Peter Kirk wrote:
Perhaps I should clarify further. SIL Ezra was designed to use "legacy Latin-1 override or similar hacks". For example, in its Windows Character Set table it uses 0x41/0x61 for "a" sounds, 0x42/0x62 for "b" sounds etc - although it goes beyond Latin-1 in using nearly every code point from 32 to 255. But for various technical reasons connected with Windows, it is encoded as a Windows Symbol font, which means that its Unicode tables are mapped not to U+0020 to U+00FF, but to U+F020 to U+F0FF, following Windows Symbol conventions. This makes it rather less of hack in that PUA characters are used rather than regular Unicode code points being reused. And it can only be called a hack in retrospect, for it was designed at a time when full Unicode Hebrew was barely defined and certainly not widely implemented.
It doesn't really make it "less of a hack" since Windows just maps the
glyphs encoded from from F020 to F0FF in the cmap of "Windows Symbol" fonts to characters x20-xFF in the Windows code page for your locale (normally "Windows ANSI" if you are in the US or UK). You still type in non PUA characters, and those non PUA characters are what gets stored in your files - *not* PUA characters.
- Chris
