Text 6152, 59 rader
Skriven 2013-01-28 15:43:50 av mark lewis (1:3634/12.42)
   Kommentar till text 6139 av Sergey Dorofeev (2:5020/12000)
Ärende: Z4
==========
 SD> orig.message to echo FIDONEWS on 28 Jan 13  03:44:30
 SD>> Its in vocabulary but we have different defaults. So CHRS kludge 
 SD>> is obligatory element to see the symbol right. Or area specific
 SD>> override must be configured manually.

 ml> your reader should know your local default and be able to 
 ml> translate the message you are reading to the same characters in 
 ml> your local default character set if they exist in it or it should 
 ml> substitute if they do not exist...

 SD> My reader uses UTF-8-able console for output so I need only to 
 SD> convert to UNICODE. But two things exist without which it is 
 SD> impossible get to UNICODE properly: 
 SD> 1) The message charset must be specified and

this is where you must assume a certain character set for those messages that
/do not/ have a character set defined in them... some editors have the ability
to do this on an area-by-area basis while others are more primitive and have
only one slot for this...

 SD> 2) Specification must not be a lie.

you takes yer chances when there's no character set specified... actually, the
sender takes their chances because others might not be able to read their
messages...

 ml> in other words, when reading, your reader should have already 
 ml> translated CP437 to CP866 by the time your eyes focus on the 
 ml> text... when you write, your reader will put CP866 in the message 
 ml> and other readers should translate that to their local character 
 ml> set when they read it...

 SD> My usage of CP866 ended about 10 years ago...

ok... but what about everyone else that you communicate with? ;)

 ml> generally speaking, you can go from most any character set to most 
 ml> any other character set... /generally/... but in some cases, you 
 ml> will loose characters unless there's a way you can translate to 
 ml> multiples (eg: ‘ = ae, the trademark symbol would become (tm), the 
 ml> copyright symbol would become (c) and similar)...

 SD> That's not my case, but such translation may be useful for that who
 SD> use 8-bit encodings. But, again, original codepage must be honored,
 SD> or we will finish replacing russian letters with (tm), (c), o:, or
 SD> some other garbage. 

yeah, and? that's what's been done naturally for 20+ years now... that's when
folks get fussed at when their editor trashes things... folks have to work
together... those wanting to ride on the bleeding edge of the razor blade can
do so with others who want to ride there as well but when communicating with
folks who don't want to be up there, then they need to back down and use
something more compliant... either that or remain in their world separated by
the same language written with different code...

)\/(ark

--- 
 * Origin:  (1:3634/12.42)