README
¶
Character Encoding for TrueType Fonts
This example generates a PDF file which can be used to explore how a viewer maps character codes to glyphs for simple TrueType fonts.
Methods
Given a single-byte code c, we consider the following method to map c
to a glyph in a TrueType font:
A. Use a (1,0) "cmap" subtable to map c to a glyph.
B. In a (3,0) "cmap" subtable, look up either c, c+0xF000, c+0xF100,
or c+0xF200 to get a glyph.
C. Use the encoding to map c to a name, use Mac OS Roman to map the name to
a code, and use a (1,0) "cmap" subtable to map this code to a glyph.
D. Use the encoding to map c to a name, use the Adobe Glyph List to map the
name to unicode, and use a (3,1) "cmap" subtable to map this character to a
glyph.
E. Use the encoding to map c to a name, and use the "post" table to look
up the glyph.
The program ttf-encoding constructs a sequence of TrueType fonts which allow
to determine which of these methods are supported in the viewer, and in which
order they are tried.
Results
| Preview | Ghostscript | Acrobat Reader | FireFox | Chrome | |
|---|---|---|---|---|---|
| 2.0/+S/+E | DBC | BDA | DBA | BE | DBEA |
| 2.0/-S/+E | DBC | DC | DC | DC | DC |
| 2.0/+S/-E | BA | BA | BA | BA | BA |
| 2.0/-S/-E | B | - | - | A | AB |
| 1.2/+S/+E | DBC | BDA | DBA | BE | DBEA |
| 1.2/-S/+E | DBC | DC | DC | DC | DC |
| 1.2/+S/-E | BA | BA | BA | BA | BA |
| 1.2/-S/-E | B | - | - | A | AB |
Conclusions
- Good choices when writing a PDF file seem to be the following:
- symbolic font, no encoding, use method B
- non-symbolic font, encoding, use method D
- For readers, implementing the behaviour of Google Chrome seems a good choice.
- Existing viewers seem to ignore the PDF version.
- I found no evidence that Adobe Acrobat Reader implements method E.
Documentation
¶
There is no documentation for this package.
Source Files