At 01:22 PM 12/31/2002, Peter Lofting wrote:
The main reasons to have short human-readable glyphnames are:These naming conventions may also be applied using uniXXXX format names, and are so used in plenty of fonts for un-encoded variants of non-AGL glyphs, e.g.:
(1) to have name identifiers for un-encoded glyphs.
e.g. A.swash1 A.swash2 A.smallcap A.initialcap A.endflourish A_period.swash
uni06AB.swash uni06AB.swash2 uni06AB_uni06D2.final
(2) to use these names in writing shaping behaviour rules. This applies to both OpenType and AAT (MIF) shaping rules.The uniXXXX format is not opaque, but it is obviously harder to work with than human-readable names *during production* unless you happen to have memorised the Unicode Standard. But the AGL and the Adobe glyph naming conventions are specifically concerned with final, shipping font glyph names, not with names used in production. This is why I am bewildered by the sudden appearance, in this new AGL, of all Adobe's internal production names. The implication of including names in the AGL is that these names will be recognised by ATM, CoolType and other Adobe text engines and mapped to corresponding Unicode values. Unless all these engines have always been able to recognise the Adobe production names included in the new AGL, it can't be a good idea to include them, because fonts using these names would not work with older versions of ATM etc. If these engines can recognise all these names, then all the new AGL really does is document a set of acceptable alternative names for some characters, that may be used in shipping fonts instead of othe AGL names or uniXXXX names. I continue to think that this version of the AGL may simply be an error, and that Adobe did not intend to publicise their internal production names in this way.
Writing glyph substitution/transformation algebra is much easier if the string is short, unbroken, and readable/recognizable. Unicode names don't fulfil all these criteria and the uniXXXX format is opaque, which prevents debugging and re-use of the shaping rule libraries.
At Tiro, we have an extensive set of human-friendly production names used for Arabic, Hebrew, Cyrillic, etc., but these are always swapped out before we ship a font and replaced by AGL 1.2 (where they differ, e.g. the AFII names for Cyrillic, Arabic and Hebrew) or uniXXXX names.
I disagree that using uniXXXX names in production would 'prevent ... re-use of shaping rule libraries'. Human-friendly production names are simply a convenience, and I can't think of anything we do with them (including deriving lookup rules from glyph repertoire spreadsheets) that we couldn't do using uniXXXX names with the same conventions for un-encoded glyphs.
Happy New Year, Peter et al.
John Hudson
Tiro Typeworks www.tiro.com
Vancouver, BC [EMAIL PROTECTED]
A book is a visitor whose visits may be rare,
or frequent, or so continual that it haunts you
like your shadow and becomes a part of you.
- al-Jahiz, The Book of Animals
