Wikipedia talk:Manual of Style/Dates and numbers

External image
Kliban: Converting feet to meters

Unofficial anagram of the Manual of Style (First runner-up: A lemony flatus.)

L is for liter

So as this MOS page notes, the lowercase letter "l" is easily confused with the number "1" or the uppercase letter "I" or the pipe symbol "I", depending on the font. The uppercase letter "L" does not have this problem. Consider the difference between these two lines:

100 L (22 imp gal; 26 US gal)

100 L (22 imp gal; 26 US gal)

When I learned the metric system in high school, for this reason we were taught to use only "L" for liter, and I'm told that's also why this standard has been adopted by some chemistry journals. Doing some Wikipedia-wide spellcheck work recently, I did find "l" being used for liter, and it is indeed not all that easy to read whether in the edit window or final display style. To improve clarity for editors and for users of Wikipedia content (which includes but goes beyond this web site and its default settings), I would like to propose changing the MOS guideline so that only "L" is recommended for liters, including constructed units like "mL". I can't think of any upside for using "l", though perhaps someone else can point some out. -- Beland (talk) 04:28, 14 February 2021 (UTC)

Context readily makes this clear. The pronoun "I" is equally easily confused with the number "1" or the lowercase letter "l" or the pipe symbol "I", depending on the font, but does not create any difficulty in the post above. 100 l is equally clear above. Doremo (talk) 04:47, 14 February 2021 (UTC)

I do not find it particularly clear; the reader can be left wondering what unit of measure is being represented by "I", especially in the U.S. where the metric system is not in common use. -- Beland (talk) 20:18, 15 February 2021 (UTC)

Sure it's clear. About as clear as mud. Dondervogel 2 (talk) 21:39, 15 February 2021 (UTC)

• Gosh, it seems impossible to imagine that this has been discussed before now, so that it would be worth checking the archives to see what arguments have been presented already, thereby saving repetition and waste of time. Nonetheless, hoping against hope, let's see ... Amazing! It has been discussed before! – WT:Manual_of_Style/Dates_and_numbers/Archive_157#Decide_on_“l”_vs_“L”_for_litre_(liter). EEng 04:50, 14 February 2021 (UTC)

  Just because it's been discussed before does not invalidate the argument presented. MOSNUM can, if it chooses, adopt a uniform symbol for use across Wikipedia, and a single uniform symbol makes articles easier to follow. Dondervogel 2 (talk) 09:53, 14 February 2021 (UTC)

  Sure, but it's a mark of courtesy (one might even say civility) to respect the time of one's fellow editors by reviewing past discussions first. "Ell can be confused with other stuff, and L can't, so always use L" has been suggested already, with much discussion, so if one wants to move things forward one should build on what has come before – not drag everyone back to the starting line. As the great J.S. Mill put it

  No one's idea of excellence in conduct is that people should do absolutely nothing but copy one another ... On the other hand, it would be absurd to pretend that ... nothing whatever had been known in the world before they came into it; as if experience had as yet done nothing towards showing that one mode of existence, or of conduct, is preferable to another.

  Of course, he wasn't talking about MOS discussion, but the point remains. EEng 14:28, 14 February 2021 (UTC)

Thank you for the link to the previous discussion; reading it was very helpful. I'd be happy to discuss the general problem of connecting current policy to previous discussions if we can do that in a collegial and non-snarky way.

In this case, it looks like the previous discussion got derailed and didn't come to any particular orderly conclusion, though it was a good brainstorm. Based on the preferences expressed, I think there were two options likely to get supermajority consensus support in an RFC. Before jumping into that, why don't I run those past people here:

• All cases - To avoid confusion with "1" and "I" and "|", "L" should be used in all cases, whether isolated like "L" or with a prefix like "mL".
• Limited cases - To avoid confusion with "1" and "I" and "|", "L" is prefered when used without a prefix ("L"), or with a prefix ("mL") if "L" is used elsewhere in the same article ("1000 mL is 1 L", "5 mL/L"). Otherwise, lowercase may be used with a prefix ("ml").

I was thinking of asking folks if they support neither, one, or both of these, and if they have a preference between them. Does that make sense? -- Beland (talk) 00:03, 19 February 2021 (UTC)

Absolutely not. When I taught in a secondary school we explained the the litre was a derived SI unit, it was a way of expressing 1 dm³. It is not derived from a persons name, like Joule or Pascal so must be written like metre, tonnes, or gram with a lowercase initial letter. It is not the Wikipedia way to take a POV, and canvas to have it established through polling users, and on the strength of that make a global change. The Wikipedia way is to reference an authoritative source. Nearest at hand was the Style Guide. THe Economist. 2005. p. 200. ISBN 1861979169. Petrol and all liquid capacity is measured in litres- and bottled beer in UK in 500ml bottles and is not legal unless marked that way.

Delving deeper, I find this explanation.

The litre, and the symbol lower-case l, were adopted by the CIPM in 1879 (PV, 1879, 41). The alternative symbol, capital L, was adopted by the 16th CGPM (1979, Resolution 6; CR, 101 and Metrologia, 1980, 16, 56-57) in order to avoid the risk of confusion between the letter l (el) and the numeral 1 (one).

"The International System of Units (SI)" (PDF). www.bipm.org. 2006. p. 124. Retrieved 19 February 2021.

So you youngsters may believe that uc L is an acceptable alternative, but anyone I have taught won't accept it, or us wrinklies. If the lc l could be confused, then write it out in full or use cubic decimeters instead. Please revisit this in 2161 but keep on smiling. ClemRutter (talk) 01:48, 19 February 2021 (UTC)

As you stated, "The alternative symbol, capital L, was adopted by the 16th CGPM (1979, Resolution 6; CR, 101 and Metrologia, 1980, 16, 56-57) in order to avoid the risk of confusion between the letter l (el) and the numeral 1 (one)." Given that it is after 1979, using "L" is a perfectly acceptable alternative. Thanks for finding the loophole. BilCat (talk) 02:11, 19 February 2021 (UTC)

• Regulars here will be unsurprised at my interjecting, at this point, the criteria given at Wikipedia:If MOS doesn't need a rule on something, then it needs to not have a rule on that thing, to wit, that something belongs in MOS only if (as a necessary but not sufficient test) either:

1. There is a manifest a priori need for project-wide consistency (e.g. "professional look" issues such as consistent typography, layout, etc. – things which, if inconsistent, would be significantly distracting, annoying, or confusing to many readers); or
2. Editor time has been, and continues to be, spent litigating the same issue over and over on numerous articles, either:
   1. with generally the same result (so we might as well just memorialize that result, and save all the future arguing), or
   2. with different results in different cases, but with reason to believe the differences are arbitrary, and not worth all the arguing – a final decision on one arbitrary choice, though an intrusion on the general principle that decisions on each article should be made on the Talk page of that article, is worth making in light of the large amount of editor time saved.

