Re: Symbol font mappings

From: Robert Herzog (
Date: Fri Oct 15 1999 - 05:52:07 EDT


> On Thu, 14 Oct 1999 18:53:18 -0700 Peter Edberg <> wrote:
> With regard to your first point: The Apple mappings on the ftp site are
> explicitly for Unicode 2.1 (as indicated in the header of each file),
> not Unicode 3.0. Also, Unicode is a character encoding (not a glyph encoding);
> nevertheless, we may take the glyph change into consideration when deciding
> whether there is enough reason to change the tables for Unicode 3.0.
> However, that has not been done yet.

I see. I guess this depends a bit on how one looks at this 'editorial change'
from Unicode 2.x to 3.0. If one thinks of it as a correction to an 'unfortunate
bug', then one could even change the 2.1 mapping with an explanatory note.
(The whole rational was discussed earlier on this mailing list.)

> With regard to your second question: The Adobe mapping tables on the ftp site
> are covered by a caveat in the readme.txt file: "[These files] were originally
> provided by the Unicode Consortium for use by NeXT implementations with DPS,
> and continue to be provided for compatibility with those implementations.
> All others should refer to the files on mentioned above."
> The files referred to are at:
> <>
> but as far as I can tell, there is no Symbol font mapping there.
> Which Adobe Symbol mapping are you referring to?

I was referring to the following two files at the Unicode site: 28/7/99 23/9/99

From the Adobe/symbol.txt file:

# Name: Adobe Symbol Encoding to Unicode
# Unicode version: 2.0
# Table version: 0.2
# Date: 30 March 1999
# General Notes:
# The Unicode values in this table were produced as the result of applying
# the algorithm described in the section "Populating a Unicode space" in the
# document "Unicode and Glyph Names," at
# to the characters in Symbol. Note that some characters, such as "space",
# are mapped to 2 Unicode values. 29 characters have assignments in the
# Corporate Use Subarea; these are indicated by "(CUS)" in field 4. Refer to
# the above document for more details.
# Contact <> with any questions or comments.

Although there is the caveat you mentioned in the Adobe/readme.txt file,
Table version 0.2 dates from March this year and has been changed
(by Adobe or someone at Unicode?) to follow the recommendations of
I assume this means that this is the 'official Adobe' table.

As I understand it the table was created in the following way:

1. Symbol font code point -> PostScript name according to
   appendix E.12 from the PostScript Language Reference

2. PostScript name -> Unicode value according to the Adobe glyph list
   (Of which the following lines might also have to be changed to
    reflect the unicode 3.0 glyph swap:
    03D5;phi1;GREEK PHI SYMBOL)
   Since all characters in the Symbol font are in the Adobe glyph list,
   some using the Corporate Use Subarea, there is no other mapping
   mechanism necessary.

The result of this procedure is a different one from the Symbol font
mapping used by Apple. So the question remains: Is this useful?

Greetings, Robert

This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:53 EDT