[Fontinst] Mapping all diacritics to actual glyphs rather than composites

Reinhard Kotucha reinhard.kotucha at web.de
Fri Jan 15 22:26:56 CET 2010

On 15 January 2010 Lars Hellström wrote:

 > Just to be clear here: By "composite" you mean that there are really 
 > two glyphs being combined to make up one letter?
 > I first felt a bit confused when reading this, but that's probably 
 > because I've been thinking about Unicode lately, where (the semantic 
 > counterpart of) this kind of thing would be called "decomposed".

Well, composed chars are supported by Unicode, though deprecated
nowadays.  Since Christopher is talking about Type1, I suppose that
the seac subroutine (standard encoding accented character) is meant.
As its name implies, it supports Standard Encoding only.  For glyphs
not supported by Standard Encoding you usually have provide a new
glyph.  More about seac here (page 56):


If only accented characters are needed, one could put both, the base
char and the accent into different subroutines.  This saves space, at
least if different base chars are to be combined with different

Ogoneks are quite problematic because they are part of the glyph.  If
the ogonek is a separate outline, it's very likely that it overlaps
with the outline of the base character, which isn't allowed by the
Type1 spec.

If outlines overlap, you get funny results.  In Gentium, the
Vietnamese glyphs with a horn accent show this problem in Adobe Reader
at some resolutions:


It's easy to add accents to glyphs if there are no overlaps.  Han The
Thanh provided separate glyphs for all the *horn characters.  The
following example shows /o from Blue Sky Research Computer Modern
(blue) and /ohorn from VNR (red):


There are no overlaps but a lot of work was necessary.

At DANTE-2006 in Berlin, Boguslaw Jackowski held a talk where he
explained why /aogenek should be a separate glyph rather than a
composite:  The ogonek is not an accent, it's part of the glyph.  And
whenever you create such a glyph, you have to keep in mind how such
glyphs are drawn using a pen.

Christopher, I'm pretty sure that TeX Gyre Pagella is a much better
starting point than URWPalladioL.  Especially because you can be sure
that Jacko made sure that Eastern European languages are supported
properly.  And Pagella provides SC and OsF already.  Pagella is
currently not hinted optimally; Jacko is aware of it.  I think that
the sole benefit of using Adobe fonts is that the latter (currently!)
contain better hints.

It's definitely worthwhile to evaluate TeX Gyre Pagella before you
invest a lot of time into your own developments.  In LaTeX, just say



Reinhard Kotucha			              Phone: +49-511-3373112
Marschnerstr. 25
D-30167 Hannover	                      mailto:reinhard.kotucha at web.de
Microsoft isn't the answer. Microsoft is the question, and the answer is NO.

More information about the fontinst mailing list