So if I may, before we proceed I'd like to ask for evidence supporting either (1) or (2) i.e. evidence that the merely suggestive (not prescriptive) language we have now ...

The symbol l (lowercase "el") in isolation (i.e. outside forms as ml) is easily mistaken for the digit 1 or the capital letter I ("eye").

is inadequate. EEng 05:49, 19 February 2021 (UTC)

For as long as people use ml/L it's inadequate. Mosnum should at least make clear that the same symbol for litre should be used throughout any one article. Dondervogel 2 (talk) 08:57, 19 February 2021 (UTC)

Have you discussed the issue with the editors of that article? EEng 09:21, 19 February 2021 (UTC)

It was discussed by MOSNUM editors in the link you posted, where it was stated that "ml/L" was perfectly acceptable. I will take it up at the article if and only if we agree here that such ridiculous and confusing notation is unacceptable. Dondervogel 2 (talk) 10:02, 19 February 2021 (UTC)

For reasons explained in the essay I linked a bit ago, it's highly useful for discussion to start in the context of actual articles. In my experience useful arguments are often uncovered that way which would otherwise go unthought in abstract discussions here. EEng 10:24, 19 February 2021 (UTC)

I understand the argument. I'm just giving notice that I do not intend to fight the nonsense without consensus here that it IS nonsense. Life is too short Dondervogel 2 (talk) 11:26, 19 February 2021 (UTC)

So you'd rather argue it in the abstract than hear from editors who are likely to be familiar with actual sources and actual usage in the real world. Noted. [Sorry, that last bit was snarky. My apologies.] EEng 02:31, 28 February 2021 (UTC)

Your post lacks your usual sharp reasoning:

1. There is no need for the discussion here to be abstract. It just helps holding it at one central location.

2. You imply I am unfamiliar with sources. I know from my own experience that some use "l" while others use "L". No serious science editor (in the real world) would sanction a mixture. We should not either.

Dondervogel 2 (talk) 07:57, 28 February 2021 (UTC)

The doctors took away all my sharp reasoning for fear I'll harm myself. Look, I'll just say it again. I don't think anyone here would care if you took 30 articles with ml/L and changed them to mL/L. Then if you wait a few days, you might get 5 inquisitive editors of those articles engaging you about it. Maybe all of them will see it your way, and you can go on changing another 50 articles, and so on, with no shots fired. Or maybe one or two object, and even have some reason you haven't heard of or thought of, or a reputable source that really does that. Talk to them about it. Maybe they'll come around. Or not. Then come here, and bring them with you. Those are the people we need in this discussion. EEng 12:41, 28 February 2021 (UTC)

Apologies not really needed (I know from experience you mean well) but gratefully accepted. I'll think about your suggestion. Dondervogel 2 (talk) 12:53, 28 February 2021 (UTC)

So I wasn't really asking for people to give their opinions on the answer to the question, just on how the question should be asked. But given there was support for the "written out" option and it's mentioned as a common solution on Litre#Symbol, I added that. "ℓ" is also mentioned there, so I'm throwing that in as well. So how about this: (Beland (talk) 20:19, 20 February 2021 (UTC))

Options

(EDITED BASED ON SUGGESTIONS BELOW)

The Manual of Style currently allows the use of "litre", "liter", "l", or "L", with a note that says: The symbol l (lowercase "el") in isolation (i.e. outside forms as ml) is easily mistaken for the digit 1 or the capital letter I ("eye").

How would you rank the following options?

• Option A - No change.
• Option B - Always use "L", including with ("mL") and without ("L") prefixes, to reduce confusion.
• Option C - Use "L" without a prefix to reduce confusion. Never mix "l" and "L", with or without prefix, in the same article. For example write mL/L and 1000 mL is 1 L instead of ml/L and 1000 ml is 1 L. E.g. "ml" is OK if "L" does not appear in the same article.
• Option D - Write out "liter" or "litre" to avoid confusion, where "l" would otherwise be used without a prefix. Follow WP:ENGVAR.
• Option E - Use "ℓ" to avoid confusion where "l" would otherwise be used without a prefix.
• Option F - Always use "l", including with ("ml") and without ("l") prefixes.

Discussion

For litres (not millilitres) {{convert}} seems to be used with "L" over "l" only about 56% to 44% in the February 1 database dump, but in free text just looking at well-spaced contexts, it's 6% "l" to 94% "L". Given that the warning has arguably dissuaded some but not all use of "l", whether it is adequate depends on how acceptable you find "l" in various contexts. -- Beland (talk) 20:19, 20 February 2021 (UTC)

• Any option that discourages a mix is OK by me. In particular, either of B and C from this list would be an improvement on what we have now. Option C can be improved further by also discouraging mL/l. Option E would make things worse because only the symbols L and l are accepted by the BIPM. Dondervogel 2 (talk) 00:02, 21 February 2021 (UTC)

  C would not allow "mL/l" because the denominator doesn't have a prefix; it would have to be "mL/L". -- Beland (talk) 03:32, 21 February 2021 (UTC)

• B - Use 'L' always. It is allowed by BIPM and is not mistaken for the number one or lowercase ell. Option A is putting our head in the sand. Option C (mix of ml and 'l') is yet another convoluted rule to remember - why bother? Option D often turns into an edit war if the article doesn't have any other instances of litre/liter, metre/meter, colour/color to base WP:ENGVAR on (I'm thinking of Japanese car articles). Option E is just so damn hard for the majority of editors to enter that it will never work in practice (technically it is just a font choice of the BIPM sanctioned lowercase ell, so it gets a lot of use on supermarket product labels).  Stepho  talk  00:59, 21 February 2021 (UTC)
• Once again, I wasn't asking people to answer the question yet, just making sure the question was phrased coherently before advertising the RFC. -- Beland (talk) 03:32, 21 February 2021 (UTC)

  The questions are fine - they cover the major bases. Let's start the RFC rolling.  Stepho  talk  04:43, 21 February 2021 (UTC)

  The question can be improved: I'd like to see an option like B, but for lower case only (plain l, not that silly curly font); and option C does not rule out mL/l, but IMO it needs to. Dondervogel 2 (talk) 09:34, 21 February 2021 (UTC)

  I added a new option F to address the first; and an option C2 to replace C. Dondervogel 2 (talk) 15:53, 21 February 2021 (UTC)

  C was intended to be the same as C2; I've unified them with more examples. Hopefully I didn't mess up the meaning? -- Beland (talk) 19:46, 21 February 2021 (UTC)

  Yep, you closed the loophole. Dondervogel 2 (talk) 21:00, 21 February 2021 (UTC)

• It seems to me that there is a missing option G, which is to say explicitly that mixing L and ml is OK, as long as it's done consistently in an article. That would be my preference. In the absence of that, I suppose I'd pick A, no change. --Trovatore (talk) 07:48, 28 February 2021 (UTC)

  Plus, by adding G we would open ourselves to the entire spectrum of musical notes–based ABCDEFG wordsmithing, with words like BAGGAGE, CABBAGE, FEEDBAG, and many more possible. (Handy words for traveling and dining, those are.) Or, restricting ourselves to rankings (with each option appearing at most once), people can select from DECAF, CAGED, FAG (sorry), FACED, BEAD, AGED, and (my favorite) EGAD. EEng 21:51, 28 February 2021 (UTC)

Litre abbreviation RFC

Should English Wikipedia stop using "l" to abbreviate litres in some or all cases? -- 04:14, 25 February 2021 (UTC)

The Manual of Style currently allows the use of "litre", "liter", "l", or "L", with a note that says: The symbol l (lowercase "el") in isolation (i.e. outside forms as ml) is easily mistaken for the digit 1 or the capital letter I ("eye").

