I think what we gain is the same capability as the current HTML browsers.
I have a beta version of Netscape that supports Japanese, but other than
that I haven't seen any HTML browser that supports anything above LATIN-1,
but they all do that. As I see it the advantages to using LATIN-1 are:
Maximizing the languages hit with a single encoding implementation and
ease of implementation (all popular platforms can get access to LATIN-1
outlines). The disadvantages are as you mention.
Here's an interesting point. The VRML header has the word "ascii" in it.
This would indicate that the only text strings you could represent
directly (without some escape mechanism) would be ascii. But in the
Inventor implementation, even though our headers also say "ascii" we allow
the full Latin-1 set in text strings! Perhaps Inventor is wrong and we
should be more strict with VRML.
Perhaps we should limit text strings to ascii for simplicity. The
argument that unavailable glyphs would be represented as polygons still
holds - it just applies to more glyphs.
--
chris marrin ,,.
Silicon Graphics, Inc. ,`` 1$`
(415) 390-5367 ,|` ,$`
[email protected] b` ,P` ,,.
mP b" , 1$'
,.` ,b` ,` :$$'
,|` mP ,` ,mm
,b" b" ,` ,mm m$$ ,m ,,`P$$
m$` ,b` .` ,mm ,.`'|$P ,|"1$` ,b$P ,,` :$1
b$` ,$: :,`` |$$ ,:` $$` ,|` ,$$,,`"$$ .` :$|
b$| _m$`,:` :$1 ,:` ,$Pm|` ` :$$,..;"' |$:
P$b, _;b$$b$1" |$$ ,,`` ,$$" ``' $$
```"```'" b$P `""` ""` ,P`
`"` '$$b,,...-'
"As a general rule, don't solve puzzles that open portals to Hell."
- excerpt from "A Horror Movie Character's Survival Guide"