RE: Language tags on root

>
> > But having an empty xml:lang does the same thing (effectively).
>
> Very much, but with a twist in the context of requiring xml:lang on the
> <html> element (with the goal of increasing the near-zero fraction of
> web pages that are tagged).  If xml:lang is required but allowed to be
> empty, common practice may very well move from "not tagged" to "tagged
> with empty", an illusory gain.  xml:lang should therefore be required to
> be non-empty, but then "mul" must be allowed for multilingual documents
> with no single primary language or sensible default processing language.
>
Another side twist on this would be that that if (when) draft-langtags
supercedes RFC 3066, a 'mul' primary language could be combined with useful
information about the document for processing purposes, notably the ISO
15924 script code. A document tagged 'mul-Arab' conveys a base direction of
RTL almost as effectively as 'ar' does.

On the other hand, I'm dubious about requiring information that is likely to
be wrong much of the time. While I like the idea of requiring language
information in documents, I'd be willing to guess that many documents will
be tagged with values that are not 'mul' and not correct either if the
attribute is required. I see this frequently with the much more important
encoding attribute in the XML document declaration and I'd hate to see us
mandating more of the same.

<span xml:lang="la">Caveat</span> user still strikes me as a better practice
than requiring users to lie about their content :-).

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 21:39:46 UTC