ietf-822
[Top] [All Lists]

re: NEW proposal on encoded-headers

1991-10-16 11:09:13
A step in the right direction, but I do not believe that `:' can be considered
a safe token to use in personal name phrases even if is properly quoted.
There are too many broken implementations that get terribly bent out of shape
by these characters.

Also, suppose you want to use more than one character set?  That is, your
message is going to different individuals whose names are expressed `right'
only in a particular character set, and each of these are different.

There is a presumption, which I think is fatally flawed, that the only
character set which counts is 8-bit ISO.  This is a major weakness of the
TEXT/charset type.  You really should define and implement non-USASCII
character sets in RichText instead.

The other fundamental conflict is that many of the character sets in question
are implemented in hardware.  People want to use their hardware support and
not bother with software details.  The problem is that everyone has his own
nationalistic view of whose hardware counts and whose doesn't count.  At least
the US view is justifiable as being the only one that works everywhere.


<Prev in Thread] Current Thread [Next in Thread>