐expand include size: 1125649/2097152 bytes
Template argument size: 56689/2097152 bytes
Highest expansion depth: 32/100
Expensive parser function count: 4/500
Unstrip recursion depth: 0/20
Unstrip post‐expand size: 60050/5000000 bytes
Lua time usage: 0.346/10.000 seconds
Lua memory usage: 17243629/52428800 bytes
Number of Wikibase entities loaded: 0/400
-->
<!--
Transclusion expansion time report (%,ms,calls,template)
100.00%  915.490      1 -total
 14.98%  137.155      1 Wikipedia:Featured_article_candidates/Philosophy/archive1
 14.93%  136.702      1 User:Mathglot/sandbox/Test_pages/FAC_page_load_time_proposal
 13.92%  127.431      3 Template:Blockquote
 13.18%  120.697    100 Template:Tq
 12.24%  112.088      1 Template:FAC-instructions
 10.77%   98.623      1 Wikipedia:Featured_article_candidates/Benty_Grange_hanging_bowl/archive1
  9.46%   86.565    100 Template:Encodefirst
  9.44%   86.410      1 Template:Lang
  7.93%   72.600      3 Template:Cite_book
-->

Test 4. NewPP limit report, with the 100-row table, and a {{tq}} template with a longer remark:

NewPP limit report
Parsed by mw2262
Cached time: 20231112012735
Cache expiry: 1636349
Reduced expiry: true
Complications: [vary‐revision‐sha1, show‐toc]
CPU time usage: 2.206 seconds
Real time usage: 2.649 seconds
Preprocessor visited node count: 14812/1000000
Post‐expand include size: 2074614/2097152 bytes
Template argument size: 312802/2097152 bytes
Highest expansion depth: 32/100
Expensive parser function count: 4/500
Unstrip recursion depth: 0/20
Unstrip post‐expand size: 60050/5000000 bytes
Lua time usage: 0.504/10.000 seconds
Lua memory usage: 15332992/52428800 bytes
Number of Wikibase entities loaded: 0/400
-->
<!--
Transclusion expansion time report (%,ms,calls,template)
100.00% 1285.678      1 -total
 23.62%  303.717      1 User:Mathglot/sandbox/Test_pages/FAC_page_load_time_proposal
 21.90%  281.583    100 Template:Tq
 18.16%  233.425    100 Template:Encodefirst
 14.97%  192.487      1 Wikipedia:Featured_article_candidates/Philosophy/archive1
 14.09%  181.140      3 Template:Blockquote
 10.54%  135.554      1 Template:FAC-instructions
  9.88%  127.058      1 Wikipedia:Featured_article_candidates/Benty_Grange_hanging_bowl/archive1
  9.56%  122.919    100 Template:Trim
  8.66%  111.398      1 Template:Lang

Ironically, the version *without* the 100-row test table containing 100 {{not done}} templates, ran slower (1.414, 1.830 seconds) than the version *with* the test table (1.245, 1.610 seconds). No surprise there, all sorts of factors (notably server load) can affect that. But I'd say that while not a proof, it's one quantum of evidence that adding 100 small icons to the table does not adversely affect page load time. Adding a {{tq}} template to each row appeared to slow it by 0.120697" (one eighth second) for one trial). The test is not ideal, as it fails to vary only one variable: the table should be redone to containg X-mark icons, without the template, to get a better A–B comparison. Mathglot (talk) 23:37, 11 November 2023 (UTC) updated by Mathglot (talk) 01:06, 12 November 2023 (UTC)[reply]