RE: Authoring Techniques Document

Chaps,

This is a draft proposed response to Charles' comments on the first WD of
Authoring Techniques. Please check before I send. If I hear nothing in a few
days, I'll just send.

RI


 

Hello Charles,

Thankyou for your comments on the first WD of  Authoring Techniques for
XHTML & HTML Internationalization.  Please find responses below.

Note also that we intend to release new working drafts before the AC
meeting.  We have, in the meantime, split the original document up into 3
topic-focussed documents.  We aim to produce more such documents as we
develop the material.  The in-edit versions of the new documents are:
		Characters and Encodings 1.0 
	
http://www.w3.org/International/geo/html-tech/tech-character.html
		Specifying the language of content 1.0 
		http://www.w3.org/International/geo/html-tech/tech-lang.html
		Handling Bidirectional Text 1.0 
		http://www.w3.org/International/geo/html-tech/tech-bidi.html



> -----Original Message-----
> From: www-i18n-comments-request@w3.org 
> [mailto:www-i18n-comments-request@w3.org] On Behalf Of 
> Charles McCathieNevile
> Sent: 11 October 2003 18:11
> To: www-i18n-comments@w3.org
> Subject: Authoring Techniques Document
> 
> 
> A handful of comments on the first draft...
> 
> 0. Congratulations!! This is a great thing, and lots of us 
> are discovering that we would really like this kind of 
> information from a reasonably authorative group.

Thanks !


> 
> 1. Where are the archives for the comments list? They don't 
> seem to be linked in the start of the document, just 
> mentioned in text.
 

We will fix this in the next version. Thanks.


> 
> 2. The section on language tagging -
> <http://www.w3.org/TR/2003/WD-i18n-html-tech-20031009/
> #ri20030218.131124329> - doesn't clarify something that the 
> HTML specification doesn't make clear. If a document contains 
> 2 languages, can you use
> 
>     <html lang="it en">
> 
> or is this wrong?


Multiple languages in the <html> tag are not allowed.  RFC3066 provides the
rules for these values.

It may be possible to do this in the HTTP header - we will look into it -
but it is not clear to us that user agents would know what to do in such a
situation.



> 
> 3. One use of hreflang is to allow for user-side addition of 
> graphic markers. A stylesheet rule such as
> 
>     a[hreflang = en]:before { content: url('engflag.gif') } or
>     a[hreflang = en]:before { content: url('OzFlag.gif') }
> 
> or its XSLT equivalent
> 
> allows people to select the way languages are represented - 
> avoiding a cultural insensitivity that causes many people to 
> otherwise avoid providing any graphic idenitfier - something 
> that would be helpful for low-literacy users in particular, 
> and provide rapid recognition for many people.


Thanks for bringing the use of hreflang to our attention.  We agree that we
should mention this type of usage in the document, although clearly signpost
that it doesn't work for IE and possibly other browsers (which makes its
application limited, unfortunately, at present, though we would like to see
it become more widespread).  

Where we disagree is the use of flags.  These are associated with countries,
rather than languages. We also feel that text would be better than graphics
on most user agents for people who need to view at larger sizes.  Karl
outlines a method whereby the value of the hreflang attribute is used in
square brackets.  This seems better, as it reduces complexity and bandwidth
(albeit slightly).



> 
> 4. No mention is made of the technique of using <link 
> rel="alternate"  
> hreflang="xx" ...
> 
> The link element is now widely implemented in browsers 
> (Internet Explorer and Safari are the only two reasonably 
> well-known browsers that don't implement it), and as a backup 
> for language negotitation (analagous to including charset 
> metadata) it seems worthwhile. It would also enable easier 
> searching for language variants rather than simply offering 
> auto-translated versions. Many large organisations make 
> extensive use of Dublin Core metadata, which would allow them 
> to generate this relatively simply...


Thanks for bringing this up.  We will look into it.  

We know that the link element is recognised in some user agents, but we are
not clear whether user agents do anything with the language related
information specifically.  Do you have more details of how this works? We
plan to write some tests for this.




> 
> 5. Is there a way to provide examples without participating 
> in the task force? I work in several groups who are 
> interested in this work, but where many participants don't 
> speak english, which makes attempting to participate in an 
> english-speaking group a waste of time. Some of them have 
> ready-to-hand examples available, but it isn't clear if there 
> is a good way to submit those without finding a friendly 
> bislama-speaking (or whatever language is relevant) member of 
> the task force.


We're not really clear about what you are asking for here, or who you have
in mind and specifically what they would want to contribute.  Could you
clarify?  

Note also that we are concerned about IPR issues.  WG members have agreed to
the IPR policy on signup.  We are not clear about how to handle general
contributions.


Thanks again for your comments,
Best regards,

Richard Ishida
For the GEO task force


> 
> --
> Charles McCathieNevile                          Fundación Sidar
> charles@sidar.org                                http://www.sidar.org
>

Received on Wednesday, 28 April 2004 12:08:00 UTC