> > >If this is the case couldn't we pick a single encoding to support in
> > >VRML and get rid of the encoding field?
> >
> > Not really, and the main reasons are:
> >
> > 1) Lack of tools
> > 2) Emotional attachment to established encodings
> > 3) No single encoding is optimal for everything
> >
> > What does it mean to encode JISX 0208 using UTF8? This is perfectly
> > possible, but the tools for data creation do not exist, and even if
> > they did, most people would not appreciate the fact that the could not
> > look at a VRML file in their PC editor, and understand what was
> > contained in the text nodes.
>
> My concern is the explosion of supported encodings required by all
> browsers. What are we talking about a few encodings (2 or 3) or many (10
> or 12). If we can limit it to UTF, JISX208 and whatever you'd call a
> simple 8 bit encoding, then I think it's implementable.
>
Actually, I wouldn't limit it. I would make the tag refer to a
font generating object with a URL pointing to where on the Net you can
get the object if you don't have it already. Browsers would have to
support a standard object format, but that's all. There would be about 3
or 4 popular objects that would cover 95% of the world, but we would
still have the ability to accomidate the rest.
Omar