Problems with character translations

[Adam H. K. <ah...@ch...>]

I'm having trouble with character translations when I'm using PuTTY terminal 
emulator, and I wonder if there's a way to deal with it in trn.

The linux server uses UTF-8 by default. PuTTY doesn't know this, so I have 
to choose this setting myself.

When I read an article, the character set won't necessarily be UTF-8. It 
might be Latin-1, ISO-8859-1 or Windows1252, or the Latin-9 minor variation 
IS0-8859-15.

Again, PuTTY doesn't know this, so I have to reset the translation to match.

Something in my setup is MIME aware, because I don't see strings of code on 
headers when MIME encoded word syntax is used for non-ASCII characters.

When I compose a followup, I invoke vi. Back under server control, so the 
server present UTF-8.

I've learned that UTF-8 translation isn't helpful when composing a followup 
to an article posted through Google Groups, as I don't see the non-ASCII 
cruft Google Groups has added. I'd rather just send ASCII, which also means 
I don't have to add MIME headers. These non-displayable characters display 
when the wrong translation is used, making it easy to delete them.

For some reason, when I have the translation set to UTF-8 and a Latin-1 
non-ASCII character is present, I see a solid box versus the relavent glyph, 
even though the character in question is in the Latin-1 character set and 
I'm using a Western script. For some reason I don't understand, the code 
isn't being translated.

How have trn users solved character translation problems? Is there an 
appropriate MIMECAP entry so I get something displayable as intended?