Module talk:bn-translit/testcases

From Wiktionary, the free dictionary
Latest comment: 8 years ago by DerekWinters in topic Outstanding failed cases
Jump to navigation Jump to search

ośobbho —Aryamanarora (मुझसे बात करो) 00:23, 11 January 2016 (UTC)Reply

Arbitrary header

[edit]

Thanks @Kc kennylau! The transliterations of Jônô gônô mônô that you have added do not comply with WT:BN TR - I'll fix them now. —Aryamanarora (मुझसे बात करो) 15:43, 22 February 2016 (UTC)Reply
Also, that isn't Bengali - it's Sanskrit written in the Bengali script, so no inherent vowels are dropped. I think it should be removed... —Aryamanarora (मुझसे बात करो) 15:46, 22 February 2016 (UTC)Reply

@Aryamanarora: Please do replace it with a better testcase. --kc_kennylau (talk) 15:51, 22 February 2016 (UTC)Reply

Outstanding failed cases

[edit]

@Kc kennylau Hi Kenny. Do you think you can fix any of those? --Anatoli T. (обсудить/вклад) 07:45, 7 August 2016 (UTC)Reply

@Atitarev Sorry for the late reply; all fixed. --kc_kennylau (talk) 00:39, 16 August 2016 (UTC)Reply
@Kc kennylau, @Atitarev, @Aryamanarora I added some more, but mostly to show that I think Bengali, especially due to its very high level of Sanskrit borrowings, is not a language that can be transliterated so easily. Its schwa ô /ɔ/ is not regularly deleted, and for a significant amount of words the pattern must be learned from a native speaker. This comes from my experience as a current Bengali learner. DerekWinters (talk) 03:13, 16 August 2016 (UTC)Reply
@DerekWinters I believe that the remaining 6 words that are borrowed from Sanskrit have no inherent vowels deleted... just make sure the editors know that the Sanskrit borrowings are not in the scheme of this module. --kc_kennylau (talk) 06:23, 16 August 2016 (UTC)Reply
The included test cases should only be typical, which follow some rules, which need to be implemented. If they are irregular exceptions, they need to be handled manually, no test cases required. --Anatoli T. (обсудить/вклад) 06:45, 16 August 2016 (UTC)Reply
@Kc kennylau They do though. The last vowel on জল is deleted, as is for the last vowel in পশ্চিম. And higher up, সর্বনাম has two deleted and রচনা has one in the middle deleted. These examples I gave are Sanskrit borrowings. Sanskrit borrowings in other North Indian languages have regular schwa deletion patterns that align with the rest of the language. Yet, due to the nature of Bengali's schwa (/ɔ/) it is is present in some words, like গাল (galô) without a way of explicitly noting it, but not in words like কাঁঠাল (kãṭhal), both of which are native words. DerekWinters (talk) 14:22, 16 August 2016 (UTC)Reply
And also, due to the Bengali Rennaissance, there are arguably more Sanskrit loanwords in daily usage than any other North Indian language, so I would be loathe to have that category be excepted.
@Atitarev I feel that there are so many such exceptions that they are in fact typical. These are fairly standard words and there are several words that will and will not delete the final schwa, making for either an extremely large exception category or no feasible transliteration system. DerekWinters (talk) 14:27, 16 August 2016 (UTC)Reply