How would you rank the following options?

• Option A - No change.
• Option B - Always use "L", including with ("mL") and without ("L") prefixes, to reduce confusion.
• Option C - Use "L" without a prefix to reduce confusion. Never mix "l" and "L", with or without prefix, in the same article. For example write mL/L and 1000 mL is 1 L instead of ml/L and 1000 ml is 1 L. E.g. "ml" is OK if "L" does not appear in the same article.
• Option D - Write out "liter" or "litre" to avoid confusion, where "l" would otherwise be used without a prefix. Follow WP:ENGVAR.
• Option E - Use "ℓ" to avoid confusion where "l" would otherwise be used without a prefix.
• Option F - Always use "l", including with ("ml") and without ("l") prefixes.

Previous inconclusive discussions:

• Wikipedia talk:Manual of Style/Dates and numbers/Archive 157#Decide on “l” vs “L” for litre (liter).
• Wikipedia talk:Manual of Style/Dates and numbers#L is for liter

-- Beland (talk) 04:14, 25 February 2021 (UTC)

• Prefer B; C is OK. D is too cumbersome when space is limited, but would be better than A. E and F would make things worse by being non-ASCII and more confusing, respectively.
  • "L" was added in 1979 to the SI system by the CIPM precisely because of these typographical issues. (See Liter#Symbol.) Some major institutions already use "L" exclusively, including NIST and some international scientific journals. CIPM may in the future deprecate "l".
  • Confusion of "l" does happen in practice. Wikipedia's use of a sans-serif font by default makes "l" after a number look like a one, especially if there is no space before the "l". With a space, it can still look like a one after a decimal space or implied multiplication or a typo. "l" is also indistinguishable from "I", which is used to represent electrical current, dynamic action, sound intensity, and moment of inertia.
  • While readers can usually figure out the meaning from context, it impedes Wikipedia's mission to be unclear and inconsistent and create a speedbump puzzle for students and people unfamiliar with the metric system.
  • A blanket rule to use "L" in all cases removes all the guesswork and is easy to enforce in a semi-automated fashion.
  • What's the downside? Some people say "mL" looks weird, but others say "ml" looks weird, and I think it just depends on what you saw more growing up. English Wikpedia readers from all countries currently encounter both in articles without restriction to any particular national variety of English. If Wikipedia consistently uses "L" and "mL", in the long term maybe that will begin to "look normal" to everyone, but much more importantly fewer readers will be confused by "l".
  • This doesn't make the Manual of Style longer; it's actually streamlining by changing a warning we have to think about in each case, to a guideline we just have to implement.

  -- Beland (talk) 04:14, 25 February 2021 (UTC)

• Prefer B. Lowercase l looks too much like a 1 in too many likely font choices. (In the ones I use, they are very different in preview, but I still can't tell I from l from | in preview, and I can't tell l from 1 in the edit source window.) If it's an official part of SI then it's an acceptable choice, and for typographic reasons the lowercase versions are not. —David Eppstein (talk) 05:11, 25 February 2021 (UTC)
• Prefer A, followed by F and D. The context almost always makes this clear, and it can be written out when needed. Doremo (talk) 05:24, 25 February 2021 (UTC)
• I can live with B, C or F. All other options are poor. Between these, I prefer to adopt a single WP-wide symbol for the litre, so my preferred options are B and F, with a preference between these for B for the reasons stated by others. My third choice is C because it deprecates bizarre combinations like “ml/L", thus improving on the present guidance. A is unacceptable because it permits "ml/L"; D is impractical because there are times when a unit symbol is needed (eg, in an equation). I guess that leaves E as just acceptable. On these grounds it becomes a (poor) fourth choice. To summarise

  1. B

  2. F

  3. C

  3. C (or Option A if modified to exclude mixture of l and L in the same article)

  4. E

• Dondervogel 2 (talk) 07:46, 25 February 2021 (UTC)

  I've just realised that Option E would permit all kinds of combinations, including ml/L, making it unacceptable to me. Dondervogel 2 (talk) 13:00, 25 February 2021 (UTC)

  The strikeout falling right on the middle arm of the E is a delicious touch. EEng 22:46, 25 February 2021 (UTC)

  LOL. Unlike I from l, E is (just) distinguishable from E. Dondervogel 2 (talk) 23:22, 25 February 2021 (UTC)

  Per my recent exchange with Peter coxhead, I would find Option A acceptable if it were modified to exclude a mixture of l and L in the same article. This would rank equally alongside Option C. Dondervogel 2 (talk) 09:52, 28 February 2021 (UTC)

• Option A – None of the others are acceptable to me. I'm not confused by 'ml' and 'L' used together (although 'ml' and 'mL' together probably should be deprecated), and I don't expect others to be confused. Besides, if you really want to make your meaning clear, you employ wiki-links or "convert" templates. Dhtwiki (talk) 09:40, 25 February 2021 (UTC)
• Option B as unambiguous. If we must have A, then the existing rubric needs extending so that it says The symbol l (lowercase "el") in isolation (i.e. outside forms as ml) is easily mistaken for the digit 1 or the capital letter I ("eye") and consequently should never be used. --John Maynard Friedman (talk) 11:20, 25 February 2021 (UTC)
• 🯁🯂🯃42 ...and what was the question ? --:GSMC(Chief Mike) Kouklis U.S.NAVY Ret. ⛮🇺🇸 / 🇵🇭🌴^⍨talk 15:23, 25 February 2021 (UTC)
• I'm still waiting for evidence that editors aren't working out this issue (if an issue it is) for themselves on article talk pages, or that such discussions are wasting time that would be saved by adding to the guidance here. Editors actually working on articles where the question arises are the best people to comment on such issues, and there are none such present here, AFAICS. EEng 19:49, 25 February 2021 (UTC)

