Problem with Mandaic shaping, IT and IN switched
Benjamin Riefenstahl
b.riefenstahl at turtle-trading.net
Mon Jun 30 15:47:39 CDT 2014
Hi everybody,
I am currently in the process of designing a simple OpenType font for
Mandaic. As some of you are probably aware, shaping in OpenType as it
is recommended by the OpenType standard requires that the application
(i.e. the text rendering engine) knows the joining behaviour of the
characters.
It seems that there is an error in the joining data for Mandaic as
defined by the Unicode standard (table 14-5 and 14-6, chapter 14.12 in
version 6.3) and by the file ArabicShaping.txt at
http://www.unicode.org/Public/UNIDATA/ArabicShaping.txt.
The tables list the character IT as dual-joining and the character IN as
right-joining. These two seem to be switched. In the table columns
with the actual characters (columns Xn, Xr, Xm, Xl) the correct
characters are given (compare the code chart at
http://www.unicode.org/charts/PDF/U0840.pdf), but the names (and the
relative positions in the tables) are wrong and that error is than taken
over into the file ArabicShaping.txt:
0847; MANDAIC IT; D; No_Joining_Group
[...]
084F; MANDAIC IN; R; No_Joining_Group
The correct characters in the table should be (in this order)
* Dual-Joining: ATT, AK, AL, AM, AS, IN, AP, ASZ, AQ, AR, AT
* Right-Joining: HALQA, AZ, IT, AKSA, ASH
And the correct data in ArabicShaping.txt:
0847; MANDAIC IT; R; No_Joining_Group
084F; MANDAIC IN; D; No_Joining_Group
Please advise what I can do to help correct this in some future version
of the Unicode standard.
Regards,
Benjamin Riefenstahl
More information about the Unicode
mailing list