[texhax] TrueType implementation question

[Brian O'Toole]

Hello TeX Users Group,

 

I have recently been exploring TrueType implementations of Computer
Modern fonts, and am looking at the TrueType format specification.  It
is clear from the spec and from apps that dump the contents of TrueType
TTF files that there could be multiple subtables in the 'cmap' area,
where 3.1 (Microsoft Unicode), 3.0 (Microsoft Symbol), and 1.0
(Macintosh Roman) are very common.  My question is, "On various
platforms such as Windows, Linux/Unix, Mac, how does the OS decide which
encoding to use?"  For example, in Windows, if I want to add a font to
the System Font Table and specify a TTF file, how does Windows know what
'cmap' subtable to use if both 3.1 and 1.0 are included in the TTF file?
Is there a way to specify the one I want?  How do Linux/Unix and Mac do
it?  I tried to display a particular character ID for a particular Font
on a PowerPC Mac running OS X, and was surprised when the Microsoft
Unicode encoding was used rather than the Macintosh Roman.  How do Java
and XWindows play a role in this?  Finally, are there equivalent issues
on the Postscript Type1 side?  Does anyone have a good reference?  

 

Thanks very much,

Brian

 

Brian P. O'Toole

Plotting & Exploration Group

The MathWorks

Phone: 508-647-7613

Email: botoole at mathworks.com

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://tug.org/pipermail/texhax/attachments/20070615/d87c0839/attachment.html