Odp: Re: RE: Why was L2/25-061 provisionally assigned?

Sat, 05 Jul 2025 22:20:02 -0700 

In  www.unicode.org UTC 183 Minutes (unicode.org) , in "D.1 1.7 Compound 
IPA stress mark" it was mentioned that " not everything that 
graphically consists of pieces should be encoded as a combining sequence ", 
which is a reasonable argument (even though the damage is already done with 
many combining sequences in Unicode that I would consider inappropriate), and I 
no longer question the provisional assignment.   What I still don't get is 
the claim that " combining marks on modifier letters generally become 
smaller and attach to the modifier letter ". This implies not only that the 
'superscript' or 'subscript' property is inherited by all 
diacritical marks above or below, but also that each use of combining 
implicitly creates a new nesting level relative to the level of the base 
character instead of relative to the base document. This seems to be bizarre to 
me and I've never seen a precedent of that occuring.   Dnia 18 kwietnia 
2025 22:08 [email protected] via Unicode <[email protected]> 
napisał(a):  I don't use Unicode normalization myself and I'm not 
really in a position to make decisions on canonical/normalized representations, 
but since stability policies prevent one of the representations from 
normalizing to the other, if the usage of combining characters were 
standardized to compose to ˈ̩ or ˌ̍ for phonetic usage, most likely one of the 
representations would be recommended whereas the other would go to the 'Do 
Not Emit' list or something.   The use of anchor points is quite font 
technology specific and therefore off topic, though the point still stands with 
any method of maintaining systematic typographical alignment of all combining 
character combinations.   Dnia 18 kwietnia 2025 21:38 Charlotte Eiffel Lilith 
Buff via Unicode <[email protected]> napisał(a):  Which would be 
the canonical representation, spacing low line + combining line above or 
spacing high line + combining line below? Any font that bothered to define 
proper anchor points for diacritics on modifier symbols would display both 
sequences identically.   Am Do., 17. Apr. 2025 um 21:41 Uhr schrieb   
[email protected]  via Unicode <  [email protected] >:  The way 
I see it is that U+02C8 and U+02CC are spacing versions of U+030D and U+0329 
diacritics, and therefore to compose a spacing character with both diacritics, 
the spacing character of one and combining character of the other could be 
used. And there is already precedent of spacing diacritics composed with 
combining characters, particularly U+0385 which is composed as U+00A8 U+0301 
(although the precomposed version is encoded as it's essential for CP869, 
CP1253, and ISO 8859-7 compatibility).   Dnia 17 kwietnia 2025 21:05 Doug Ewell 
via Unicode <  [email protected] > napisał(a):   
[email protected]  wrote:   I really don't get why [the character 
proposed in] L2/25-061 would be  provisionally assigned to U+208F when it can 
be composed with  combining characters (ˈ̩ U+02C8 U+0329) or (ˌ̍ U+02CC U+030D) 
which  should be equivalent to the proposed character, and the potential use  
of the existing combining characters is not mentioned in the proposal,  but the 
proposal owner was informed of the compositions before the  Recommendations to 
UTC #183 were made.   While the quoted passage on the Submitting Character 
Proposals page makes sense for “normal letter with diacritic” proposals, which 
were once commonplace, I don’t think it’s typical to attach combining marks to 
a modifier letter such as U+02C8 or U+02CC, or for UTC to recommend composition 
in such cases.   The NormalizationTest file does not include any instances of 
combining characters used with modifier letters, except for a few wacky, 
cross-script, stress-test cases involving a combination of Latin letters, 
Hebrew accents, and Adlam modifiers.   Perhaps someone has authoritative info 
on whether the difference in handling is policy or just the way it’s been.   -- 
 Doug Ewell, CC, ALB | Lakewood, CO, US |  ewellic.org ewellic.org

Reply via email to