RE: U+1889 Isolate Plus 1828/182D/1833

Hi Badral,

Assignment of the individual FVS to a given function has always been somewhat erratic (as I know it), but then I was not involved in the early days of assignment, so cannot really say for sure. The U+1822 medial over-ride (needed to handle NAI+FVS2+MA) was raised in this discussion - I am not sure how it was handled in the very earliest encoding standards. The U+1828 medial over-ride case of ANAR is in the current Chinese Standard as requiring FVS1, but was not in the MGWBM. The U+182D medial over-ride needed for compounds like COG+FVS2+A+FVS1+GULA was not mentioned in the earliest of documentation. The assignment of U+182D+FVS1 to over-ride the default U+182D was made as early as 2000 in the MGWBM. The means of handling SAY+FVS2+IQAN (with the curved Y) is inferred in the MGWBM, but the documentation could be interpreted several different ways - one might argue that the early FVS2 assignment was inferred. So, it appears to me that there was no early attempt to say that a given FVSx was to meant carry out a certain function. Others might comment here and fill in the gaps of the history.

As for the U+182D+FVS2 assignment, I am told that there are several words that are masculine with a leftward-sweeping orkitz. If anyone can help with itemizing this list, it would be helpful. Even without such a list however, we need some device to display the final feminine GA - the U+182D+FVS2 assignment is justified therein.

[cid:image002.png@01D10FDD.9BC36D00]

Greg



From: Badral S. [mailto:badral@bolorsoft.com]
Sent: Monday, October 26, 2015 1:08 AM
To: public-i18n-mongolian@w3.org
Subject: Re: U+1889 Isolate Plus 1828/182D/1833

Hi Greg,
For 1,2,3, I am still very sceptical. Because it might destabilize current *unstable* data.
4. Ok, understand. I thought that always FVS2's used for overriding. Is 182D final+fvs2 not redundant, if we use 182D final + FVS1 as overrider?

Badral

On 25.10.2015 14:54, Greg Eck wrote:
Hi Badral,

OK, I will mark 1889 as green.


1.)    / 2.) / 3.) I tend to agree with you that what we have now is not necessarily the ideal choice of the default. However, it is what the last 15 years of Mongolian font development has given us. The parameters that seem important in a proposed change/clarification (to me at least) have been:

*         We should try to upset the current font implementations as little as possible

*         Where the implementation does not match the specification, if reasonable, look at changing the specification (eg. U+182C-Medial/Final where the specification said that the two glyphs were to be implemented in the medial position but the font developers all implemented at the final position)

*         If there are sound examples of a new variant and there is seeming agreement between the font developers, add the variant (eg. U+1822-Final+FVS1)

*         Regarding isolates, native script forms should hold a higher value than that of the Unicode-ascribed glyph forms (eg. U+1824/U+1826 where the Unicode-ascribed presentation form does not match the most common isolate form). Where the Unicode-ascribed presentation does not match the desired isolate form, the Unicode glyph should at least be assigned an FVS so that it is displayable as a variant isolate.

*         Surely there are other principles that each of you might want to suggest ... Please comment here

*         My thoughts on the matter are that to change the defaults in the suggested code-points (1828/182D/1833) are massive enough to cause great disruption of font acceptance/usage.

4.) There is no difference in the glyph. Earlier we defined several over-rides - U+1822-Medial, U+182D-Medial, and U+182D-Final (discussion is dated around September 9, 2015). This is the situation where the default context needs to be over-ridden to shape the words where the common context does not match. In other words, given that the shilbe is considered an "I", then "eight" is spelled NAI+FVS2+MA. AN+FVS2+AR spells the archaic ANAR. Feminine "similar/like" is spelled S(H)IG+FVS1 where the expected left-swing orkitz is actually a right-swing suul.

Greg
>>>>>>
Sent: Sunday, October 25, 2015 11:03 AM
Subject: Re: U+1889 Isolate
Hi Greg and all,
I agree with the 1889. I just reviewed http://r12a.github.io/scripts/mongolian/variants and raised following questions.
1. Why we should not switch U+1828 medial and U+1828 medial + FSV1?
2. Why we should not switch U+1833 medial and U+1833 medial + FSV1?
3. Why we should not switch U+182D medial and U+182D medial + FSV1?
Main writing rule of Mongolian is the alternation of consonants and vowels. All subscribers in this list know it.
Hence, we can easily recognise (without any statistics etc.) which variants are more frequently occur especially in the middle of a word. What do you think? Before vowel variant or before consonant variant?
In Mongolian, consonants are not usable as separate character. That's why we spell all consonants as syllable like DA, NA, GA etc. Here A means a vowel and seems as vowel. Then, why we don't select the form before A(Vowel) but select the exceptional (devsger) variants as default?
I think, it makes no sense and longer typing as well as inefficient linguistic processing.
4. What's difference between U+182D final and U+182D final + FSV1?

Badral
>>>>>>>


[imap://badral%2Ebolorsoft@mail.bolorsoft.com:143/fetch%3EUID%3E.INBOX%3E4948?header=quotebody&part=1.2&filename=image001.png]


--

Badral Sanlig, Software architect

www.bolorsoft.com<http://www.bolorsoft.com> | www.badral.net<http://www.badral.net>

Bolorsoft LLC, Selbe Khotkhon 40/4 D2, District 11, Ulaanbaatar