On Fri, 25 Mar 2005 20:35:51 -0500, Andrew Cunningham
<andj_c@...> wrote:

> My first and main point, is that UTF-8 is just as suitable as UTF-7.

> I suspect that the errors you are seeing in the UTF-8 tests aren't
> because of UTF-8.

That's how it looked to me.

I had been sending in utf-8 by default, but one frequent correspondent was
seeing an annoying (iirc) "=20" at the end of every text line. I compose
(in Opera) in flowed format, but Opera somehow wraps when it sends, while
(afaik) complying with the flowed-format (MIME?) spec. (Maybe that "=20"
was unrelated to my sending in utf-8.)

> collides with C1 control characters? the codepoints corresponding to c1
> in Unicode like the ISO-8859 series are empty.

{quiet voice} help, please?

yrs trly is puzzled. y.t. just looked at the first two charts in the
Unicode 3.0 book, and it seems that what you say ("empty") also describes
C0 control codes, too; correct? I think I understand that Unicode is not
concerned, any more than it needs to be, with control codes, and, surely,
no printable characters are defined for either C0 or C1 control-code
codepoints. Am I OK?

> Content-Type: text/plain; charset=ISO-8859-1
> Content-Transfer-Encoding: base64
>
> essentially, the utf-8 message was sent out not as utf-8, but ISO-8859-1
> (actually most windows browsers and email clients will really use
> Windows-1252 instead)

That sort of thing is one reason I don't use such software. While I'm not
utterly certain that what I do use does it right, I do have a lot of
confidence.

[African languages]

> additionally in Firefox, we setup some Yahoo specific style rules in
> userContent.css file to force yahoo to use appropriate fonts for the
> languages we are working with.

Very interesting. Nice to learn that FF can do that. Just about sure
Opera could, too.

--
Nicholas Bodley /*|*\ Waltham, Mass.
The curious hermit -- autodidact and polymath