[rfc-i] open issues: character sets of examples

[Martin Rex]

John Levine wrote:
[ Charset UTF-8 unsupported, converting... ]
> >I have my LANG variable set "correctly" in order to display
> >20 years of EMail, News, docs and Sourcecode and occasionally filenames
> >that contains iso8859-1 chars (predominantly german umlauts, but
> >also some others).   Switching to UTF-8 locale is simply not an option,
> >because it would break a LOTS of things with exactly ZERO benefit,
> 
> Wow, what a lot of broken software.  I use Alpine in a gnome-term in
> UTF-8 mode, and it works fine.  Alpine is quite able to look at the
> character sets in message headers and render them appropriately, even
> when they're not UTF-8.

Nope -- your assumptions are flawed.

A lot of my Mail and News archives do not have mime headers or charset
specifier, and a fraction of it uses an incorrect charset specifier.

C89-Source code, TXT files and filenames simply do not have metadata
that indicates a charset, it is a historical convention.

LANG is a stupid "pass-the-buck" game that breaks for lots of reasons
in lots of places.  Backward-incompatible changes to existing APIs
and existing software is an unconditionally bad idea because it
assumes a flag day where everything (software AND data) stops
to exist as "old installed base" and starts to exist as
"new installed base".


-Martin