Wikipedia talk:Manual of Style/Superscripts and subscripts: Difference between revisions

Revision as of 07:48, 5 August 2025

Revision as of 08:05, 5 August 2025

Line 32:

:: I'd love to be proven wrong and enlightened, but as I see it currently, I beg to differ. I doubt "modern typesetting" involves using the HTML <nowiki><sub>/<sup></nowiki> tags. Browsers just take the ordinal digits and make them smaller, which is already wrong. The Unicode subscript and superscript numerals are styled differently from the ordinal digits. Fonts that employ old-style numerals by default with descenders and ascenders (and miniature 0,1,2) use the even-sized tabulating numerals for its sub/sups. The forms are then adjusted for clarity and consistency, the strokes are proportionately thicker, serifs are reduced or even absent, and they have their own kerning and layout rules. Dumping all that logic out to the browser's HTML renderer is just silly in 2022. Sensible semantic search results are solvable problems using appropriate [[Unicode equivalence]] and [[text normalisation]] rules in the search engine. I agree with {{user|Marc Lacoste}} that this part of the MOS should be reconsidered. It is inconsistent with how most modern fonts work, produces ugly typography, and conflicts with use of Unicode elsewhere in Wikipedia (e.g. ♭, ♮ and ♯ symbols in music articles). Anyone needing to routinely type all these sorts of characters can use a compose key; it has eleventeen bazillion shortcuts by default in Linux, and there's WinCompose (open source) which gives you the same thing in Windows, instead of horsing around trying to remember loads of obscure 4 digit numbers, and there's bound to be something on a Mac (I have no idea, sorry, good luck with that). For instance, I can type ₂ using the following key strokes: <code><nowiki>[RAlt] [_] [2]</nowiki></code> and similarly, <code><nowiki>[RAlt] [~] [n]</nowiki></code> produces ñ. Anyway, I think in the last several years, there has been a lot of progress and change in the uptake of Unicode both within Wikipedia and everywhere else. — [[User:Jonathanischoice|Jon]] ([[User talk:Jonathanischoice|talk]]) 23:34, 16 March 2022 (UTC)

:::Pick any modern professional word-processing software, HTML-editor, or other professional typesetting editors (such as [[LaTeX]]-based editors), when you select the superscript function, what you get is the regular characters raised and shrinked, not converted to corresponding unicode entities. Unicode entities have long been deprecated on Wikipedia and elsewhere, let's not re-introduce them. As for {{xt|searching for "2" (two) should not find "²" (square), they do not have the same meaning}}, it should, because both mean 2. 2 does not differ in meaning depending on its position. 2 means 2 in 22, 1/2, 3², {{radic|134|2}} or {{nowrap|a₂ {{=}} 1.375}}. &#32;<span style="font-variant:small-caps; whitespace:nowrap;">[[User:Headbomb|Headbomb]] {[[User talk:Headbomb|t]] · [[Special:Contributions/Headbomb|c]] · [[WP:PHYS|p]] · [[WP:WBOOKS|b]]}</span> 00:00, 17 March 2022 (UTC)

::::Word processing isn't the only consideration, and rich text isn't always relevant.

::::There are still plenty of use cases where "plain text" rules, and HTML tags are (rightly) ignored.

::::We're ''still'' encountering ''new'' page summaries and audio page readers that claim such nonsense as that the mass of the sun about 2 tonnes, specifically 1.9889×1030 kg. see [[#Different guidance for non-HTML rendering]]

::::[[User:Martin Kealey|Martin Kealey]] ([[User talk:Martin Kealey|talk]]) 08:05, 5 August 2025 (UTC)

== Add exception to allow Unicode super/subscripts in COinS fields in {{tl|cite xxx}} templates? ==