Louise Bremner wrote:
> Jean-Marc Desperrier <jmdesp@alussinan.org> wrote:
>>So when it receives a message, it identifies what it it, and converts it
>>to the latin-1 before displaying, replacing everything that can not be
>>represented in latin-1 by '?'.
> 
> That doesn't explain how another pair of messages were displayed

I did explain that just after.
Read my message carefully, do not read what you think I wrote :-).

>>When the mesage is missing the correct header for identification like
>>Haluk's was, MacSoup don't know it should convert, so displays it raw.
>>And then your external program can do the conversion.
> 
> Um.... Haluk posted two messages. The first (entitled "Kill Bill" and
> dated: Thu, 8 Jan 2004 23:07:30 +0200) presumably was without the
> correct header and I saw it as ????. 

OK, I will explain more carefully, and taking into acount the case of 
badly sent messages like this one.
That makes three cases :

1 - OE without correct header and wrong encoding selected before
sending :
This is the case of Haluk's first message.

In that situation, OE convert the japanese to '?' before sending.
Nobody on earth will see anything else than '?', because '?' is actually 
what has been sent, the japanese never got out of Haluk's computer, only 
the '?' were sent on the network.

You see the '?' just like anybody else, because that's all there *is* in 
the message.

2 - OE without correct header and correct encoding selected before
sending :
This is the case of Haluk's second message, the one at top of this 
thread. Nothing in the header is different from first case.

In that situation, OE send the japanese correctly, but does not put the 
header, so the software has no elements to guess what encoding it is.

In that case, MacSoup assumes it is ISO-8859-1, assumes it can display 
it without changing anything and display ESC[BrandcomcharsESC[J.

Then your external programm detects that this looks like japanese and 
converts it for display.

You see the japanese.
Other people can see the japanese, but will have in many case to 
manually select japanese encoding, because there's no information in the 
header that the message is in japanese.

3 - OE with correct header and correct encoding selected :
This is not the case of any of Haluk's message, but some other people 
have this correctly configured.

In that situation, OE/some other software send the japanese correctly, 
and puts the header, so receiving software can know what encoding it is.

In that case, MacSoup knows it's japanese, decides it can't display and 
converts every japanese string to '?'.
Your external programm can't do nothing, because all it sees is '?'.

You can't see the japanese.
Other people can see the japanese, without any manipulations.

This is the correct situation and correct way to send messages.
If your setting can not handle it, it is bad.

>>Submit the problem to the MacSoup author, and ask him to ask a mode to
>>force the display of the message as "raw" without conversion to avoid
>>this. 
> 
> He keeps promising full two-byte support in some future version, but I
> have my doubts that will ever happen. 

"raw" display is much, much easier to implement.
This is not a difficult feature, I will send him a message suggesting 
that, and explaining the rationals.