  Well, if I went ahead and systematically changed all instances of "l" to "L" because I think that's universally clearer, I'm confident a bunch of editors would object (either in general or for the articles they are watching) and ask me to get consensus for such a change first. It would be a bit disruptive and potentially a waste of time to change a bunch of articles only to potentially change them back, for the sole purpose of proving there's a dispute. While a manual of style may be useful for settling disputes between employees (or in this case volunteers), in professional publishing they are also used to ensure visual consistency of arbitrary choices and to document best practices that copy editors should enforce. Here I'm not trying to settle active disputes, I'm trying to increase clarity; but first I need agreement that lack of clarity is actually a problem and that this is an acceptable solution. Option B doesn't "add" guidance in the sense of making the MOS longer; it would actually make it shorter. -- Beland (talk) 09:05, 7 March 2021 (UTC)

  Option B doesn't "add" guidance in the sense of making the MOS longer – No, but it adds guidance in the sense of telling editors what to do instead of letting them decide among themselves on particular articles. EEng 10:34, 7 March 2021 (UTC)

  Precisely, and that is the purpose of mosnum. Dondervogel 2 (talk) 11:19, 7 March 2021 (UTC)

  Where necessary, not for its own sake. EEng 11:51, 7 March 2021 (UTC)

  Indeed, not for its own sake, but for the sake of "clarity and cohesion ...[especially] within an article". Dondervogel 2 (talk) 12:27, 7 March 2021 (UTC)

  If and where guidance is needed to achieve that. EEng 05:47, 13 March 2021 (UTC)

  No need to be hypothetical here. I once tried doing what Beland suggests. I was immediately reverted. Dondervogel 2 (talk) 10:15, 7 March 2021 (UTC)

  And did you engage those editors in a discussion to elicit their reasons for preferring a particular style? That might be useful information to have here in this discussion. EEng 10:34, 7 March 2021 (UTC)

  I believe the editor concerned preferred to follow the sources that use ml (disregarding those that follow mL?), and was willing to sacrifice uniformity to achieve that. There are plenty of editors already here arguing precisely that, so adding one more would not change things. Dondervogel 2 (talk) 11:19, 7 March 2021 (UTC)

  Editors with skin in the game are often more effective advocates. EEng 11:51, 7 March 2021 (UTC)

  That particular editor was one who would resort to bullying, sometimes in alliance with a known sockmaster, to get his way. Not someone we'd want contributing to a civil discourse. Dondervogel 2 (talk) 12:25, 7 March 2021 (UTC)

  OK, well how about the next article editor you engaged on this issue? EEng 20:40, 7 March 2021 (UTC)

  I recall a discussion here. Many years ago. Before the one you mentioned. I wouldn't know how to find it though. How about pinging the editors in the discussion you found? Dondervogel 2 (talk) 20:44, 7 March 2021 (UTC)

  I found this snapshot on 25 July 2008. Does that help us track down the archive? Dondervogel 2 (talk) 21:02, 7 March 2021 (UTC)

• Option A - to leave it alone - is far preferable to all alternatives. Per Litre#Symbol, this is a case where usage genuinely varies between countries. That means, if we have to have a firmer rule, then we need to go down an WP:ENGVAR-type route. If we mandate obscure Unicode characters as in E, we'll be ignored. There will always be cases where a symbol is needed (i.e. where space is limited) so D is inappropriate. The others inappropriately mandate uniformity. Kahastok talk 20:34, 25 February 2021 (UTC)

  FTR, I do not object to Option G, advising against mixing L and l in the same article without requiring one or the other. Per User:Peter coxhead, this is not the same as option C. Kahastok talk 12:04, 7 March 2021 (UTC)

• I'm am still waiting for referenced sources, from the editors that regular use litres to add a little expert guidance. ClemRutter (talk) 20:38, 25 February 2021 (UTC)

  BIPM permits "L" or "l". NIST requires "L". What more do you need to know? Dondervogel 2 (talk) 22:18, 25 February 2021 (UTC)

• Option A Requiring B will never work with the many people used to lower case—this is a collaborative community of volunteers, not a bunch of paid hacks who take orders from on high. Encouraging Engvar wars is not useful. Johnuniq (talk) 23:04, 25 February 2021 (UTC)

  @Johnuniq: Engvar does not permit constructions like "The honor of modelling aluminum behaviour", and for good reason. Option A permits writing ml/L for millitre per litre. Why is mixing OK for the litre but not for Engvar? Dondervogel 2 (talk) 23:16, 25 February 2021 (UTC)

  Show me a page with mL/l and I'll correct it. We are allowed to use common sense. Johnuniq (talk) 23:29, 25 February 2021 (UTC)

  Bigeye tuna

  Murashige and Skoog medium

  Sea toad

  Wastewater discharge standards in Latin America

  Blue cod

  They're not all of the form "ml/L", but they do all mix l and L in some way. I expect there are many more. Dondervogel 2 (talk) 23:53, 25 February 2021 (UTC)

  We do lots of things that some people aren't used to and never write, like straight ' and ". It's not a problem for a second editor to come along and make a contribution MOS-compliant (and there are volunteers willing to do that), as long as the first editor doesn't change it back. I'm not sure if that's what you're saying will happen, or that people just won't bother to write MOS-compliant text up front? -- Beland (talk) 02:03, 26 February 2021 (UTC)

  @Beland: If they were non-compliant I would fix them myself. The problem is they comply. My purpose is to encourage a change to MOSNUM to make them non-compliant. Dondervogel 2 (talk) 09:00, 26 February 2021 (UTC)

  @Dondervogel 2: Sorry, I was replying to Johnuniq's original comment. -- Beland (talk) 00:24, 27 February 2021 (UTC)

  I moved the full list further down, leaving only the original 5 examples. Dondervogel 2 (talk) 09:49, 27 February 2021 (UTC)

• Prefer B+D, ie use 'L', 'litre' or 'liter' but never 'l'. Option A has too much scope for ambiguity. Option B is legal everywhere (even for those countries like mine that were taught to use little ell back in the 1970s - if we can understand colour/color, litre/liter, autumn/fall then this is nothing). Option C is too complicated for many of our editors (which is a sad commentary but true). Option D should be allowed as an alternative to B (WP:ENGVAR works like it has done for years) . Option E is good for reading but the average editor still struggles with entering hyphens, let alone other symbols. Option F has too much scope for ambiguity.  Stepho  talk  07:26, 27 February 2021 (UTC)

• Prefer B+D, per Stepho-wrs. Seems reasonable to me. BilCat (talk) 23:28, 27 February 2021 (UTC)

• Option B (and D as an acceptable alternative) - it leaves no room for confusion. -- Carlobunnie (talk) 00:52, 28 February 2021 (UTC)
• Option A with a requirement to not mix within an article is my preference, as per Johnuniq. Peter coxhead (talk) 08:15, 28 February 2021 (UTC)

  @Peter coxhead: What's the difference between 'Option A with a requirement not to mix' and 'Option C'? Dondervogel 2 (talk) 08:46, 28 February 2021 (UTC)

