Re: How can I check for the existence of a glyph in TeX?

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

At 06:25 AM 98/09/15 +0100, Rebecca and Rowland wrote:

>>The fonts themselves are identical.

>Not in my experience they're not.

You are wrong.  The Font Manipulation Package for example includes MACtoPFA
and 
PFBtoMAC which convert between Mac and PC Type 1 formats making it easy to
verify
that, for example, any font in the Adobe Type Library has the same internal
information
on the Mac as it has on the PC.  You can take the Mac font file and
automatically
create a PC font file that is byte-for-byte identical to the one you get
>From Adobe for PC.
Going in the other direction, again you get byte -for-byte equality except
for the 
addition of the icon, version and bundle resources which show you the nice
slanted
`A' (or other) icon.  Of course, none of that affects the rendering of the
glyphs in any way.

>>Adobe text fonts have 228 `standard'
>>glyphs.  You can convert the actual font file from Mac to PC format back
>>and forth without losing anything.  The text font does have all those
>>glyphs --
>>on any platform.   A stand-alone PS driver like DVIPS can trivially get at
>>all 228 if it wants to (if DVIPS could understand the Macintosh Type 1 font
>>file format).

>It's a nice theory; what about the point that Mac founts often have glyphs
>that don't exist in normal founts (encoded or otherwise): things like pi,
>sigma, delta, rabbit, apple, candle, and so on?

Mac text fonts do not actually have any of those glyphs.  On the Mac, 15
`mathematical' symbols are imported from the Symbol font into text fonts:

% notequal, lessequal, greaterequal, approxequal, partialdiff, integral
% summation, product, pi, infinity, Delta, Omega, radical, lozenge, apple

As a result these look identical, no matter what font you select.  Hence are
essentially a useless artifact that wastes 15 slots in the Mac standard roman
encoding. A few more `mathematical' glyphs used to be treated this way
originally, 
but are now `real' - part of the text fonts.

% degree, plusminus, mu, logicalnot, divide, fraction

>>Clearly the `right' solution is for the DVI viewer to work around the
>>operating system
>>limitation and allow access to all 228 standard characters.  Why should
one be
>>prevented from using ff, ffi,ffl ligatures on the Mac if the font has them
>>(e.g. the
>>Lucida Bright fonts) just because Mac standard roman encoding doesn't
>>include those?

>Because it's difficult to work around the OS, obviously.  If it were an
>easy job, lots of dvi drivers other than Y&Y's would be able to do it.  As
>it is, I gather that no (free|share)ware MS-Windows dvi driver (aside from
>dvips) can even do re-mapping; and while OzTeX can manage re-mapping, it
>can't do re-encoding (unless it uses dvips).

Shame... :-)

Regards, Berthold.

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