The accept-language http header and the document encoding are separate
concepts.
Internet Explorer 4 and 5 do not care about the document encoding, as long
as the system support to decode the octets the document is composed of, is
present. The decoding support, if not present in a default install, is
provided by the language pack add-ons.
The accept-language header does not indicate a preferred encoding, it only
indicates the preferred language. The document author needs to decide on the
preferred encoding and can take into accout:
- the content of the document (which scripts are used, which language(s),
size, document authoring environment, sources of document content, etc....)
- the version of the browser (i.e. do not send UTF-8 to version 3 and below
browser)
The version of the browser is also transmitted in the http header, sent with
every GET.
----- Original Message -----
From: Christopher John Fynn <cfynn@dircon.co.uk>
To: Unicode List <unicode@unicode.org>
Cc: Jonathan Rosenne <rosenne@qsm.co.il>
Sent: Friday, December 03, 1999 3:37 AM
Subject: Re: Multilingual Documents [was: HTML forms and UTF-8]
> Of course UTF-8 is not a language but the only way of configuring IE5
> to request a particular encoding from sites which serve multiple versions
> of page seems to be by setting preferred language(s). The user can choose
> the encoding used to read the page once you have received it if IE5 gets
> it wrong.
>
> My point was simply that a site serving different encoded versions of
> a page is unlikely to get requests for UTF-8 as long as most client
> browsers cannot even be set to request UTF-8 by preference and that
> even new browsers don't always have this option.
>
> - Chris
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:56 EDT