  @Dondervogel 2: the status quo allows the use of "l" without a prefix so long as it's clear; Option C does not. Text like Its volume is about 6.7 L (1.5 imp gal; 1.8 US gal) is perfectly clear and hence acceptable in my view. Peter coxhead (talk) 09:38, 28 February 2021 (UTC)

  Thank you for explaining. I agree that consistent use of "l" throughout an article is acceptable, which is why I am also OK with option F. Perhaps the option you favour (explicitly ruling out the mixture) should be added to the list. Dondervogel 2 (talk) 09:49, 28 February 2021 (UTC)

• Option A. There is nothing wrong with ml/L. --Trovatore (talk) 19:24, 28 February 2021 (UTC)
• Option B – if "L" is acceptable per SI, and results in less confusion, which seems really clear to me because I frequently struggle with glyphs that look similar, like - (hyphen-minus) − (minus sign) – (en-dash) — (em-dash), then let's standardize on the least confusing acceptable option. —Joeyconnick (talk) 18:44, 6 March 2021 (UTC)
• Option B When comparing "1 l" with "1 L" out of context, the lowercase "l" can lead to confusion with the numeral "1" for certain fonts, and so the capital letter "L" is used. To keep things consistent, adding a prefix, such as "milli", should not change the unit symbol. The SI Brochure (9th edition) shows that both "l" and "L" are accepted. The unit symbols are lowercase letters unless they are derived from a proper name, but an exception for "L" was made by the CGPM in 1979. However, in the United States, the use of "mL" is preferred by NIST and the FDA. It seems like the CIPM's recommendation of using SI units rather than liters may explain their indifference between "l" and "L". Somerandomuser (talk) 03:16, 7 March 2021 (UTC)

  Units and measurements on Wikipedia are very rarely - if ever - provided without context. MOS:UNITNAMES says that the first few instances of measurements in litres in prose should be spelt out in prose. We also mandate that a non-breaking space be used between a number and the unit symbol. Plus, in our default font used by the vast majority of our readers, the "l" and "1" are pretty easily distinguishable. The likelihood of genuine confusion is minimal.

  Meanwhile, upper-case "L" and "mL" are unusual enough to look jarring in significant parts of the English-speaking world. Arguing for NIST is not convincing in areas outside the United States where it is just another foreign national standards body. Kahastok talk 12:02, 7 March 2021 (UTC)

• I've already posted these examples in the other section below, but I would like some opinions on some interesting cases used in medicine. These units can be found in charts as well. The units for measuring human chorionic gonadotrophin (hCG) levels, thyroid-stimulating hormone (TSH) levels, or other hormone levels: mIU/ml or mIU/mL (milli-international units per milliliter). See also IU/l or IU/L and mU/ml or mU/mL. The units for measuring blood mean platelet volume: fl or fL (femtoliters). The "fl" is also used as "fluid" in "fl oz". What should be done for these cases? Somerandomuser (talk) 06:20, 15 March 2021 (UTC)

  Present advice for fluid ounce is to use either "US fl oz" or "imp fl oz" (not "fl oz"). When written like this it's hard to see fl being read as femtolitre. Dondervogel 2 (talk) 08:15, 15 March 2021 (UTC)

• Option B, although I think 'liter' is acceptable too. I just think we should get rid of "l" since it can be confused with I, I guess that would make D acceptable too. Eccekevin (talk) 03:40, 9 March 2021 (UTC)
• Option D - per making Wikipedia more accessible to our visually impaired readers and editors. Screen readers (at least mine) do not recognize the letter L, upper or lowercase, as an abbreviation for liter. Other abbreviations like ml. kg. km. etc. are recognized by a screen reader, but only if the period is included, otherwise it just reads as letters. Isaidnoway (talk) 19:12, 12 March 2021 (UTC)

  I'm sorry, but this makes no sense at all. I'm sighted, and my browser doesn't recognize the letter L, upper or lowercase, as an abbreviation for liter. All I see is the naked letter, and I have to know that it means liter. So you're on exactly the same footing as everyone else. EEng 20:02, 12 March 2021 (UTC)

  I think the point is that "500 l" or "500 ml" exist in real life text. So, if you're in a place where those symbols are used, you'll interpret them easily because you're used to seeing them in that context. The problem for a person with a screen reader is that basically nobody ever refers to "500 litres" in speech as "five hundred ell". Kahastok talk 20:29, 12 March 2021 (UTC)

  A screen reader isn't supposed to be imitating speech; it's supposed to read out what's on the screen. EEng 05:47, 13 March 2021 (UTC)

  Ideally, it should read out the screen in a way that is readily understandable to the user. That is inevitably going to imitate speech. Otherwise, why not just have it spell out all the words? Don't get me wrong, I accept the argument that says that writing out literally every instance of the unit is unworkable in practice. But it's worth bearing in mind that our guidance already says to write out the unit in most situations. Kahastok talk 11:26, 13 March 2021 (UTC)

  Spelling them out in full each time is ok for text but gets unwieldy in tables and infoboxes. This would make the each entry in the {{{engines}}} field in {{infobox automobile}} overflow to 2 lines in many car articles. Stepho  talk  22:29, 12 March 2021 (UTC)

• Option D. I would prefer to see most, if not all, units spelled out. Avoiding any potential confusion is well worth the extra characters it takes to do so. -- Calidum 19:43, 12 March 2021 (UTC)

  Surely you jest. EEng 20:02, 12 March 2021 (UTC)

  How dare you call me Shurley! Dondervogel 2 (talk) 20:09, 12 March 2021 (UTC)

• [1] EEng 20:26, 12 March 2021 (UTC)

• Option B. There's a reason why it was added to the official standards and why most scientific contexts have adopted it exclusively. And this is not some recent change. The 70s were half a century ago. Heck, even the bottle of Coca-Cola sitting next to me uses "500 mL". The use of lowercase for liters is properly considered archaic. The idea of using a script lowercase l is also objectionable on lack of ease of input, and again, the modern standards call for an uppercase L; the cursive lowercase is even more archaic. As for spelling out, it seems good in theory, except the whole ENGVAR thing and the principle of WP:COMMONALITY. Why introduce additional places where there can be spelling issues, especially for a unit so common that even Americans use it daily (see, again, Coca-Cola or any hard liquor, the one area that has been fully metricized). Nah, just keep it simple. Use the modern standard: L for liter. oknazevad (talk) 11:01, 13 March 2021 (UTC)

  Which suggests that you're based in North America or Australia. Coke bottles and other drinks bottles in Britain and Ireland say "500 ml", with a lower case l. This isn't "archaic" any more than it's "archaic" to spell it as "litre". It's just different standards in different places. Kahastok talk 11:26, 13 March 2021 (UTC)

