[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

`low slots'


> TrueType comes to mind when you think about restrictions below 32

(1) Actually, there is nothing in the TrueType spec about this and I have
many TrueType fonts that work just fine with characters in 0 - 32.
However, the Macintosh and Windows 3.1 and Windows 95 
implementations (and most font generation software) do not implement
the spec - only Windows NT does.

(2) More importantly, the problem with 0 - 32 is not restricted to TrueType
fonts, but to arbitrary scalable fonts, since most applications have
some problem with this.  On Mac, 13 is hardwired to return and not
a glyph, many applications treat strings as null terminated so 0 is
problematic. Acrobat has troubles with 0, 9, 10, 13, 32, and 127.  etc.

(3) The way this can be handled currently is to use the `Lucida remapping'
invented by Barry Smith in about 1988 when Adobe did the original
Lucida Math fonts.  There, 0 - 9 is remapped to 161 - 170 and 
10 - 32 is remapped to 173-195, and 127 to 196.  The glyphs can be
accessed using either character code (which caused a problem for
DVIPS for a while since it overwrites the metrics of the font).  The 
BSR/Y&Y/AMS... CM fonts are set up that way.  Of course,
this doesn't do you any good in a font that already uses those slots.

(4) I assume you are talking about math fonts (I get this only
intermittently since it goes to a CompuServe account I hardly
ever look at), since for text fonts it makes no sense to enforce the
encoding in the font in the first place.

> ...commercial vendors are more likely to go with such a setup
> if they can offer those fonts for other software.

I don't think you have to worry too much about math fonts being used in
other applications.  They are essentially useless when prepared for TeX.
This is because the advance width is not the advance width, but where to
put the subscript, the `true' advance width is the advance width of the
character plus the italic correction, etc.  You know the story.

> If such a restriction applies, could one not have that in mind, making
> one set of dumb fonts with no chars below 32 and vritual

But what it the point of moving things around using virtual fonts?
The only reason I can see for characters below 32 is so you can
have 256 glyphs in a font.  And if you leave out 0-32 you only have 
224 anyway...

Regards, Berthold.