Module talk:my-pron

@Wyang, Mahagaja: Moved from Module talk:ko

Burmese tangent

@Wyang: Over at Module_talk:links#Adding_ts=_param we've been talking about |ts= for Burmese, and Wikitiki89, JohnC5, and Victar all argued against using it for Burmese since the MLCTS is sufficient most of the time. I had wanted to to use |ts= for the IPA transcription, but I was willing to settle for one of the two pronunciation-based transcriptions, but the others were unenthusiastic even about that. If we ever were to use a transcription system for Burmese, I'd actually rather use Okell's than ALA-LC, simply because I find Okell's more intuitive: I feel that c hc j are preferable to ky ch gy for /t͡ɕ t͡ɕʰ d͡ʑ/, and indicating the aspirates with hp ht hc hk hs hy is more consistent than hp ht ch hk hs sh, and using the period and colon to mark tone is confusing in running text. (Note that the original version of ALA-LC doesn't actually mark tone at all; I myself was the one who co-opted the period and colon from MLCTS and pressed them into service for Wiktionary's adaptation of ALA-LC.) —Mahāgaja (formerly Angr) · talk 14:26, 8 April 2018 (UTC)

@Mahagaja Either BGN/PCGN or Okell is fine with me. I support the idea of having some kind of transcription alongside the transliteration in Burmese links, including the scenario of having both by default, since the transliteration and transcription may correlate poorly with each other, e.g. the transliteration of hkrangse. makes little sense in the etymology of chinthe. BGN/PCGN has the slight advantage of a reduced complexity (chindhe.) vs. Okell (hciñtheí), but any of them would be useful for Burmese. The same holds true for other SEA languages written in Indic scripts; cf. the practice in Jenner and Pou (1980), A Lexicon of Khmer Morphology. Wyang (talk) 23:54, 8 April 2018 (UTC)

@Mahagaja, Wyang: Considering the negative community reaction, I won't insist on possibly displaying the IPA using |ts=. The minimum target is to get agreement on displaying a more phonetic transliteration, perhaps only applicable to a few languages, such as Burmese or maybe Tibetan, which have multiple transliterations with different approaches. --Anatoli T. ^{(обсудить}/^вклад) 03:28, 9 April 2018 (UTC)

I do actually like the idea of using some kind of transcription scheme in Burmese links. I recommend a system that is compatible as much as makes sense with the transliteration system. --WikiTiki89 13:46, 9 April 2018 (UTC)

My vote is still for Okell. It may be a bit more complex, as Wyang points out above, but it's more systematic and unambiguous than ALA-LC. —Mahāgaja (formerly Angr) · talk 14:25, 9 April 2018 (UTC)

Words with သျှ

@Wyang, Atitarev, Hintha: both this module and Module:my-translit are failing on the cluster သျှ (hsya.), so that the entries ဗြိတိသျှ (bri.ti.hsya.), သျှင် (hsyang), and အသျှင် (a.hsyang) are throwing module errors. MED has nine words beginning with this cluster; apparently it can be pronounced either /ʃ/ or /j/. I have no idea what the MLCTS and ALA-LC transliterations would be; maybe syh? Can the modules be updated to accommodate this cluster? (It's very weird-looking; I never knew that ◌ှ could be used on any consonant other than nasals and approximants.) —Mahāgaja · talk 20:30, 3 April 2019 (UTC)

@Mahagaja, Hintha: Hi, I can't help with that yet but I will try to find out more. We need more on the usage of ျ and ှ and all other symbols. Wyang has retired from Wiktionary, unfortunately but hopefully we can email him and ask questions. @Octahedron80: Do you think you can help? @Mahagaja, could you give some examples of words that are actually in Sealang? I tried the words you supplied but they are not there. --Anatoli T. ^{(обсудить}/^вклад) 00:28, 4 April 2019 (UTC)

@Mahagaja, Hintha: According to Burmese alphabet: သျှ (hsya.) (hsy) and လျှ (hlya.) (hly) are pronounced [ʃ]. --Anatoli T. ^{(обсудить}/^вклад) 01:14, 4 April 2019 (UTC)