• Option C, which resolves the central matter of avoiding a single ‘l’ whilst not forcing unnecessary article change in the many uses of cl, ml, etc. in lower case, which remains standard usage in much of the world. Minimising the extent of change to resolve the key concern should be a criterion when considering any change to the MoS. I don’t give too much weight to concern about understanding, since if that were the criterion then the MoS would say next to nothing about punctuation. The important thing is that whatever we write is as clear as possible, so that it is understood when cited. Incidentally, if we can turn the clock back, the mistake was using a single letter; it would have been better if litre(s) were lt. MapReader (talk) 06:54, 15 March 2021 (UTC)
• Option B Should be the preferable option.Sea Ane (talk) 11:48, 18 March 2021 (UTC)

WP:NOTVOTE MapReader (talk) 07:46, 20 March 2021 (UTC)

• Of the proposed options, I'd agree most with B, as WP is (within reason) entitled to choose to standardise on formats based on its need to communicate information unambiguously. We're not proposing any innovation or deviation from the relevant international standards; indeed, standardising on one symbol for the litre is perfectly compatible with BIPM guidance, and it is entirely representative of the practice of reputable reference publications – I'd argue more so than allowing a hodgepodge mixture of "l" and "L", which is unnecessary and IMHO unprofessional. I see ENGVAR as irrelevant, since we are discussing international mathematical symbols that do not form part of any natural human language, let alone specifically English, let alone specific dialects of English. Some sort of de facto rule amounting to "use l and L according to the relative frequency of their use" (i.e. the situation we are in without standardising) is needlessly overcomplicated and sacrifices unambiguity to a chimerical notion of neutrality. I do not think that compromise is warranted. Archon 2488 (talk) 11:08, 22 March 2021 (UTC)

  You need an exceptionally narrow definition of "natural language" to argue that the word or symbol for litre is not natural language. People in metric-using countries speak and write them all the time, including dialect-specific features like pronouncing "ml" as "mill". User:GKFX^talk 15:47, 27 March 2021 (UTC)

• Prefer A Per NIST, "both l and L are internationally accepted symbols for the liter, to avoid [the ambiguity] the preferred symbol for use in the United States is L." As such enforcing the capital L would be an Americanism, and the whole point of WP:ENGVAR is that all common forms of English are permissible. Consistency within each article is good, but the aim for consistency is again something covered by ENGVAR (and common sense). I oppose E, F on respectively the NIST guidance cited and the same argument as above that UK English should not be favoured either. User:GKFX^talk 15:47, 27 March 2021 (UTC)

  If the use of capital L was purely a stylistic thing then, yes, it would be an Americanism. But it serves a very real purpose of avoiding a point of ambiguity for all readers and hence is not an Americanism at all. Note: I'm not American and usually argue against Americanisms.  Stepho  talk  16:04, 27 March 2021 (UTC)

  Option A does not require consistency. There are MOS-compliant articles that mix l and L in the same article and others that mix them in the same unit symbol (ml/L). This makes Option A unacceptable to me. I can accept it if it is modified to require consistent use of l or L within an article. Dondervogel 2 (talk) 16:13, 27 March 2021 (UTC)

References

Litre abbreviation examples

Dondervogel 2 provided some interesting examples where l and L are mixed: Bigeye tuna • Murashige and Skoog medium • Sea toad • Wastewater discharge standards in Latin America • Blue cod. I had been confident that mixing was obviously wrong but now I'm very unsure. From the first link, compare "oxygen content is below 1.5 ml/L" with "oxygen content is below 1.5 ml/l" or "oxygen content is below 1.5 mL/L". In the second link, changing the subheadings to "Major salts (macronutrients)/ 1l" gives an awful result but "1,650 mg/l" is correct for many people. On reflection, I would leave these for editors maintaining the articles to manage (although the Blue cod example needs a simple consistency fix). The standards body probably allows l or L for reasons similar to Wikipedia's ENGVAR pragmatism. In medicine, it is essential that doses be unambiguous and I believe L is often used for that. Yet a large number of people trained in other science areas are used to lower case and consistency should give way to the reality of a volunteer community. I'm posting here to invite opinions on these specific examples. Johnuniq (talk) 02:32, 26 February 2021 (UTC)

• BIPM originally preferred "l" because (like the metre (m) and unlike the newton (N)) the litre is not named after a person. When people complained that "l" could be confused with "I" an exception was made to permit "L" for this one unit. Aberrations like "ml/L" were never intended.
• I fixed example two by spelling out "per litre" in full

Dondervogel 2 (talk) 07:29, 26 February 2021 (UTC)

Aberrations like "ml/L" were never intended – How do you know? It seems quite natural to me. EEng 20:31, 27 February 2021 (UTC)

It stands to reason. The CIPM is driven by a desire to remove ambiguity. This is why there is one and only one symbol for 99 % of units (I know of just 2 exceptions). Permitting a mixture would result in some readers thinking that the symbols "l" and "L" represent different units. Who would benefit from that? And if the font in use does not distinguish between "l" and "1", it also would not distinguish between "ml" and "m1".

Having said this, regardless of the CIPM's intentions, using "L" only (Option B) is clearly permitted, as is using "l" only (Option F). These remain my first and second preferences, with C a close third.

Dondervogel 2 (talk) 20:54, 27 February 2021 (UTC)

Sure it stands to reason. And since it's clear that lots of apparently at least some high-quality sources apparently use ml/L, that apparently stands to reason too. Do you have any actual sources for the statement that ml/L was not "intended"? EEng 21:05, 27 February 2021 (UTC)

I know of no sources that state "ml/L" was not intended. I also know of no sources that use that combination, and I'm surprised to learn that you do. What sources are they? Dondervogel 2 (talk) 21:16, 27 February 2021 (UTC)

I was inferring from the discussion that the use we see in articles followed sources, and I should have made that clear; I've modified my post. EEng 02:31, 28 February 2021 (UTC)

It is tempting to observe that any source that used an abomination like that is ipso facto not a reliable source. --John Maynard Friedman (talk) 22:06, 27 February 2021 (UTC)

Some interesting examples would be the units for measuring human chorionic gonadotrophin (hCG) levels or thyroid-stimulating hormone (TSH) levels: milli-international units per milliliter (mIU/ml or mIU/mL). See also IU/l and mU/mL. Somerandomuser (talk) 03:50, 8 March 2021 (UTC)

The blood mean platelet volume is measured in femtoliters: fL or fl. Somerandomuser (talk) 04:10, 8 March 2021 (UTC)

List of examples

Blenders Pride

Examples are listed of articles in which "l" and "L" are mixed. Feel free to add your own examples.

