W3C home > Mailing lists > Public > www-international@w3.org > July to September 2005

Re: Comments on Authoring Techs for XHTML & HTML I18N

From: Martin Duerst <duerst@it.aoyama.ac.jp>
Date: Wed, 21 Sep 2005 10:23:01 +0900
Message-Id: <6.0.0.20.2.20050921101027.094038f0@localhost>
To: Christophe Strobbe <christophe.strobbe@esat.kuleuven.be>, www-international@w3.org

Some simple personal replies below:

At 02:12 05/09/21, Christophe Strobbe wrote:
 >
 >Hi,
 >
 >I have a few comments and questions on the 24 February 2005 working draft 
of "Authoring Techniques for XHTML & HTML Internationalization: Specifying 
the language of content 1.0" [1]. I hope this does not duplicate anything 
that has been discussed previously (I am new to this list).
 >
 >Example 1 in section 4 and example 11 in Technique 5: there is a space in
 >         xml ns="http://www.w3.org/1999/xhtml"
 >that is not supposed to be there (xmlns="http://www.w3.org/1999/xhtml").
 >
 >Technique 1 or 8: What would you recommend for content that has no 
natural language, e.g. type samples that include Latin, Greek and Cyrillic 
characters? (Joe Clark brought this issue to the attention of the WCAG WG: 
http://lists.w3.org/Archives/Public/w3c-wai-gl/2005AprJun/0144.html.)

Use lang='' / xml:lang='', i.e. the empty string.

 >Technique 1, 8 or 11: What would you recommend for languages that 
currently have no language code?

Register one with IANA (actually, first try with the ISO-639 committee).

 >Technique 3: What would you recommend for transcriptions from other 
languages, for example when transcribing a Russian name in French or Arabic 
text in English? Use French and English, respectively? (There may be more 
than one convention for transcribing e.g. Russian into French, but I don't 
think this is relevant.)

Names are tough; sometimes a name is very clearly attached to a language,
sometimes it's very generic, not really linked to a specific language
(I like to think of my first name that way), and sometimes it's a name
that clearly shows its origins but has long ago been taken over by another
language. So I think it's very much a judgement call.

For running text that is transcribed, you should be able to use just
the language code (the script is obvious from the text anyway) or you
should be able (once the revision of RFC 3066 is completed) to use
a language code combined with a script code. You may also want to
(register and) use something identifying a particular transliteration
or transcription if you care about that level of detail.

 >Technique 14: if both language versions of the same document are under 
control of the same author, it should be much easier to keep the hreflang 
and the actual language of the target document in sync than with documents 
located on different servers. Something along these lines could be added to 
the discussion of this technique.
 >
 >Technique 15: adding the two-letter code for a language by means of the 
CSS content property does not strike me as particulary user friendly. I 
wonder how many users of the Web are familiar with these language codes. 
Also, language codes use Latin alphabet; are users of other scripts 
supposed to know this alphabet?

Good point. The text says "You could just as easily append text
or even a graphic after the link." but doesn't say how to.
It's one point to just copy an attribute value, it's another
to try to change that into a text or graphic. An example would help.


By the way (not for Christophe, but for the GEO WG):

When I was looking at the white and black dots at the start
of the technique, I interpreted the black dots as being okay,
and the white dots as being not okay. Probably, more easy
to interpret icons should be choosen.

Regards,   Martin.

 >
 >[1] http://www.w3.org/TR/2005/WD-i18n-html-tech-lang-20050224/ 
Received on Wednesday, 21 September 2005 04:00:24 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 19:17:05 GMT