Talk:将

Spacing modifier in {{ja-r}}

@Poketalker I'm not sure why you removed all of the "%" in {{ja-r}}, but can you compare the following output? KevinUp (talk) 14:35, 24 August 2019 (UTC)

1. 将棋(しょうぎ) (shōgi) vs. 将(しょう)棋(ぎ) (shōgi)
2. 武将(ぶしょう) (bushō) vs. 武(ぶ)将(しょう) (bushō)
3. 知将(ちしょう) (chishō) vs. 知(ち)将(しょう) (chishō)

The first one is without "%" while the second one is with "%". Can you observe any differences between the two?
Also, I hope you don't mind the addition of "%" even if it is not required. This is done for uniformity, so that when new compounds are added by other users, it will be easier to check whether the spacing modifier is required or not. KevinUp (talk) 14:35, 24 August 2019 (UTC)

@KevinUp: Erutuon removed the automatic kana dividing feature for {{ja-r}} in this revision. See also this discussion. One difference is when highlighting them, it will yield something like this: 将棋しょうぎ ‎(shōgi‎) vs. 将しょう棋ぎ ‎(shōgi‎) ～ POKéTalker（═◉═） 00:26, 25 August 2019 (UTC)

Pinging @Dine2016, Erutuon, Suzukaze-c: Is the "%" parameter in {{ja-r}} deprecated? KevinUp (talk) 00:42, 25 August 2019 (UTC)

@Poketalker: Seems like the automatic kana dividing feature (which intelligently decides which kana belongs to which kanji) was removed because this feature caused the kana in words such as 昨日(きのう) (kinō) to be wrongly distributed. I don't think the "%" feature (custom division of kana between kanji) was removed at all. KevinUp (talk) 00:55, 25 August 2019 (UTC)

Ping also @Eirikr for comment. KevinUp (talk) 00:55, 25 August 2019 (UTC)

(edit conflict) @KevinUp: Right. I removed the feature by which, for instance, {{ja-r|知将|ちしょう}} was treated as {{ja-r|知%将|ち%しょう}}. Now the percents are required. — Eru·tuon 00:59, 25 August 2019 (UTC)

Thanks for the clarification. So the percents are required now, and not the other way round. KevinUp (talk) 01:04, 25 August 2019 (UTC)

I think this is a step backward. Automated division should be the unmarked state, and jukujikun marked. Wyang (talk) 01:23, 25 August 2019 (UTC)

• FWIW, I do see different display in the examples above. Using up-to-date Chrome on 64-bit Windows 10.

@Wyang, @Erutuon, I agree that automatic would be ideal, but how reliable was it? If we get incorrect results often enough, that misleads users. Meanwhile, not apportioning kana by kanji at all (when not marked manually) is not misleading, and seemingly the safer default behavior. ‑‑ Eiríkr Útlendi │^{Tala við mig} 16:55, 27 August 2019 (UTC)

@Eirikr: It might take some kind of statistical analysis to determine how often the automatic division was correct, or perhaps you can guess, if I explain it. The function in Module:User:Erutuon/ja divides the series of kana into syllables (syllables, not morae) consisting of an optional consonant, long or short vowel, and optional moraic n or sokuon. Module:ja used this division if it was equal to the number of kanji that corresponded to the series of kana. So the automatic division would be useful if kanji more often than not correspond to a full syllable. 18:18, 27 August 2019 (UTC)