Re: SM

Dave J Woolley:

> > > [DJW:]  Browsers are not required to understand
> > > the CP 1252 encoding,

> > Neither are they required to understand any other encoding,
> > not even iso-8859-1 (Latin 1). So what's the special issue
> > with CP 1252?

> [DJW:]  HTTP text/html without a charset is ISO 8859/1, which
> more or less means that every browser has to support
> 8859/1.

Latin 1 is in practice the most important character code, and
it's indeed the default as far as servers go, but it's not
a requirement that browsers understand Latin 1. The HTML norm
does not state which encodings browsers must understand.
Actually it explicitly states that it does not make any such
requirements. So as far as XHTML is concerned, using Latin 1
is neither more or less correct than using Windows-1252.

There are many more character codes than Latin 1, Windows-1252
and Unicode. Some of them use the code points 128-159 that are
undefined in Unicode. It's perfectly correct and alright to
use those encodings. The main character code for Russian, KOI8-R,
is one of them. It's not bad practice to use KOI8-R, and it's
not wrong to use Windows-1252, as long as you correctly announce
the encoding you use.

Of course I would advice to use Latin 1 rather than Windows-1252
for most normal cases. _But_ if a display of a trademark character
is important, then Windows-1252 (correctly used) will give better
results across browsers and platforms than Unicode (in any form) -
today.

I doubt there is any browser on any platform that chokes on a 
Windows-1252 trademark character, but is able to display it if 
another encoding is used. I would be interested to hear of any 
exceptions.

#####################################################################
                         Bertilo Wennergren
                 <http://purl.oclc.org/net/bertilo>
                     <bertilow@hem.passagen.se>
#####################################################################

Received on Friday, 5 May 2000 18:30:30 UTC