Preserve source - Thaana & Perso-Arabic consonants

[GokulNC]

Is it possible to preserve source for Dhivehi & Urdu consonants which are of Perso-Arabic origin?

For your reference:

Thaana	Perso-Arabic	Devanagari	Roman
ޙ‎	ح	ह़	h̤
ޜ	ژ	झ़	ž
ޘ‎	ث	थ़ / स़़	ṯ / s̱
ޛ‎	ذ	द़ / ज़़	ḏ / ẕ
ޞ‎	ص	स़	s̤
ޟ‎	ض‎	ध़ / ॹ	ḋ / ż
ޠ‎	ط‎	त़	t̤
ޡ‎	ظ	ध़़ / ॹ़	d̤ / z̤
ޥ	व़	w

Please note that:

• Some Devanagari consonants are approximately made-up to enable one-to-one map (some with multiple nuqtas, though it may not be separately visible in default font)
• Some phonemes are transcribed differently in Arabic (as in literary Dhivehi) and Persian (as in Urdu)

I might have missed something, please feel free to implement it the way you see it right. Thanks!

---

Edit: Suggestions for Thaana--

• For diphthongs (like ޤައުމަކަށް -> qa:umakaṣ), please use the U+A789 colon instead of the regular colon.
• For word final /އް‎/, please use U+02BE instead of ʔ

[virtualvinodh]

@GokulNC Thanks for this!

Isn't there an established way to uniquely (as in one-to-one) transcribe Urdu to Devanagari (that doesn't use multiple Nuktas)? I am sure somebody should have done standardized mapping.

V

[GokulNC]

So far I haven't come across any standardizations/proposals to uniquely map all Perso-Arabic consonants to Devanagari. I think this is mainly because there is no necessity to do that. Only when it comes to lossless transliteration (for computational purposes), it might make sense to have such a map.

For example, there are 4 graphemes in Persian & Urdu (ز , ذ , ض , ظ) for the same phonetic (IPA /z/) , for which the Devanagari "ज़" is sufficient. Similarly for स, त, ह.

Some of the above consonants in table (ह़/ح, झ़/ژ, थ़/ث, स़/ص, त़/ط‎) comply with the latest ISO 15919 Indic Transliteration standard.

---

There seems to have been some established mapping for Hindustani which I came across in the book: The Hindee-Roman Orthoepigraphical Ultimatum (See pages cxxxvi & cxxxvii)

This features the double-nuqta स for ث (स़़ - Shows 2 nuqtas in some fonts) I am not sure about the other Nagari characters listed for the unique mapping (does not seem to be in modern Devanagari)

[GokulNC]

For Mahl Dhivehi (of Lakshadweep), I found this mapping for Devanagari: https://dv.wikipedia.org/wiki/ވިކިޕީޑިއާ:How_to_read_Devanagari_script (See additional consonants section in the image)

Not sure of its origin and authenticity.

[virtualvinodh]

Fixed. Will be available in the next update.

I had to make up my own mapping a bit but this should work.

You have to enable preserve script to access the extra nuktated consonants.

V