tables-06b.txt: A.8 Gershayim

[Cary Karp]

Quoting Mark:

> I think the simplest course of action is just require the character
> before to be Hebrew; that is a sufficient limitation on usage.

I strongly agree with Mark's suggestion, which has the additional
benefit of permitting the rules for the geresh and gershayim to be
reunited. This also furthers the general effort toward streamlining the
aggregate rules to the greatest extent possible.

Although adopting that approach should moot further discussion of the
ways in which prospective name holders might wish to use the geresh
and gershayim, I would like to note that it is incorrect to assume that
the orthographic conventions of the Hebrew language are applied in all
other languages written with the Hebrew script. The YIVO rules for
Yiddish orthography, for example, use the gershayim as the first
character in a sequence of letters to indicate that the initial portion
of the word has been clipped, thus attesting to a situation where the
gershayim is not word internal.

I'm not suggesting that we support leading gershayim in IDNA, but this
case does highlight the risks the involved in taking the orthographic
rules for a dominant language written in a given script to be normative
for all other languages that use it.

/Cary

> On Thu, Jul 23, 2009 at 02:41, Matitiahu Allouche <matial at il.ibm.com
> <mailto:matial at il.ibm.com>> wrote:
> 
>     I totally agree with Ken's analysis of Gershayim usage, and with his
>     simplified pseudo-code.
> 
>     However, I seem to remember somebody mentioning using Gershayim at the
>     boundary between preceding Hebrew letters and succeeding letters from
>     another script.  Personally, I see no need for this, and such a label
>     would probably be disallowed anyway by the rules for Bidi domain names.
>     Still, if anybody thinks there is such a use case, he/she should speak
>     now.
> 
>     Shalom (Regards),  Mati