Unicode 11 Georgian uppercase vs. fonts

[Michael Everson via Unicode]

On 27 Jul 2018, at 09:35, Alexey Ostrovsky via Unicode <unicode at unicode.org> wrote:

> On Fri, Jul 27, 2018 at 8:54 AM, James Kass via Unicode <unicode at unicode.org> wrote:
> https://unicode.org/wg2/docs/n4712-georgian.pdf
> 
>> The revised proposal to change the Georgian encoding model from caseless to casing was convincing and compelling.  (It's bilingual, too, English and Georgian.)
> 
> It may look so, but my statement is still correct.

No, it’s not.

> This is not the first time, when the consortium mistreats Georgian (one can remember a story of encoding the ecclesiastic minuscule).

You have me to thank for undoing that mistake. And some other mistakes. We all make mistakes. 

> Just two points:
> 1) "compelling" (less important). The supporters are either font designers or non-specialists organizations. There are several institutions in Georgia that had to be involved IMHO (like Institute of Georgian Language, Institute of Manuscripts and Academy of Sciences; Ministry of Economy is not an institution competent in the script issues).

Those institutes were consulted. I met with representatives of both of them on my trip to Tbilisi to work with the font designers who spearheaded this project. The analysis in N4712 is correct. 

> 2) "convincing". I will not discuss all the controversies here, but will only cite §1.1 and §8:
> §1.1, on "Mkhedruli… is caseless, and no casing behaviour is expected or permitted by Georgian users. The mtavruli titling style of Mkhedruli… is not case; it is a style analogous to small caps or bold or italic. <...> Mtavruli-style letters are never used as “capitals”; a word is always entirely presented in mtavruli or not. Mtavruli-style is used in titles, newspaper headlines, and other kinds of headings." of the original encoding (N2608R2):

I wrote N2608R2 and I said explicitly in N4712 that this was mistaken. It was mistaken because I had never seen the 19th-century title-casing material, and because I made assumptions about how to handle it as a style. The UTC also made a mistake in taking this at face-value, at least insofar as the suggestion that “small caps” styling be used, since small caps is dependent upon encoded capitals. So don’t go quoting me in 2003 in order to argue against me in 2016.

> — "This statement was not correct."
> At the same time, §8 on successful implementation of the proposal in question: "Within a sentence a given word might be written IN ALL CAPS (MTAVRULI) for emphasis. An entire sentence or header may also be written in Mtavruli." And all the sample photos of the modern books and journals demonstrate exactly the same behavior as described in N2608R2: " Mtavruli-style is used in titles, newspaper headlines, and other kinds of headings".
> (I can provide more information if needed)

The mistake in interpretation is yours. Here:

1) Latin script cases. It has capital letters and small letters.

1a) English orthography uses capital letters at the beginnings of sentences and of names and of the names of the months and weekdays. Sometimes ALL CAPS are used.

1b) French orthography uses capital letters at the beginnings of sentences and of names but not at the beginnings of the names of the months and weekdays. Sometimes ALL CAPS are used.

1c) German orthography uses capital letters at the beginnings of sentences and of names and of the names of the months and weekdays. Sometimes ALL CAPS are used, but not in Fraktur font styles.

2) Georgian script cases. It has capital letters and small letters.

2a) When Georgian orthography uses capital letters it uses them on every letter in the word where they are used, regardless of what kind of word it is.

2b) In the 19th and early 20th century there was an orthography which used capital letters at the beginnings of sentences and of names, as well as full-word ALL CAPS. 

> The key question is whether Georgian is caseless or not in plain text encoding, and N2608R2 does not provide any evidence for casing in modern Georgian. 

N2608R2 was written in 2003 and has been superceded by N4712. Mtavruli is ALL CAPS. Mtavruli is not small caps. 

> Basically, the issues addressed are the low level of technical support for implementing small caps in Georgian typesetting (but this must not be Unicode issue) and incorrect idea that small caps must be preserved in plain text encoding (just because someone loves it), it is obvious from §1.1 (right after the text I cited).

It is now possible to use Georgian small caps, since both capital letters and small letters are encoded. Previously it would not have been possible to do so, since small caps is a fancy-text style of presenting lowercase letters with uppercase glyphs. 

Mtavruli is not small caps. Mtavruli is ALL CAPS. 

Michael Everson