[tex-k] Feature requests: otf/ttf, font expansion

[Michael Zedler]

Hello,

Even though running pdftex in pdf mode has become tremendously popular, 
I prefer to go the traditional dvips way for several reasons. It'd be 
very nice if you would consider adding the following features
  1) otf inclusion (opentype cff)
  2) ttf inclusion (plain ttf/opentype ttf)
  3) automatic generation of expanded metrics (for use with 
\usepackage[DVIoutput,expansion]{microtype}

Reasons for 1,2: pfb are slowly dieing out, many useful fonts are made 
available these days as otf/ttf only.
These are the common workarounds: convert the otf to pfb (this may be 
prohibited in the license), for ttf not letting dvips include the font 
but to do this in the ps->pdf conversion stage. The latter involves 
having different map files for dvips and pdftex; this is problematic to 
impossible with updmap of today's distributions. (I don't consider 
ttf->pfb conversion an alternative because hinting information is lost 
and it may be prohibited in the license, too.)
The workarounds are not nice, I think.

Reason for 3: Right now one needs to have all expanded metrics locally 
stored. This adds extra work for the user to generate these metrics, and 
the tfm/vf pile can become extremely large: For a book I created 
expanded metrics of the MinionPro package on CTAN, this accumulated to 
some 1.5GB...

I'm not a programmer so I cannot provide patches, but the code does 
exist already:
  1) cfftot1 from the lcdf typetools
  2) Type42 conversion (from, e.g., fontforge)
  3) the autoexpand feature of pdftex

The second item might be less trivial than it sounds at first. Shall one 
a) follow the partly erroneous glyph naming of pdftex (not interpreting 
the cmap of the ttf), or b) use the cmap as fontforge does or c) use 
dvipdfmx' unicode<->glyphname mapping (making use of Adobe's 
glyphlist.txt)?
In order to avoid an enc-file mess, coordinating the glyph naming with 
the pdftex developers might be a good idea (that is, do they plan to 
correct the currently erroneous glyph naming in the future).

What do you think?

Best,
Michael