Yair Friedman <yfriedma@JohnBryce.Co.Il> writes:

> > [5  <text/plain; iso-8859-8 (8bit)>]
> > Hebrew			ωμεν
> > Italiano		Ciao, Buon giorno
> 
> I'm not sure it's wise to display the Italian greetings in Hebrew
> charset, most of the readers probably doesn't have it installed, so they
> won't see it.

But the Italian greeting was in all ASCII, so...

When encoding a message (or a message part), then one chooses a
charset that can encode that message (or part).  If one were to only
use the charset for the non-ASCII words, then one would have to send
gazillion-part messages, which is not what one would want.  One wants
as few parts as possible.

> iso-8859-8 is not a "valid" character set.  The message seems to be
> using iso-8859-8-i (According to RFC 1556), but the whole point of
> sending r2l language messages in (X)Emacs is mute.