[MlMt] Writing in a format that is not utf 8

[Benny Kjær Nielsen]

On 9 Feb 2012, at 16:31, Bill Cole wrote:

> On 9 Feb 2012, at 3:11, Benny Kjær Nielsen wrote:
>
>> On 9 Feb 2012, at 8:19, Alan Schmitt wrote:
>>
>>> I interact with a computer program by email (a Diplomacy judge to be 
>>> precise) that mangles text with accents. It seems it's because I 
>>> send the messages as utf 8 but it interprets and forwards them as 
>>> ISO-8859-1. Is there a way to tell MailMate to send a particular 
>>> message in a different encoding?
>>
>> No. I was also hoping such problems were an issue of the past.
>
> That's a refreshing bit of optimism, but sadly it does not yet reflect 
> the real world. Despite a wealth of obviously right ways to do things 
> like picking a message charset and encoding, legacy code and sclerotic 
> wetware persist in perverse bugginess.

Given the number of workarounds in MailMate already I'll probably soon 
lose any sense of optimism :-)

>> Anyone else on the list with similar problems?
>
> Yes and no. I have not encountered this yet with MM because I had a 
> very slow transition from TBird, but there is a small subset of my 
> *human* correspondents whose mail clients know nothing of utf-8 and 
> who instead use either some Windows 'code page' (CP1252 being one, aka 
> 'almost Latin-1', but I've also had need to use Indian and Chinese 
> encodings ) or can only handle pure ASCII as the Elder Gods intended. 
> This only forces me into a specific encoding randomly a few times per 
> year but when I need a non-utf-8 encoding, I REALLY need it. It would 
> probably be adequate for my needs to have a mode to 
> force/squish/transliterate a message into ASCII since even when I am 
> mailing someone who prefers Hindi or Cantonese, I am always only 
> writing US English. (I am cultural imperialism in action!)

Since ascii is a subset of utf-8 then it may be “good enough” in 
many cases even if the recipient client does not support utf-8. Unless 
of course it refuses to display the content.

Just to be clear, a message which only contains ascii has no charset in 
its Content-Type, so that should always work.

And also, please don't get me wrong. I do not oppose some kind of 
charset feature (as long as the default is utf-8), but it is not (yet) a 
high priority.

-- 
Benny