Re: Unicode and math symbols

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

   > Martin

   > > Please note that for one area where you would assume that
   > > dealing with "meaning" would be especially easy, the
   > > character/glyph model that is the general base of
   > > Unicode/ISO 10646 had to be changed somewhat (to
   > > come closer to "glyph" than for other scripts) to
   > > accomodate for present practice and user expectations.

   > Sounds interesting: to what do you refer?

   Well, there is a draft for an explanation of the character/
   glyph model with respect to CJK, which will probably
   become an informative annex to 10646.
   The main point where "shape" is preferred over "meaning" is
   where characters have been simplified so much as to lead
   to totally different shapes that cannot easily be identified,
   and where simplifications have lead to merging of several
   originally different characters. In both cases, "simulating
   paper" in the sense that there are different codepoints
   for the major shape differences is the better solution,
   because users mostly are not aware anymore e.g. about
   originally differentiated characters.

I think I have read this although I was so confused I don't
quite recognize your reference.  They spend 26 (twenty-six) pages
not quite defining the difference between glyph and character.
It is on the UNICODE 2.0 CD-ROM.

Let me stick my neck out here:  I know this was not the intent 
of UNICODE, and UNICODE has many features that make it non-ideal
for this, but UNICODE *is* a de facto glyph standard.  

(1) Which is why we have the `alphabetic presentation forms' 
ff, ffi, ffl, fi, fl, slongt, st etc. in UNICODE.

(2) Software like Framemaker and MathType could not use anything
but the Adobe Symbol font *until* they got UNICODE numbers for
glyphs in other math fonts from Adobe, Y&Y, Monotype etc.
They do not want to have the TeX nightmare of complex metric 
structures to be built for ever new font they come across.
And yes: they can't use half of the glyphs in LucidaNewMath
because they don't have UNICODE numbers.  Which is a serious 
failing of UNICODE, IMHO.

(3) It is what happens in Windows NT.  Yes, you can access glyphs
in a font by their numeric position within the font, but I don't
know of any software that does that.  Access is via UNICODE.
The exception are `pi' fonts. And of course right now we
mostly deal with math fonts in that mode.

Yes, you want font switches to cover stylistic variations, but
wouldn't it be nice if `contour integral' as accessible to same
way in a regular font and a bold font, in a font from Adobe
and a font from Monotype. etc.

Berthold.