Re: RFC 1556

From: Jonathan Rosenne (100320.1303@CompuServe.COM)
Date: Thu Jul 04 1996 - 13:02:25 EDT

Could I get hold of an electronic copy of the relevant pages of Unicode for
the preparation of a proposal to ammend ISO 10646?


Jony Rosenne

---------- Forwarded Message ----------

From: Olle Jarnefors <ojarnef@ADMIN.KTH.SE>
Subject: Re: RFC 1556
To: Multiple recipients of list ILAN-H <>
In-Reply-To: <> (Tue, 02 Jul 1996 08:58:28 GMT; From:
     (John Clews))

John Clews wrote, quoting Doron Shikmoni:

> > I believe this also represents Machon Hatkanim's view these days
> > (in the relevant committees). There seems to be a consensus that
> > the only solid, reliable and specifiable implicit bidi algorithm
> > is Unicode, and that there should be an effort to specify it in
> > all places, to help it become a de-facto standard and not just
> > a de-jure one.
> Is it _even_ a de jure one? Are there any international standards which
> relate to this (i.e. ISO/IEC standards?)

If you regard only standards under the ISO/IEC umbrella
as _de jure_ standards, then the Unicode bidi algorithm is
not such a standard. ISO 10646 _allows_ its use by
including the special characters needed, but their
semantics is only incompletely described in 10646, in the
_informative_ annex D. Alternatively, bidi text can be
coded in 10646 by using control functions from ISO 6429.

The Arabic and Hebrew parts of ISO 8859 specifies 8-bit
code values for Arabic and Hebrew letters but are silent
about the practically not unimportant issue of in which
order these should be placed when representing texts
longer than 1 character.

If Israeli or Arabic standard organizations are convinced
that the Unicode bidi algorithm is superior to whatever
ISO 6429 offers, they should definitely propose its
transformation into an ISO standard.

> Does anybody know if ISO/IEC
> JTC1/SC22/WG20 is dealing with this as well as IETF?

I don't know for sure if JTC1/SC22/WG20 is looking into
this, but JTC1/SC2 seems to me to be the proper ISO
committee for this fundamental issue in character coding.
The IETF will soon publish a specification of
internationalization extensions to HTML as a proposed
Internet standard. It is based on the Unicode method.
The latest version is available e.g. at:



This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:31 EDT