@Suzukaze-c, perhaps it would be easier to deal with if trying to hyphenate particles was abandoned? The Korean module is unable to do so too.--Tibidibi (talk) 06:17, 13 March 2021 (UTC)
- @Tibidibi From my understanding, support for hyphenation (and quotation marks??) was removed from the Korean module because it was interfering with the romanization of jongseong-choseong clusters after a redesign of the module. (Fixing that would be rad.)
- On the contrary, I don't think that is a problem for Middle Korean (aside from the need to strip the hyphens from the hangeul text before it is shown to readers).
- What I'm stumped on is /ɣ/. Can it be romanized automatically without manually marking its location? —Suzukaze-c (talk) 06:27, 13 March 2021 (UTC)
- @Suzukaze-c, the general rule as given by Lee and Ramsey 2011 is that ㅇ should be /ɣ/:
- After ㄹ
- After ㅿ
- After ㅣ (including ㅐ, ㅔ, etc.) but not before /j/ (so not e.g. 머리예 (meli-yey))
- The main issue is that some cases of ㅇ are actually from the loss of /β/, e.g. 글왈 (kulwal) < 글ᄫᅡᆯ (kulWal) was probably not [kɨl.ɣwal] but [kɨl.wal], with the orthography representing an intermediary period in which the vestige of the /β/ blocked the expected surfacing as [kɨ.ɾwal] that we end up having anyways in MdK 글월 (geurwol). While Lee and Ramsey use <G> for this anyways, I don't consider this ideal. This should probably be manually marked.--Tibidibi (talk) 07:01, 13 March 2021 (UTC)
- @Suzukaze-c, the general rule as given by Lee and Ramsey 2011 is that ㅇ should be /ɣ/:
- Cool. I've made it so that manually typing
gbefore choseong ieung produces ⟨G⟩, and currently all of the testcases pass, except for the one where a plain ieung should produce ⟨[ng]⟩, which probably needs manual intervention like ⟨G⟩. —Suzukaze-c (talk) 07:28, 13 March 2021 (UTC)
- Cool. I've made it so that manually typing
Pinging @LoutK. Please treat this testcases page as a wishlist and as a place to stress-test the module. —Suzukaze-c (talk) 07:35, 13 March 2021 (UTC)
Currently the module changes behavior if there are tone marks: it automatically marks low tone and romanizes a jongseong plain ieung in hanja readings as null.
A way to manually mark low tone would probably be useful (as it stands, a jongseong plain ieung in a standalone 平聲 reading would not be removed: 之징→cing), but I'm not sure what the best way is. —Suzukaze-c (talk) 07:59, 13 March 2021 (UTC)
Not ready for usage since the hyphen silliness needs some sort of special attention. {{l|okm|길--헤}} → 길헤 is a bad link. —Fish bowl (talk) 06:51, 21 February 2022 (UTC)
- Identical treatment would be useful for links like
{{l|ko|자유-의 여신상}}instead of{{l|ko|자유의 여신상|자유-의 여신상}}. Can we simply remove hyphens that aren't at the beginning (prefix) or end (suffix) of a link? (@Tibidibi, LoutK) —Fish bowl (talk) 12:19, 6 March 2022 (UTC)- @Fish bowl This would be great. It should be safe to strip them wholesale since only affixes and particles should have links with hyphens. — LoutK (talk) 14:00, 6 March 2022 (UTC)
Done: Word-central hyphens are removed for ko, okm, and ko-ear; and Module:okm-translit is enabled for okm and ko-ear. —Fish bowl (talk) 08:36, 20 April 2022 (UTC)
@Erutuon The module is apparently disgustingly slow, judging from Category:Pages with module errors where pages encounter the 10 second processing limit. Can you look over it? (So much for ease of maintenance...) —Fish bowl (talk) 23:31, 20 April 2022 (UTC)
- Never mind, I found a solution. (????) —Fish bowl (talk) 23:51, 20 April 2022 (UTC)
@Theknightwho I don't know how much is on your plate right now, but would you be interested in applying your expertise to make this module less disgusting? —Fish bowl (talk) 23:12, 1 April 2024 (UTC)