W3C home > Mailing lists > Public > www-international@w3.org > October to December 2002

Accept-Language and W3C-Translations

From: Stefan Mintert <stefan@mintert.com>
Date: Wed, 09 Oct 2002 19:44:12 +0200
Message-ID: <3DA46AEC.2B20279D@mintert.com>
To: www-international@w3.org


I have a question/suggestion regarding the translations of W3C specs. I

1) For many W3C specs there are translations available
2) HTTP has a Accept-Language request-header field, which can be used to
get a web document in a preferred language

Now my question: Why is 'Accept-Language' not considered when serving
W3C specs?

Of course, I have an answer: If a client request a document in another
than the original (default) language (let's say 'de' (german), instead
of 'en'), then the german document should be "the same" document as the
english; the only difference should be the language. 

Obviously this is not true for W3C specs, because the only *normative*
version is the english document. If the W3C server would always deliver
the german translation when the preferred language is 'de', then no one
with this setting could see the normative (english) document.

But, wouldn't it be possible to consider 'Accept-Language' *and*
preserve the special status of the normative document? My idea is to
evaluate 'Accept-Language' on the server side and to give the reader the
option to get the translation. When accessing a W3C spec there could be
a small pop-up window (uuh, i don't like pop-ups ;-) reading "This
document is available in *your* language" (of course this sentence
should be in the preferred language). Or instead of the pop-up I'd like
to have a short note in the W3C spec, somewhere at the top, i.e. near
the abstract; the note should contain a link to the other language
version ("Would you like to read <a href="...">German</a>?").

I think this would be appreciated and the idea of the i18n and i18n-geo
initiative, isn't it?

What do you think?

Kind regards,


Received on Wednesday, 9 October 2002 13:41:45 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 21 September 2016 22:37:22 UTC