Re: lynx-dev HTML4.0 and default charset

[David Woolley]

> ...
> You right here, the most notable problem is a proxy which may be in aware
> of document content, and _can_ translate the text/* document to another 
> charset
> and should change META, and we probably may get a situation where
> HTTP charset and META charset both present but different,

In that case, HTML 4.0 says that the HTTP charset still stands and 
the META one has no effect!  This is of course the only sensible way
of doing in transit translation, as META can only be used for HTML
and only if the character set is sufficiently compatible with ASCII,
whereas changing the real header allows translation of any text/*
media type without regard to content.

> ...
> > This is probably politically unacceptable for many service providers
> not a politically inacceptable, but as you wrote previously

There seem to be lots of places where local character sets are assumed
and there is a lot of reaction against MIME (e.g. the Ukranian writer).
It's in that sort of context where a service provider being correct may
lose customers.

Generally, though, I think character sets will always be a problem because
it is difficult to make people think globally even if they are vaguely
aware of characters sets.  It's probably still difficult to make them
think there is any software platform except the one they use.