W3C home > Mailing lists > Public > public-i18n-core@w3.org > January to March 2005

RE: Language tags on root

From: Richard Ishida <ishida@w3.org>
Date: Thu, 27 Jan 2005 20:06:31 -0000
To: <aphillips@webmethods.com>, "'Misha Wolf'" <Misha.Wolf@reuters.com>, <public-i18n-core@w3.org>, "'i18n IG'" <w3c-i18n-ig@w3.org>, <www-html@w3.org>
Message-Id: <20050127200630.997914EF78@homer.w3.org>

> From: w3c-i18n-ig-request@w3.org 
> [mailto:w3c-i18n-ig-request@w3.org] On Behalf Of Addison Phillips [wM]
> Sent: 27 January 2005 17:35
> To: Misha Wolf; public-i18n-core@w3.org; i18n IG; www-html@w3.org
> Subject: RE: Language tags on root
> A couple of notes on Misha's comments.
> > The code for multiple languages is "mul":
> >    http://www.loc.gov/standards/iso639-2/englangn.html#mn
> Language tags and ranges containing the primary language 
> "mul" (and its friend "und" for undetermined) are deprecated 
> by RFC 3066 (item 6 in section 2.3). Tagging a document's 
> root element with "mul" is essentially meaningless. It 
> doesn't convey any information about how to render or process 
> the document, content base directionality, and so forth. It 
> does indicate that multiple languages exist in the document 
> and that none of them are to be considered the primary 
> language. But having an empty xml:lang does the same thing 
> (effectively).
> I think it is important to separate (as W3C-I18N-GEO does) 
> the processing language (the <html> element) from meta data 
> about the page (the <meta> element). 

Or the HTTP header. (Which is actually used in some instances by browsers,
unlike the meta tag.)  For the information that Addison is referring to, see

> The latter might very 
> well contain multiple languages. Content authors must then 
> consider how best to label a document.
> For example, an English document commenting on a French novel 
> might include more text in French than English if substantial 
> text is quoted. The <html> element would probably be set to 
> "en" and the metadata about the document might also say "en" 
> (or "en, fr; q=0.8", since knowledge of French seems to be at 
> least useful in reading the document).

You could do this, though I'd be inclined to say this was an English
document, since the intended consumer is English - not French (otherwise the
commentary would be in French).

Where you really get into multilingual declarations is when you have
parallel content in multiple languages.

(Note, also, as an finniky aside, that there are pages aimed at, say, an
Indian audience where the majority of true content is in,say, Kannada, but
the navigational wrapper is in, say, English. In such documents it is less
clear which language should be used for the <html> element.  This is an edge
case, though.)

> > Looking quickly through the public XHTML 2 draft, I couldn't find a 
> > mention of http-equiv.  Is it still there?
> No, I don't believe it is in there. If I understand 
> correctly, the <meta> element changes how it deals with 
> metadata in XHTML2, divorcing itself somewhat from HTTP. The 
> ability to describe the content language can still be 
> accomplished by defining a property. However, I don't believe 
> the XHTML2 draft contains a pre-defined property for document 
> language and we should probably add that to I18N-Core's 

See comment #38.

> comments. Perhaps something like:
> <meta property="documentLanguage" content="en" /> <!-- the 
> primary language, could be a list of languages -->
> and maybe also:
> <meta property="contentLanguage" content="fr, tlh-AQ, mn-Cyrl-MN" />

Or we could ask for a new element such as <documentlanguage> that supports
multiple language values, rather than go this indirect route.

> Note that I18N-Core WG's comments on XHTML2 are at:
> http://www.w3.org/International/2004/10/xhtml2-i18n-review.html

Note (particularly the www-html folks who are copied on this email) that
those comments are still being formulated by the i18n WG, and have not yet
been formally supplied to the HTML WG for consideration.)

> We welcome additional comments on this and other 
> specifications. For a list of documents in various stages of 
> last call, please refer to our review
> radar:
> http://www.w3.org/International/core/reviews.html
> Best Regards,
> Addison
> Addison P. Phillips
> Director, Globalization Architecture
> http://www.webMethods.com
> Chair, W3C Internationalization Core Working Group 
> http://www.w3.org/International
> Internationalization is an architecture.
> It is not a feature.
Received on Thursday, 27 January 2005 20:06:33 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:01:07 UTC