From: Jony Rosenne (jr@qsm.co.il)
Date: Thu Jan 10 2008 - 00:00:22 CST
See Unicode Standard Annex #9, The Bidirectional Algorithm, clause 3.3.1.
"The Paragraph Level".
It seems there are many non-conforming applications out there, but the
Unicode standard is clear and sufficient in this respect.
Jony
-----Original Message-----
From: unicode-bounce@unicode.org [mailto:unicode-bounce@unicode.org] On
Behalf Of Murray Sargent
Sent: Thursday, January 10, 2008 4:44 AM
To: Behnam
Cc: 'Waleed Oransa'; unicode@unicode.org; verdy_p@wanadoo.fr
Subject: RE: Directionality Standard
Behnam, I have to admit that I'm tempted to agree with you. Unicode has
run-level directionality format characters, e.g., RLM and LRM, but depends
on a higher-level protocol for paragraph directionality. In working on the
RichEdit engine over the years, I have run into situations where it would
have been convenient to be able to specify the desired paragraph direction
unequivocally in plain text. We even have a BiDi context feature that
chooses the paragraph direction based on the first strong character
encountered. Works the way you'd like most of the time.
But I wonder how many display engines would break at this late date if we
added paragraph directionality control characters. One scenario that comes
to mind is copying an RTL paragraph to the plain-text slot on the clipboard.
If an application prefixes the text with an RLPM (right to left paragraph
mark), will target applications get confused and display a missing-character
glyph? The same problem could occur with any unaware editor when opening a
plain-text file containing an RLPM.
So for a time at least, BiDi display would be messed up by the introduction
of RLPM and LRPM format characters. One has to gauge whether the gain is
worth the temporary confusion. If you stick with a higher-level text format
like HTML, you don't have any problem.
Thanks
Murray
This archive was generated by hypermail 2.1.5 : Thu Jan 10 2008 - 01:04:03 CST