mtxtoafm

Han The Thanh thanh@informatics.muni.cz
Wed, 4 Oct 2000 09:19:40 -0400


> I do not understand you completely here.
> 
> CC instructions are part of AFM files if the fonts actually use the SEAC
> operator to print accented characters. Some more recent fonts do no
> longer use SEAC and therefore there are no CC data in the metrics -
> these fonts use subroutining for compactness.
> You can however provide CC data in the metrics files to allow Tex (via
> fontinst) to construct those characters despite they are not available
> in the font program or are not provides in the encoding of the "raw"
> fonts. This is very useful for producing virtual fonts covering Tex T1
> encoding from the Adobe standard character set.
> 
> However this means those accented characters are constructed by the
> virtual font and not searchable in a PDF. To avoid this one must simply
> provide in the font program (and "raw" encoding as well) the complete
> character set.
> 
> You must create for every font without CC data the correct numbers by
> hand (or a script which simply centers accents). If you are at this
> point you can use the program COMPOSIT from the Y&Y font manipulation
> package to generate in a given Type1 font the SEAC instructions for the
> CC data given in an AFM file.

that's exactly what I would like to do: create the CC data and use COMPOSIT
by Y&Y to composite the accented glyphs. The FMP package also allows having
glyphs from more than one font; it is extremely useful as I would like to
make the accents for uppercase letters different from the accents for
lowercase (the uppercase accents should have smaller heigh, which could be
done using horizontal scaling).  I have created such virtual fonts by
fontinst, so I think it would be feasible to use the corresponding data to
compose them using COMPOSIT instead of virtual fonts.  Generating the CC
data by hand is however a pain, while fontinst could do a much better job.

> The problem with these fonts is that the SEAC composites are not
> correctly rendered by ATM on MS-Win if reencoded to e.g. CE encoding -
> Adobe Acrobat Reader however does display all the added SEAC composites
> correctly without problem. I.e. you can use the fonts within PDFs but
> not install them on MS-Win for ATM.

the problem with SEAC can be avoided using another program from FMP as
well.

Regards,
Thanh