From: 	Mansour, Kamal[SMTP:kamal@ca.monotypeusa.com]
Sent: 	30 September, 1997 18:50
To: 	Hart, Edwin F.
Subject: 	RE: Comments on Character/Glyph Model (ISO DTR 15285)

Ed,

I finally was able to set aside time to read through DTR 15285 and your recent
comments. It's pretty impressive to see all these ideas so well formalized in
the document. In general, it expresses the variety of concepts relating to
characters and glyphs very well. As is true in most cases, if you're not already
steeped in a particular discipline, standards make for some dry reading. In
contrast, I think people working in various areas related to computer-aided
publishing (especially, multilingual) will find this document quite readable
because the problems make themselves evident. The document is written
generically enough so as not to trip people over small technical details.

Re Comment 8 (renaming 'coded-font model'): I'm in full agreement.

C.3.2 The last sentence caught my eye:
'The glyph identifiers used in a character-to-glyph mapping must be the same as
those ...'. 
I think that requirement is overstringent. It not identical, the glyph
identifiers need to be mappable by an explicit method to those used in the font
resource. If my font calls a glyph 'U017c' or 'zdot' instead of 'zdotabove' or
'afiiixxxx', it shouldn't matter. The currently used rigid encodings have saved
us from this problem, but it's always lurking. 

This standard takes us one step towards recognition of char-glyph differences.
At a later stage, the interface between the various pieces, as well as the
various char-glyph translators, will need to be further defined by another
standard or document.

Best regards,

Kamal Mansour