Re: bidi discussion list was: Bidi Markup vs Unicode control characters

I agree on the substantive issue (lacking complete mappings): moreover, if
you have just a language tag like "az", it could be written with either a
RTL or LTR script.

Thus getting the overall direction from the language tag may be a heuristic,
but cannot be depended on.

‎Mark

----- Original Message ----- 
From: "Martin Duerst" <duerst@it.aoyama.ac.jp>
To: "Stephen Deach" <sdeach@adobe.com>; "Jony Rosenne" <rosennej@qsm.co.il>;
"'Stephen Deach'" <sdeach@adobe.com>
Cc: <www-international@w3.org>
Sent: Tuesday, August 16, 2005 19:06
Subject: RE: bidi discussion list was: Bidi Markup vs Unicode control
characters


>
> At 01:05 05/08/16, Stephen Deach wrote:
>  >
>  >Yes, they are scripts (I said so).
>  >
>  >Martin indicated he had a list of language-to-script correlations. I'ld
> like to see it.
>
> No, I didn't. What some people (including to some extent you)
> claimed is that knowing the language can be used to determine
> directionality.
>
> What I said was that for most scripts, including some of those
> written RTL, the list of languages using that script is essentially
> open. I.e. I very much claim that I don't have such a correlation,
> and I also claim that nobody else has such a correlation that is
> complete. This prohibits implementation of generic language-to-direction
> mappings that would be needed in browsers to be able to substitute
> language information for explicit directionality information.
>
> Regards,    Martin.
>
>  >At 2005.08.15-18:45(+0200), Jony Rosenne wrote:
>  >>These are scripts, not languages.
>  >>
>  >>The Hebrew script, for instance, is routinely used for at least three
>  >>languages and rarely used for many more.
>  >>
>  >>The Arabic script is used for a number of languages today, and
historically
>  >>for many more.
>  >>
>  >>Jony
>  >>
>  >> > -----Original Message-----
>  >> > From: www-international-request@w3.org
>  >> > [mailto:www-international-request@w3.org] On Behalf Of John Cowan
>
>  >> > Stephen Deach scripsit:
>  >> >
>  >> > > The only scripts identified as RTL in Unicode are Arabic
>  >> > and Hebrew.
>  >> >
>  >> > In fact, Syriac and Tifinagh are already encoded in Unicode,
>  >> > as well as the
>  >> > archaic scripts Cypriot and Kharoshthi.  Phoenician has been
>  >> > fully blessed
>  >> > and will be in the next version.
>  >> >
>  >> > Still in the long tail are Old Hungarian (aka rovasiras),
>  >> > Avestan, Mandaic,
>  >> > Samaritan, Manichaean, and perhaps others.
>
>
>
>

Received on Wednesday, 17 August 2005 15:33:03 UTC