Re: Fallback to UTF-8

Andreas Prilop wrote:

> There is *no* disastrous side effect with the validator.
> The validator reports "non SGML character number ..."
> and this is a fine and useful error report.

It's not when the page in question clearly says that it's
windows-1252, overruled by a stupid Web server claiming
that it is Latin-1, resulting in said error report with
say Nikita.  

It's a disaster that HTML 5 is forced to replace Latin-1
by windows-1252, because nobody uses Latin-1 outside of
your test page.  It's an unmitigated disaster if 2616bis
sticks to "default Latin-1" with a barrage of SHOULD and
MAY roughly stating that "explicit Latin-1 means dunno,
and no charset implcitly means Latin-1".

Apart from RFC 1123 5.3.6(a) destroying the concept of
responsibility built into RFC 821 that "HTTP default" is
the most egregious disaster in standards I ever heard of.

The usual I18N suspects fixed it one year too late, now
we have to live it for four more *decades* (in 2048 or
later the whole world uses UTF-8 per decree in RFC 2277).

 Frank