• Aldosterone-to-renin ratio
• Assessment of kidney function
• Beer in Canada
• Bigeye tuna
• Bitburger Brewery
• Blenders Pride
• Blood test
• Blue cod
• Bong Spirit Vodka
• Breastfeeding
• Calvert Extra
• Cuauhtémoc Moctezuma Brewery
• Cider in the United Kingdom
• Cirrhosis
• Citric acid
• Coca-Cola Vanilla
• Container deposit legislation in Australia
• Crush syndrome
• Cup (unit)
• Customs declaration
• Diagnostic peritoneal lavage
• Donkey milk
• Drunk driving law by country
• Effective renal plasma flow
• Eosinophilia
• Estradiol
• Exercise intensity
• Fifth (unit)
• Filtration fraction
• Fluid ounce
• Gestational diabetes
• Glomerular filtration rate
• Glossary of winemaking terms
• Glossary_of_wine_terms
• Glucokinase
• Hemoglobin
• HIV and pregnancy
• Hyponatremia
• Ketamine
• Kingfisher (beer)
• Labatt Brewing Company
• Latent iron deficiency
• Loop diuretic
• Louis_XIII_(cognac)
• Lysogeny broth
• Methanol
• Methylphenidate
• Minimum inhibitory concentration
• Mitsubishi Outlander
• Mother (drink)
• Mountain Dew
• Murashige and Skoog medium
• Olanzapine
• Olde English 800
• PAH clearance
• Pint
• Postpartum physiological changes
• Pregnancy test
• Reference ranges for blood tests
• Renal blood flow
• Royal Stag
• Sea toad
• Shot glass
• Sinfire
• Sprite Zero Sugar
• Statin
• Urge (drink)
• Uric acid
• Urine sodium
• Wastewater discharge standards in Latin America
• Weir formula
• Zinc nitrate

Non-breaking space templates

"Insert non-breaking and thin spaces as named character reference (  or  ), or as templates that generate these ({{nbsp}}, {{thinsp}})"

But these templates are very .. unfortunate, in the Visual Editor:

Regular character references display fine in the Visual Editor:

though I don't know of a way to insert them — Omegatron (talk) 15:56, 9 March 2021 (UTC)

• The sooner you stop using VE, the better. EEng 01:48, 13 March 2021 (UTC)

ISO 8601

Footnote c of MOS:DATE refers to ISO 8601 for the formatting of numeric dates. Since ISO 8601 is not freely available and can only be purchased from the ISO in two volumes each with triple-digit prices (whether in CHF, USD, or Euro), I question whether this is an appropriate standard to require Wikipedia editors to use. Is there some other more free standard we can replace it with? The footnote is only for a minor point (the interpretation of numerical dates) but I think the principle that editors should be able to read the standards we set for them is important. —David Eppstein (talk) 07:09, 12 March 2021 (UTC)

Editors can read the Wikipedia article. What more do they need to know? Dondervogel 2 (talk) 09:28, 12 March 2021 (UTC)

If we're going to refer people to a standard, the Wikipedia article's description of the standard is a poor substitute. It is not a standard and it is not written as a standard. If we're merely going to tell people to use numeric dates only in 4-2-2 format and only for Gregorian dates, we can do that without pointing them anywhere else. —David Eppstein (talk) 18:51, 13 March 2021 (UTC)

WP uses the relatively straight forward parts of ISO 8601. For the purposes of use within WP, the ISO 8601 article covers everything we need to know. Do you know of something in the standard that we use but don't cover in the article ?  Stepho  talk  22:57, 13 March 2021 (UTC)

Stepho-wrs, I'll mention one small point. In articles Wikipedia uses dates in the YYYY-MM-DD format in a somewhat hazy way. The time zone or time scale are usually not explicitly stated, and must be inferred from context. That's perfectly OK and consistent with ISO 8601. But a person who does not have access to ISO 8601 may be wondering if the standard is really that lenient, and have no affordable, reliable, way to check. (By the way, I had access to an older version but not the current version.) Jc3s5h (talk) 23:28, 13 March 2021 (UTC)

I wonder if we are muddling two issues here?

• The MOS sets Wikipedia's style: unlike normal articles, there is no obligation to cite any external authority. So the justification for using any style of date should be to the en.wiki MOS and nowhere else.
• Quite independently, if an article in main space needs to cite an ISO standard, the editor should not have to defend that choice: many articles cite books that are extremely expensive or even unobtainable outside national reference libraries. In these cases, we should assume good faith: it is sufficient that the citation is verifiable, not that every reader has to do so. Certainly it would be better to cite an accessible source rather than an inaccessible one, if that choice is available, but access difficulty is not a good reason not to use the best (most authoritative) citation possible.

--John Maynard Friedman (talk) 23:44, 13 March 2021 (UTC)

We are not talking about what we cite in mainspace as a reference for content. Those need only be published and reliable, not online or easily available. We are talking about what we point editors to when we tell them that dates must be formatted in a certain way. I think that it is out of keeping with the philosophy of the project to use paywalled non-open standards for that. —David Eppstein (talk) 00:35, 14 March 2021 (UTC)

No, we are telling them that it is so because the MOS says it is, like we do with any other style guidance. --John Maynard Friedman (talk) 08:46, 14 March 2021 (UTC)

We do not tell editors to follow ISO 8601. The editor using the MOS is able to format dates entirely correctly without any access to ISO 8601 at all.

It is sometimes useful, however, to explain to the editor why certain rules are in place. This applies particularly in a document that "is best treated with common sense" and where "occasional exceptions may apply". If the editor does not know why the rule is in place, they do not have the information they need to determine if an exception ought to be made.

For YYYY-MM-DD, we add specific limits that do not apply to any other format (i.e. Gregorian calendar, no earlier than 1583). These limits are in place because readers might interpret YYYY-MM-DD dates as conforming to ISO 8601. This reasoning is not self-evident, and we cannot sensibly explain it without mentioning ISO 8601. Kahastok talk 09:29, 14 March 2021 (UTC)

So why must we assume good faith everywhere else when a citation is given, but not here? The citation is amenable to verification at a good reference library if anyone really doubts its veracity. That only leaves ideological objections in the way. --John Maynard Friedman (talk) 11:02, 14 March 2021 (UTC)

I don't disagree, except to point out that it isn't really even a citation. The point of this footnote is to explain a restriction on use of YYYY-MM-DD dates that is not otherwise obvious. We aren't citing ISO 8601, we're mentioning it.

