[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Undeclared glyphs

> Perhaps this is sane, if the spirit of TeX/LaTeX is that only `often
> encountered text glyphs' are present in the main font (*),
> 	(*) Of course, then we have to decide which ones are and which
> 	    ones aren't, or at least believe the decisions made in Cork
> 	    over those of the designers of ISO-Latin1.

    [Please note: I'm one of those whom you can blame for the Cork
    encoding. E.g., the `cwm' is my personal invention... :) ]

Reading all those mails about the evil Cork encoding, I would like to
remind people that it is *not* a font encoding. It is a TeX internal
encoding, i.e., an encoding TeX can work with. 

In the discussions we had during its setup the most important point
was: ``How probable is KERNING information needed between two
characters of this encoding?'' The second most important question
was: ``How many languages may we HYPHENATE with this encoding?''

Those of you who have an English language background should please
keep that in mind. TeX needs all characters of a word to be letters,
there must not be intermediate moves. Addition of arbitrary symbols
that are not letters to the Cork encoding was never considered, as it
is not probable that they must be kerned automatically. (Explicit
kerns by macros are not a problem, as we normally don't need to
hyphenate such constructs.)

The Cork group expected that the mapping from the TeX internal
encoding to an external font encoding is done by virtual fonts. (a)
It's portable over many different font techologies. (b) We did never
expect that all characters of the Cork group has direct glyph
equivalences in one font. I.e., one must use the glyph combination &
the font remap capabilities of virtual fonts.

We further expected that for certain font technologies further
remapping (from the external font encoding to an internal font
encoding) is still needed. This is particularly true for PostScript
where their reencoding is something virtual fonts cannot express. (And
should not express, I want to add.)

Hope this sheds some light on the ``historic part'' of that issue

        [member of the Cork group,
        participant of the Cork meeting where the final decision felt]

Joachim Schrod			Email: schrod@iti.informatik.th-darmstadt.de
Computer Science Department
Technical University of Darmstadt, Germany