Re: A few notes on text

Gavin Nicol ([email protected])
Tue, 6 Jun 1995 12:54:30 -0400

>I've talked to our i18n guys and they've expressed concern about
>specifying family in this way. If I want to intersperse Japanese and
>English, I can do so by setting the coded character set to Unicode, which
>contains glyphs from both languages. For the font family I want to use
>Palatino for the English glyphs but my Palatino font contains no glyph for
>any of the Japanese characters. I do have Mincho on my machine which does
>contain Japanese glyphs.

I have assumed that English and Japanese would be represented by
seperate nodes. Given that the amount of text represented in VRML will
not be large, I assume that the slight inconvenience of linguistic
separation, will be more than offset by the handling of CJK issues in
Unicode.

>The i18n group suggested we make the family an MFString, allowing a list
>of fonts to be specified. When searching for a glyph, each font is
>searched in order until one is found satisfying the slot in the coded
>character set.

This will not work when Chinese and Japanese are interspersed using
Unicode (or rather, high-quality glyph disambiguation will not be
possible). Please take note of the language field on the text node,
which, when combined together with my outlined font specification,
would solve the CJK issues in Unicode (or rather provide a foundaton).