So in that context, David Eppstein's question becomes, can we explain why we disallow YYYY-MM-DD for non-Gregorian dates and dates before 1583, without mentioning any paywalled non-open standards? Well, the reason for the restriction is that the reader might think we're using ISO 8601, even though we are not. It'll be tricky to explain that without mentioning ISO 8601. Kahastok talk 12:08, 14 March 2021 (UTC)

No we can't, nor must we. The evidence is capable of verification, just not from the safety and comfort of our favourite armchairs. It is still a wp:reliable source and that is the only criterion we really must satisfy. --John Maynard Friedman (talk) 17:16, 14 March 2021 (UTC)

Correct; the ISO is indeed more reliable and reputable than such horrific unethical rags as The Times (et hoc genus omne), which have stained this MOS with their cretinous nonsense, but yet perhaps not less unethical, given the use of practices such as paywalling, which are little more than legalised criminality. But FWIW I would note that anyone computer-literate enough to edit WP markup is readily able to negotiate such trivial hurdles as these paywalls, and as (for example) people who have to access academic journals know, in the real world, the inaccessible is hardly so inaccessible. Archon 2488 (talk) 23:01, 19 March 2021 (UTC)

User:Jc3s5h, are we talking standards or double standards here? Are you telling me that a reference such 2021-03-14 requires a time zone and/or timescale but references like 14 March 2021 or March 14, 2021 do not?  Stepho  talk  02:25, 14 March 2021 (UTC)

Indeed. --John Maynard Friedman (talk) 08:46, 14 March 2021 (UTC)

If 2021-03-14 is given as a publication date and the publication is The New York Times, there is no need to mention the time zone. But if an editor was under the misapprehension that we had to give the UTC date of publication, it would require access to expensive publications to reliably learn that ISO 8601 imposes no such requirement. Jc3s5h (talk) 17:34, 14 March 2021 (UTC)

If 2021-03-14 is given as a publication date and the publication is The New York Times, there is no need to mention the time zone No? In any scheme where one believes that the ISO version of the date must have a time zone, then one is likely under a similar misapprehension about all other dates. So, I agree, seems like a double standard. --Izno (talk) 20:54, 14 March 2021 (UTC)

An editor who knows ISO 8601 well enough to know that it can have time zones also knows it well enough to know that time zones are optional and that 2021-03-12 means an unspecified time zone (ie exactly equivalent to 12 March 2021 or March 21, 2021). The naïve editor who doesn't know much about ISO 8601 will also assume that 2021-03-12 is exactly equivalent to 12 March 2021 or March 21, 2021. Therefore, there is no conflict - everybody is happy!  Stepho  talk  21:58, 14 March 2021 (UTC)

Not March 21, 2021, surely. — JohnFromPinckney (talk) 22:13, 14 March 2021 (UTC)

Endianness. EEng 22:49, 14 March 2021 (UTC)

Oops, fat fingers will be the end of me. 2021-03-12 vs 12 March 2021 vs March 12, 2021.  Stepho  talk  10:53, 15 March 2021 (UTC)

---

Could we add long ISO date support? They currently generate errors in citation templates, because those accept only the formats specified here. — kwami (talk) 23:42, 16 March 2021 (UTC)

By long I assume you mean including hours and minutes and possibly time zone. I wouldn't. The reference date is to give the reader a feel for the general time that something happened (eg last month or middle of 2013) and to help relocate references when the website owner shuffles the pages around (linkrot). The exact time of day isn't very important and brings in complications like time zone (some websites make it real hard to find which country they are from because they are pushing for a global presence). So, too hard for little gain.  Stepho  talk  22:38, 17 March 2021 (UTC)

Kwamikagami indicated elsewhere that he is looking for YYYY MMM DD. --Izno (talk) 00:08, 18 March 2021 (UTC)

What do we gain from adding that to the mix? EEng 00:56, 18 March 2021 (UTC)

Agreeing with EEng, YYYY MM DD adds nothing that YYYY-MM-DD doesn't already have.  Stepho  talk  10:24, 18 March 2021 (UTC)

Yes, it does. Example: 2021-03-12 can be interpreted as two different dates to "the naïve editor that knows nothing about ISO 8601: as March 12, 2021 or as 3 December 2021. Using 2021 Mar 12 eliminates any possibility of March 12 being confused for 3 December. Firejuggler86 (talk) 10:24, 31 March 2021 (UTC)

Can you show any example in real life where 2021-03-12 is used to mean 3 December 2021? In other words, where would a reader get such an idea? And even if that worries you, why not use the 3 Mar 2012 format we already have? EEng 11:29, 31 March 2021 (UTC)

I request that kwami strike out his post made at 23:42, 16 March 2021 (UTC) and start over, on the grounds that there is no standard published by the International Organization for Standardization that would write today's date as 2021 Mar 31. (Or, kwami could provide a link to the standard that does suggest such a format.) Jc3s5h (talk) 16:10, 31 March 2021 (UTC)

The reader might well not have seen YYYY-MM-DD either. Many of our readers are not software developers - or even technical specialists - and it's not reasonable to assume an intimate knowledge of technical data formats.

And it's not necessarily obvious without having seen any XXXX-XX-XX format that the first XX is month and the second one the day. Particularly when you come from a country where XX-XX normally means DD-MM. Kahastok talk 17:48, 31 March 2021 (UTC)

Yes, most people use dmy, so if unfamiliar with ymd might very well interpret it as ydm. ISO is of limited use on en.WP. Tony (talk) 13:12, 1 April 2021 (UTC)

Which is why it's best to spell out the month. ISO order is superior when ordering events that span centuries chronologically. — kwami (talk) 18:58, 31 March 2021 (UTC)

ANY DATE EITHER OBEYS SOME RELEVANT ISO STANDARD OR DOES NOT. SPELLING OUT MONTHS, OR USE A THREE-LETTER ABBREVIATION FOR A MONTH, VIOLATES ISO 8604 AND IT IS UNTRUTHFUL TO STATE OTHERWISE! Jc3s5h (talk) 21:50, 31 March 2021 (UTC)

@Jc3s5h, ISO 8604 Plastics – Prepregs – Definitions of terms and symbols for designations ? I assume you mean 8601.

@Firejuggler86, @Kahastok on the plus side, 2021 April 1 is non-ambiguous but offers no benefits over the 1 April 2021 or April 1, 2021 (except avoiding comma issues). On the negative side, it is not used by any country or standard that I know of. 2021-04-01 is non-ambiguous (there is only 1 date order that starts with 4 digits, nobody uses yyyy-dd-mm). If people are used to other date orders and have difficulty, then seeing the many other dates in the table or references that have dates like 2020-12-31 will correct their thinking. 2021-04-01 also takes up less space - hence its use in tables and other similar space constrained uses. Spelling out the month robs it of a major use (space wise), is not used by anybody and has no other benefits not provided by the other formats.  Stepho  talk  23:09, 31 March 2021 (UTC)

The user should not need to do detective work in order to parse a date.

The space saved by preferring "2021-04-01" over "1 Apr 2021" or "Apr 1, 2021" is marginal at most, and does not come close to outweighing the cost in intelligibility. Kahastok talk 07:30, 1 April 2021 (UTC)

Forgive me for not WP:SHOUTING back at you in my reply. If you insist on yelling at us, at least make sure you do it about something relevant to the discussion. ISO 8604 is about plastics, and I see no connection between the writing of month names and that topic. — JohnFromPinckney (talk) 23:14, 31 March 2021 (UTC)