News of AFII: international standard registry of glyphs and their identifiers

["Berthold K.P. Horn" <bkph@ai.mit.edu>]

Peter_Constable@sil.org wrote:

>       The AFII site can be found at

>       http://www.unicode.org/afii/

>       AFII is not part of Unicode, but the site was being neglected
>       (I gather) at its former location and the Unicode Consortium
>       offered to maintain the site on their servers.

It was being irregularly maintained by one person.  We got our copy of the CD
and hardcopy version just before he went on vacation...

>       I have been wanting to get info about AFII, specifically to
>       what extent the industry is paying attention to the glyph
>       registry, but I haven't seen any indication that any major
>       companies are all that interested. (If that isn't so, I'd be
>       glad to see some of their representatives give such indication
>       in response to this message!)

I think: (1) few have realized the need for a separate `output encoding'
from `input encoding,' despite Unicode Consortiums determined efforts
to propagate the character/glyph distinction. (2) the world is not ready
for 32 bit glyph codes  - it's not even yet seriously coping with 16-bit
characters.  (3) For many purposes Unicode is adequate as a` glyph
encoding' even if that is not what the Unicode Consortium intended.

>       In fact, I saw recently that,
>       whereas Adobe had been making use of AFII glyph identifiers,
>       they have specifically decided not to do so any further.

AFII identifiers in Adobe Cyrillic fonts show the disadvantage of not
using mnemonic glyph names.  It makes it much harder to work with
such fonts.  By the way, do you have a reference on `Adobe decided 
not to do so any further'?  I would be most interested.

>       One potential candidate for this are the makers of Postscript:
>       Adobe have told us, in a recent message on this list, that
>       Postscript deals with glyph identifiers and not character
>       codes. They have also indicated, however, that they don't have
>       any need for an international, font-independent registry of
>       glyph identifiers.

This may be because they do it themselves.  Not waiting for the slow and
tedious standards process.  See
http://www.adobe.com/supportservice/devrelations/typeforum/unicodegn.html
and
http://www.adobe.com/supportservice/devrelations/typeforum/corporateuse.txt
Their current approach seems to be to populate the 
`corporate use subarea' of Unicode space with the glyphs they need.
To some extent this is coordinated with other players like Apple
and MS, although corporate animosities may limit how well this works.

>       Another group that might be interested in such a registry are
>       type designers: they need a way to document glyphs they have
>       designed, and this registry should be useful to them. Adobe are
>       involved in type design, but again don't seem to need this
>       registry. 

See above.  They invent mnemonic glyphnames and maintain a mapping
to Unicode plus the corporate use subarea.

Given that the only working systems using Unicode (AT&T Plan 9 and Windows NT)
do not use a separate encoding scheme for output, it is imperative that
Unicode accept ligatures and `alphabetic presentation forms'
and some `glyphic variants' - despite the dogma on glyphs versus characters.

There are some mysteries in this regards as far as I can see.  For example,
as far as I can tell, there are not Unicode assignments for all glyphs needed
for polytonic Greek (tell me if I am wrong).  Since Unicode is not about
composing characters by combining base and accent characters, this
seems like an omission.  Also, only a few `blackboard bold' symbols 
are in Unicode. And so on.

Regards, Berthold.

Berthold K.P. Horn		mailto:bkph@ai.mit.edu