RE: Regarding update of language declaration tests (I81NWG)

Hi Leif, all:
 
> Date: Wed, 21 Apr 2010 00:47:49 +0200
> From: xn--mlform-iua@xn--mlform-iua.no
> To: cewcathar@hotmail.com
> CC: www-international@w3.org; ishida@w3.org
> Subject: RE: Regarding update of language declaration tests (I81NWG)
> 
> CE Whitehead, Mon, 19 Apr 2010 19:47:57 -0400:
> > I looked at your proposal Leif:
> > http://www.w3.org/html/wg/wiki/ChangeProposals/lang_versus_contentLanguage
> > "The value of the content attribute of the last occurring meta 
> > content-language element must be the empty string."
> > { MY COMMENT: no not really;
> > I think it should optionally be lang="" and some single language 
> > declaration tag;
> 
> Yes. Thanks to your mention of the QA article about "no language", I 
> think I will make some drastic changes to it.
> 
 
This should be changed, yes, so that it can either be lang="und" lang="" or lang="fr" (or "en" or "fr" or "zh" or "ar" or "no" etc.)

> [...]
> >> Thanks for the pointer. Since XHTML5 and HTML5 support the empty 
> >> string, the consequence of the advice in that article, must be that one 
> >> should *not* use "und" in XHTML5 and HTMl5.
> >> The problem, however, is browser support ... They do not seem to care 
> >> about the so called "schema". 'und' has better support than the empty 
> >> string.
> > So -- if the emptry string is disallowed for the meta 
> > content-language element -- 
> > lang='und' will be the best option?
> 
> Absolutely. As I said, I will make some drastic changes to my Change 
> Proposal. It is much easier for me to agree with the *current* text in 
> HTML5, as long as I can use "und". Thanks for bringing that into the 
> debate!

 

Great! But note:
I do not agree with disallowing multiple languages in the meta http-equiv Content language element -- unless there is more than one meta http-equiv Content language; then I would 'disallow' these -- to the extent that a w3c recommendation can do so -- in the last element.

> 
> > (It sort of bugs me to declare the language as 'und' when there are 
> > two known document languages -- neither having preference of course 
> > or that could be the language declaration in the meta 
> > content-language element in this case --; but if 'und' works . . .)
> 
> If you do
> 
> <meta http-equiv="Content-Language" content="und" />
> 
> then all you do is that you defines the *audience* language (because 
> *that* is what - oddly enough - "Content-Language" refers to) as 
> undefined.
> 
> So this,
> 
> <!DOCTYPE html>
> <html 
> lang="<whatever-including-empty-or-deleted>"
> >
> <meta http-equiv="Content-Language" content="und" />
> <!-- et cetera -->
> 
> would be super for me. It basically gives me all I want and need.
 
Glad that you are happy with this.
 
Best,
 
C. E. Whitehead
cewcathar@hotmail.com
> -- 
> leif halvard silli
>

Received on Thursday, 22 April 2010 01:09:56 UTC