Re: [Freetype] Unicode charmap for symbol fonts

[Steve Hartwell]

On Jan 31, 2004, at 12:36 PM, Garrick Meeker wrote:
> I'll try Font Inspector and keep looking, but FT_Get_Char_Index('A') 
> returns 0 with the unicode charmap, but a valid glyph with 
> FT_ENCODING_APPLE_ROMAN.

Well, you're right that the Unicode map for Symbol.dfont has no entry 
for U+0041 LATIN CAPITAL LETTER A (i.e. FT_Set_Charmap(face, 
FT_ENCODING_UNICODE); FT_Get_Char_Index('A')).  Strictly speaking, 
Symbol could map it to glyph ID 73, which looks like an 'A', but note 
that the font doesn't have glyphs for several other characters in the 
Latin alphabet range, such as U+0043 LATIN CAPITAL LETTER C, U+0044 
LATIN CAPITAL LETTER D, etc.).

Therefore I agree with the font designer that it would be inappropriate 
to have a Unicode mapping for U+0041 just because it happens to have a 
glyph which looks like a Latin uppercase A while lacking glyphs for 
most of the other Latin characters; in essence, since the Symbol font 
does not have mappings for the complete Roman alphabet, it ought not to 
have a mapping for any of those characters.

The MacRoman encoding, on the other hand, is a legacy encoding with 
more concern for being able to enter Greek capital letters with a US 
keyboard than purity of encoding, so it maps 'C' (0x43) to glyph ID 93, 
which looks like an uppercase 'X', and is in fact the glyph for the 
Greek character CHI; 'F' maps to PSI, and so on.  So in a sense it's 
MacRoman which is an "invalid" mapping, so to speak, but it's forgiven 
because it's legacy-- a kind of Respect for the Elderly :-)

Hope this helps,

Steve