[ On Tuesday, September 10, 2002 at 13:03:19 (+0900), Jun-ichiro itojun Hagino wrote: ]
> Subject: Re: finger 
>
> >Ok, I put in code to only enable 8-bit display for ISO8859-* codesets.
> >Are there other single-byte 8-bit codesets?
> 
> 	you still aren't sure if the sender (fingerd) and receiver (finger)
> 	are using the same locale - for instance, sender is using iso-8859-5
> 	while receiver is using iso-8859-1.

It is no worse than getting whatever encoded streams of data one gets
from the WHOIS servers for .kr or .jp or no doubt many others too.

If the "sender" really wants to assist the recipient in interpreting the
octets properly then it's trivial to include a wee bit of "clear" text
that tells the human user what LANG setting to use when running the
command, which is no different in practice than being told to use a
special switch to "see" English when making a querying the .jp WHOIS
server.

Indeed allowing the tools to send "arbitrarily encoded" data enables the
users so that they will not only be able to work around the issues of
different encodings at some other level, but also that they will perhaps
come to understand the issues caused by having so many different
character encodings.

Someday maybe everyone will have and regularly use a terminal capable of
handling full multi-byte encoded character streams and which can
simultaneously display any character glyph in every window and these
issues will "go away".  However today many of us still don't really have
that luxury.

-- 
								Greg A. Woods

+1 416 218-0098;            <g.a.woods@ieee.org>;           <woods@robohack.ca>
Planix, Inc. <woods@planix.com>; VE3TCP; Secrets of the Weird <woods@weird.com>