[html-tech-char] "Always" too strong for inline declarations

Hi,

  http://www.w3.org/TR/i18n-html-tech-char/#ri20040215.101337371 in
the 9 May 2004 Working Draft states

[...]
  If declaring the character encoding in the HTTP header,
  always declare the encoding inside the document too.
[...]

While there are good reasons to do so, as the text points out, it is
certainly not always a good idea or possible to do that. For example,
an author knows that all documents on his server are ISO-8859-1 encoded
but got convinced to switch to UTF-8. He could just transcode all
documents using a generic transcoding tool and configure the server to
declare the proper encoding. There is little wrong with doing so, but
the text clearly implies that this would be a major flaw. So please
change the text to say something weaker than "always".

It should also be clearly pointed out that it is the responsibility of
the content provider to properly declare the encoding of the content,
and if he does so, everything should be fine. For example, if the
content provider provides his web site on CD-ROM media, specifying
the encoding inline is a good idea. However, if the concern is that
someone might download the web site using some HTTP mirror software and
then use the documents locally, it is the responsibility of that
software to take care that the encoding information is not lost.

Please change the text to reflect this to a reasoable extend.

regards.

Received on Tuesday, 13 July 2004 20:54:53 UTC