Thanks for raising this point. I had tried to force the template to generate /ʃ/ by manually using ရှ, but it did not work. Yes, both generally map to /ʃ/, but both သျှ and လျှ are finicky. Some anomalies: သျှ may be pronounced /j/ (သျှောင်), while လျှ may be pronounced /l̥j/ for a handful of rhymes (လျှပ် = /l̥jaʔ/). --Hintha (talk) 03:24, 4 April 2019 (UTC)

@Mahagaja, Hintha: The new initials need to be added to Module:my-pron, in this format:

[""] = { "", "", "", "", "" } with values for:

["INITIAL"]: IPA, MLCTS, ALA-LC, BGN/PCGN, Okell. --Anatoli T. ^{(обсудить}/^вклад) 04:00, 4 April 2019 (UTC)

@Mahagaja, Hintha: What is going to be for "သျှ"?

["သျှ"] = { "ʃ", "hsy", "???", "sh", "???" } --Anatoli T. ^{(обсудить}/^вклад) 21:44, 4 April 2019 (UTC)

I'd say ["သျှ"] = { "ʃ", "hsy", "syh", "sh", "hy" } as well as ["+သျှ"] = { "j", "hsy", "syh", "y", "y" } for the voiced initial equivalent. —Mahāgaja · talk 08:23, 5 April 2019 (UTC)

@Mahagaja: Done, thank you! Do you think you can update WT:MY TR at some stage? Also, do you know some examples with "+သျှ"? --Anatoli T. ^{(обсудить}/^вклад) 09:22, 5 April 2019 (UTC)

@Atitarev: Yes, Hintha mentions သျှောင် (hsyaung) above as a case where the cluster in question is pronounced /j/. I'll try to update WT:MY TR this weekend. At the moment I'm working on a computer that shows all Burmese characters as little boxes, which is suboptimal for working on Burmese. —Mahāgaja · talk 09:28, 5 April 2019 (UTC)

Editing the module

@Verdy p: Hi, are you able to make a test module in your sandbox first and test it before creating all these errors? --Anatoli T. ^{(обсудить}/^вклад) 07:38, 25 May 2019 (UTC)

I tested it... it was an error for just a couple of seconds !!! And already corrected immediately before your reverted !!! Verdy p (talk) 07:39, 25 May 2019 (UTC)

There are still bugs in your original code because you mapped the reverse Latin to Myanmar with twice the same key, only one is used in the table, the second one overwrites it. Verdy p (talk) 07:40, 25 May 2019 (UTC)

@Verdy p: The errors are everywhere, in every Burmese entry! --Anatoli T. ^{(обсудить}/^вклад) 07:41, 25 May 2019 (UTC)

They were already there, and that's what I was looking for... I've not changed the logic in fact, the result are still the same as before. I want to clarify the code and make tables more "visible". In fact the "errors" are those from the reverse (Latin back to Maynmar) which is obviously wrong and forgets many cases. Verdy p (talk) 07:43, 25 May 2019 (UTC)

As well the reverse conversion is only tuned for BGN, not for IPA or other transcriptions, because they are not in tables. Verdy p (talk) 07:45, 25 May 2019 (UTC)

@Verdy p: The module worked fine. You have introduced them. This is a heavily used module. Can you switch to Module:User:Verdy p/my-pron, make your test modules as well? --Anatoli T. ^{(обсудить}/^вклад) 07:49, 25 May 2019 (UTC)

@Verdy p: I have created a playground module for you. Don't edit, if you can't avoid and fix errors. The module is working now. --Anatoli T. ^{(обсудить}/^вклад) 07:54, 25 May 2019 (UTC)

Well you've got errors, and only two seconds of glitch from me??? You say it is working but it was not already.

Apparently the "gsub" cause problems (not UTF-8???) in your test page. but not seen elsewhere. Verdy p (talk) 08:01, 25 May 2019 (UTC)