UTC 164-A32: Make the changes to section 17.3 of TUS as outlined in L2/20-150 and discussed in section 19 of L2/20-169 (Script Ad Hoc Recommendations)

[lianghai]

UTC action item 164-A32:

• Liang Hai, Norbert Lindenberg, Editorial Committee
• Make the changes to section 17.3 of TUS as outlined in L2/20-150 and discussed in section 19 of L2/20-169 (Script Ad Hoc Recommendations).

[lianghai]

Derived Edcom 14.0 task:

• D32. Liang Hai, Norbert Lindenberg, Ben Yang: Per 164-A032: Make the changes to section 17.3, Balinese as outlined in L2/20-150 and discussed in section 19 of L2/20-169 (Script Ad Hoc Recommendations).

[NorbertLindenberg]

I hear that the actual source of the Unicode Standard is maintained in Adobe FrameMaker.

How is text in scripts such as Balinese represented in the source? Does it use a Unicode-based font? Are OpenType fonts using the Universal Shaping Engine supported? How about AAT fonts? Is it possible to add a stylistic set to the Balinese font used, and use that stylistic set in the source?

If the above doesn’t work, what are the available workarounds? Hacked Latin fonts? SVG images? PNG images?

[NorbertLindenberg]

I’m discussing with Julie, Ken, and Rick what’s possible on the FrameMaker front.

[NorbertLindenberg]

Here are the changes I’d make to section 17.3:

– Replace the paragraph “Behavior of ra” with and the first sentence of the following paragraph with:

Behavior of ra. Unlike most Brahmi-derived scripts, Balinese includes a final consonant -r, U+1B03 balinese sign surang. This sign is derived from repha in the parent script Kawi, which represented a cluster-initial consonant r-, but has been reanalyzed because many syllables end in -r in the Balinese language. However, when transliterating Kawi, the same sign may still be used as repha. As a surang used as final -r in most cases cannot be visually distinguished from a surang used as repha, it is encoded the same way. As shown in Figure 17.2, the sequence <dha, ma, surang> is pronounced as dhamar when read in Balinese, but can be pronounced as dharma when transliterating Kawi. In modern Balinese, the word dharma is written <dha, surang, ma> with surang as the final consonant of the first syllable. When combined with another above-base sign, a surang used as repha may be placed to the left of the other sign rather than to the right; this can be implemented as a font feature.

Because of its relationship to ra, surang should be treated as equivalent to ra for searching and sorting purposes. For transliterated Kawi it may need to be reordered before the cluster-initial consonant.

– Replace figure 17.2 with this:

①	ᬥ +	ᬫ +	◌ᬃ	→	ᬥᬫᬃ	(Kawi with surang as repha)
	dha 1B25	ma 1B2B	surang 1B03		dha-rma
②	ᬥ +	ᬫ +	◌ᬃ	→	ᬥᬫᬃ	(Balinese)
	dha 1B25	ma 1B2B	surang 1B03		dha-mar
③	ᬥ +	◌ᬃ +	ᬫ	→	ᬥᬃᬫ	(Balinese)
	dha 1B25	surang 1B03	ma 1B2B		dhar-ma

[lianghai]

Reviewing your proposed edit now, before forwarding it to Julie and Ken.

[lianghai]

@NorbertLindenberg, the following is my revision:

Behavior of ra. U+1B03 BALINESE SIGN SURANG typically represents a final consonant -r. This sign is derived from the cluster-initial sign r- (also known as repha) of the parent script Kawi, and still represents a repha when transliterating Kawi, but it has been reanalyzed to represent a final -r in the Balinese orthography. As shown in Figure 17.2, the same written form, pronounced as dhamar in the Balinese language, represents dharma in transliterated Kawi. Because a surang used as a final -r cannot be visually distinguished from a surang used as repha, they are encoded in the same way. When combined with another above-base sign, a surang used as repha may be rendered to the left of the other sign rather than to the right.

For searching and sorting purposes, surang should be treated as equivalent to ra. When the processed text is transliterated Kawi, surang also needs to be reordered to precede its orthographic syllable. Two other combining signs …

Highlights:

• Avoid unnecessary linguistic curiosity (“unlike most Brahmi-derived scripts”, “because many syllables end in -r”, “because of its relationship to ra”…).
• Avoid too much cautious language (“may”, “can”, …) when it’s not the key of the problem. My understanding is, the whole identification of a piece of text as “transliterated Kawi” is already subjective, so it’s not that necessary (in the Core Spec’s scope, which isn’t about NLP) to be further careful about if surang occurrences in transliterated Kawi can be safely assumed as repha. Let me know if the caution is actually very necessary.
• Talk from a point of view that how analog texts (“the same written form”) should be represented in Unicode, rather than how Unicode strings (“the sequence <dha, ma, surang>”) are read by human.

As for the figure, I plan to drop the third example, which is good to know but not necessary and potentially confusing. And I plan to change the first parenthesized note to simply say “transliterated Kawi”.

[lianghai]

Because it was too late when I finally proposed the text, this task has been pushed from Unicode 14.0 to 15.0.

[NorbertLindenberg]

To avoid repeating much of the first line in the second, we might change to:

						Balinese	Kawi
①	ᬥ +	ᬫ +	◌ᬃ	→	ᬥᬫᬃ	dha-mar	dha-rma
	dha 1B25	ma 1B2B	surang 1B03
②	ᬥ +	◌ᬃ +	ᬫ	→	ᬥᬃᬫ	dhar-ma
	dha 1B25	surang 1B03	ma 1B2B