Re: Abkhasian pe with descender?

At 17:00 -0800 2007-11-09, Kenneth Whistler wrote:

>And the GHE WITH DESCENDER was proposed for Siberian Yupik, where
>there was no indication of the usage of U+0494/5 GHE WITH MIDDLE HOOK
>at all.

Why would there be, if MIDDLE HOOK and DESCENDER are merely 
allographs? (They are not.)

>So the situation is rather different from the situation
>where characters are being proposed for a language where there
>clearly *are* alternate typographic traditions that use both
>forms.

That's an assertion on your part.

>I'm not sure what you mean by the trailing off with "and there is
>no question..."

Oh well. Sometimes we trail off.

>  > I've got an Abkhaz who has just last week contacted me and he has
>>  said that he considers the letters to be different.
>
>You can get anybody to look at fonts and tell you that "letters" are 
>different.

I didn't solicit this gentleman's mail.

>The question is whether the representation of Abkhaz text would
>be better served by treating these as typographical variants
>of the same abstract character or not.

That's one question. Another question hinges on what value the 
encoding of historical and modern letters has. I imagine some people 
disliked the encoding of Lhuydian turned Gaelic g for early Cornish. 
The letters were encoded for their historic value.

>You are simply asserting based on typographic form that things are 
>distinct and that therefore they must be encoded, when it seems 
>pretty clear that we *are* talking about a single letter here for 
>Abkhaz, with stylistic glyph variation.

The UCS does not "unify" the DESCENDER modifier and the MIDDLE HOOK. 
Indeed, these two marks are regular and distinct. There is no 
justification for saying that, in the case of the letter PE, they are 
allographic.

>  > There are many orthographies for Abkhaz. It may be inconvenient for there
>  > to be so many letters (Latin and Cyrillic) used for it, but that's history
>  > for you.
>
>Yes, but going too far in insisting on encoding every mangled 
>Cyrillic sort that was used for printing Abkhaz and other minority 
>languages may artificially fragment orthographies more than they 
>reasonably need to be -- which I think was Eric's point.

MIDDLE HOOK and DESCENDER can't be confused by "mangled Cyrillic 
sorts". I've seen my fair share of Cyrillic sorts, and printed on 
bumaga No 2, too, which can laughingly be called "paper". For my part 
I will pursue this with my Abkhaz informant and our Dutch colleagues 
and see what there is to see.
-- 
Michael Everson * http://www.evertype.com