jurgen85
榜眼
(I am aware that you have already put a lot of work into Cantonese readings, so I'm not saying to replace those, but to use this as an automatic layer with higher priority than other automatic layers.)
I stumbled upon this interesting Unihan data type added in 2023 with ~10,000 readings from 《商務新字典》, kSMSZD2003Readings, which seems uniquely useful for Pleco, to automatically generate more accurate Cantonese readings. 2003 was a while ago, and 商務新字典 has been revised since, but I still think this should be very useful data.
What's interesting is that it specifies not only Mandarin and Cantonese readings, but their correspondence, e.g. specifying for 強: qiáng=koeng4, qiǎng=koeng5 and jiàng=goeng6 (where PLC says all are goeng6, and CCC says all are koeng4), and it even specifies some obscure readings for other characters like e.g. 龜 qiū=gau1. It also supplies a sometimes overlapping and overwhelming amount of variant readings. All this to say that it's quite comprehensive.
Z-variants would need additional consideration, as the data will be missing from one or more variants.
I stumbled upon this interesting Unihan data type added in 2023 with ~10,000 readings from 《商務新字典》, kSMSZD2003Readings, which seems uniquely useful for Pleco, to automatically generate more accurate Cantonese readings. 2003 was a while ago, and 商務新字典 has been revised since, but I still think this should be very useful data.
What's interesting is that it specifies not only Mandarin and Cantonese readings, but their correspondence, e.g. specifying for 強: qiáng=koeng4, qiǎng=koeng5 and jiàng=goeng6 (where PLC says all are goeng6, and CCC says all are koeng4), and it even specifies some obscure readings for other characters like e.g. 龜 qiū=gau1. It also supplies a sometimes overlapping and overwhelming amount of variant readings. All this to say that it's quite comprehensive.
Z-variants would need additional consideration, as the data will be missing from one or more variants.
Last edited: