ietf-822
[Top] [All Lists]

Re: Headers: the universal character set option in detail

1991-10-18 05:44:06
  Header-Charset := charset "/" encoding

One of the problems with this proposal is that it would not be very
optimal for Japan. If a Japanese person would put his/her name in
Japanese (e.g. for email within Japan), the encoded name would be
unreadable for people with old mail software. E.g.:

        From: gd0p4Jq0N2Yt09jU524c1p 
<ynk(_at_)ome(_dot_)toshiba(_dot_)co(_dot_)jp>

I think that the following could be the basis for a more reasonable
method:

        Header-Encoding: 1=text/iso2022=base64=gd0p4Jq0N2Yt09jU524c1p
        From: Yasushi Nakahara *1 <ynk(_at_)ome(_dot_)toshiba(_dot_)co(_dot_)jp>

The human receiver could tell his/her software to display both or just
one of the forms of the name. (I.e. this is for DISPLAYING only, the
file itself should never be converted, except for ASCII->EBCDIC, etc.)

This method solves the problem of only being able to use one character
set in each header (which Harald's method does not solve), because you
can have several Header-Encoding numbers:

        To: Keld Simonsen *1 <keld(_at_)dkuug(_dot_)dk>,
                Yasushi Nakahara *2 
<ynk(_at_)ome(_dot_)toshiba(_dot_)co(_dot_)jp>

Of course, some of you may complain that we shouldn't force people to
implement the Japanese form of ISO 2022.

Well, I don't want to force them to do so. ISO 2022 could be supported
in Japan. Outside Japan people can just read the romanized form of the
names (as above).

I think this method solves most of the problems. Let me know if you
have any comments.


Cheers,
EvdP