Re: "character encoding indication"

On Fri, 4 Feb 2005 paravelli-ext@airfrance.fr wrote:

> I just tried your validator with several pages taken
> from various web site. It seem that the validator
> does not correctly detect the "character encoding
> indication" or/and always detect/force it to "us-ascii"

I think it's more probable that the validator is correct,
or even hypercorrect, about character encoding information,
and the site is not.

> (or maybe did I always made the same mistakes with
> all the differents pages I tried ?) ...

Maybe. In the absence of any specific URLs, one can just make conjectures.
It seems to me that the airfrance.fr site contains some pages that
have their encoding declared as utf-8 in the HTTP headers, yet
claimed to be iso-8859-1 in a <meta> tag. This need not cause any
real harm if all the characters are in fact Ascii characters, but
it's still an information conflict and should be pointed out.
This might contain a time bomb: if anyone enters accented letters into
a document and if it is in fact in iso-8859-1 encoding, then browsers
and other user agents will get (and show) wrong data, since they rely on
the HTTP header (as they shall).

> Exept from that I must say this validator is really a nice
> tool for all web users and designers ! Good job guys !!

Well, I'm not one of the good guys, I'm a critic. I think
the validator is useful to authors who understand the basics
of SGML or XML, useless or even harmful to other authors (though they
might accidentally find some error messages understandable and
useful), and almost surely useless or worse to web users who are not
authors themselves.

-- 
Jukka "Yucca" Korpela, http://www.cs.tut.fi/~jkorpela/

Received on Friday, 4 February 2005 12:34:09 UTC