From: Raymond Mercier (RaymondM@compuserve.com)
Date: Mon May 23 2005 - 09:27:51 CDT
Gregg Reynolds wrote:
> So it sounds like the Hebrew text in your original document must have been
> sequenced "backwards" when you typed it in; i.e. if you wanted "foobar" in
> Hebrew you had to type "raboof" to get the right typeset ordering. So the
> sequence of codepoints in your file matches the sequencing of the typeset
> representation, whereas in Unicode text RTL text runs are reversed when
> typeset. Is that right? If so, your macro must not only convert to
> Unicode code units, it must reverse the Hebrew strings so that the Unicode
> bidi algorith will apply.
Years ago I wrote a little program that used keyboard hook techniques, which
ran in the background, to enable me to type 'backwards', as for Hebrew and
Arabic. This program just moved the cursor to the left automatically, so
that I could type the Hebrew letters in the order in which you would read
them. Of course since the font was just ASCII, with the usual L2R
directionality.
The problem now is that when the conversion to Unicode is carried out, this
runs into the directionality inherent in Hebrew, and this has to be
cancelled out.
Raymond
This archive was generated by hypermail 2.1.5 : Mon May 23 2005 - 09:28:46 CDT