Re: ISO charsets; Unicode

Richard L. Goerwitz (
Thu, 29 Sep 1994 05:45:38 +0100

> ...You have to reckon with the possibility that you will use one
> encoding scheme, but different transliteration systems, depending
> on the language.... I certainly would not want to render Persian
> with the same system used for classical Arabic, though they are
> written with the same basic glyphs.
>I have certainly agree that having _both_ encoding information and
>language information is useful. In my judgement, encoding information
>is crucial (without it, the bits are meaningless) while language
>information is simply useful (having it allows additional
>functionality). Neither is a good substitute for the other.

I agree with this 100%. The problem that remains to be solved is
whether we need a directionality switch. It's been claimed that
MIME's default "visual" encoding standard just doesn't fit into
HTML and SGML's philosophy, because it anticipates too many aspects
of the presentation - the way clients will choose to display things.
But can we do without it?

It would be theoretically cleaner, for sure, if we all just bit
the bullet and did things "right", i.e., used normal byte order
for all languages, relying on the clients and servers to negotiate
how direction changes will be handled.

Richard Goerwitz