W3C home > Mailing lists > Public > www-international@w3.org > January to March 2007

RE: Comment on working draft "Specifying Language in XHTML and HTML Content"

From: Richard Ishida <ishida@w3.org>
Date: Wed, 14 Mar 2007 17:17:43 -0000
To: "'Mark Davis'" <mark.davis@icu-project.org>
Cc: "'Jonathan Rosenne'" <rosennej@qsm.co.il>, <www-international@w3.org>
Message-ID: <012001c7665c$ae2197d0$6401a8c0@rishida>

Responses below...

============
Richard Ishida
Internationalization Lead
W3C (World Wide Web Consortium)
 
http://www.w3.org/People/Ishida/
http://www.w3.org/International/
http://people.w3.org/rishida/blog/
http://www.flickr.com/photos/ishida/
 
 

> -----Original Message-----
> From: mark.edward.davis@gmail.com 
> [mailto:mark.edward.davis@gmail.com] On Behalf Of Mark Davis
> Sent: 13 March 2007 20:27
...
> <meta name=" dc.language" content="en" />
> 
> This is extremely obscure. If you really need very obscure 
> stuff, it should be in an appendix, and clarified ("Dublin 
> Core notation"?)

It is, but it's there because people requested it in the past.  I added a paragraph in the next section, which I realised was missing, that should tell designers and developers that this is not currently a highly recommended approach.


> 
> 
> In the meantime, we recommend that you use HTTP headers and 
> meta elements to provide document metadata about the language 
> of the intended audience(s), and language attributes on the 
> html tag to indicate the default text-processing language. 
> Furthermore, we recommend that you always declare the default 
> text-processing language.
> 
> 
> Getting people to do one thing -- correctly -- is hard 
> enough. Asking them to distinguish between the language of 
> the document and that of the intended audience is way too 
> far. Best practices should be aimed at getting the 99% case 
> right, then point to a different document for the edge cases 
> that won't matter anyway because nobody will do them 
> correctly so they can't be relied on. 

I rewrote that whole section to make things simpler, clearer and more to the point, while still addressing the key questions people have.

I also slightly adapted some of the document headings to better focus the attention on the idea that one SHOULD always use attributes, and MIGHT specify metadata only if they feel it useful. The document does have to cover that latter stuff, though.



> Best Practice 13: Using Hans and Hant codes 
> <http://www.w3.org/International/geo/html-tech/tech-lang.html#
ri20040429.113217290> 
> 
> 
> This also goes for some other cases like uz_Arab. You might 
> list the common cases where a language uses multiple scripts.

Yes, I could, but that would entail a significant amount of extra work that I don't want to do at this point. We should consider it for the next version. 



> Best Practice 15: Using hreflang with CSS 
> <http://www.w3.org/International/geo/html-tech/tech-lang.html#
ri20030112.224458239>  > <http://www.w3.org/International/geo/html-tech/tech-lang.html#
ri20050128.152033553> 
> 
> 
> I'm really leery about this one. It is extremely fragile. If 
> you really wanted to mark the language of an HTML document 
> pointed to, this is something that the browser would have a 
> much better job of doing, since it could fetch the start of 
> the page (working in the background) and pick up the actual 
> language used on the page. So if you are evangelizing anyone, 
> I'd think it'd be browser vendors. 

I agree that it's something you should be very careful about, but people are doing it and recommending it.  Note that the best practise doesn't recommend that you do this - it says that you should carefully consider the pros and cons - and I think we point out quite a few cons.  I take the approach that we can't simply dismiss this out of hand, but we can make people think carefully about whether it's the right thing to do.

As for the user agent peeping into the top of a page, that's outside the scope of this document, which is aimed at content authors.

For Martin, I also expanded into an example my existing reference to alternate ways of doing this in addition to using the hreflang value for display.


> 
> Saw a couple of spelling errors also, so you might spell-check.

Oh. Actually I did check.  Please point me to them, if you remember.

Thanks, 
RI

-- 
No virus found in this outgoing message.
Checked by AVG Free Edition.
Version: 7.5.446 / Virus Database: 268.18.11/721 - Release Date: 13/03/2007 16:51
 
Received on Wednesday, 14 March 2007 17:17:41 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 19:17:09 GMT