W3C home > Mailing lists > Public > www-font@w3.org > October to December 2010

Re: i18n-ISSUE-2 (r12a): Why not using xml:lang? [WOFF]

From: Martin J. Dürst <duerst@it.aoyama.ac.jp>
Date: Sun, 12 Dec 2010 18:21:27 +0900
Message-ID: <4D049417.2050905@it.aoyama.ac.jp>
To: Richard Ishida <ishida@w3.org>
CC: www-font@w3.org, public-i18n-core@w3.org
Hello Richard,

I haven't studied things in great detail (and I'm not a lawyer!), but I 
tend to disagree with you.

On 2010/12/10 23:10, Richard Ishida wrote:
> Having given this some more thought, I suspect that the lang attribute in the WOFF spec may not actually be equivalent to the xml:lang attribute.  The xml:lang attribute really says "this is the language of the text in this element".  The woff lang attribute may be saying that by default in many cases, but is actually saying "this is the element to use for a given user". These things could be very different if, say, you had to use slightly different licensing text for Hungarian users but you didn't have a means to translate your license text into Hungarian. [1]

Well, hypothetically, such a case may indeed exist. But my guess is that 
a) it's a bad (essentially unworkable) idea for *Web* fonts to have 
different licenses for different countries,... and b) tweaking a license 
usually involves lawyers. While translation is expensive, lawyers can 
easily be much more expensive.

In many cases, it may also be easier to express things that apply to 
Hungary in Hungarian, first because you can then use the Hungarian legal 
terms rather than some translation with potentially unclear legal 
meaning. Also, you can make a better claim that your target audience was 
actually able to read and understand the license.

> In that case, I think it would be better to rename this attribute.  Perhaps locale would be a better name.  (The values would remain the same.)

I don't think that locale would fly, because it usually implies a 
combination of language, script, location, and so on, and because the 
location isn't the legally binding location, but the location associated 
with a certain language variant.

So what you would need is something like "legally-applicable-region", in 
most cases just a country code. You might then have subelements for 
various languages, where you could go back to use xml:lang, and would 
address cases such as providing German, French, and Italian text for the 
license applicable to Switzerland, and labeling your "English for 
Hungary" as English.

However, once we would get that far, we would realize that this is the 
start of a structured (rather than textual) license description. As far 
as I understood previous discusssions, the WG didn't want to go in this 
direction. That would bring us back to use xml:lang, to label English 
license text as English and Hungarian text as Hungarian, and (if really 
necessary) to clarify in the text itself what specific provisions of the 
licence may be different for different places around the earth.

Regards,   Martin.

> NB: We then have an additional discussion about whether xml:lang ought to be supported by the schema (even though it is an attribute defined by the XML spec, it still needs to be defined in the schema when used).  xml:lang would be used by things like spell-checkers, voice browsers, etc that need to understand the language of the text they are processing.
>
> RI
>
>
> [1] For more on this kind of distinction, see http://www.w3.org/International/questions/qa-when-xmllang xml:lang in XML document schemas.
>
> ============
> Richard Ishida
> Internationalization Lead
> W3C (World Wide Web Consortium)
>
> http://www.w3.org/International/
> http://rishida.net/
>
>
>
>
>

-- 
#-# Martin J. Dürst, Professor, Aoyama Gakuin University
#-# http://www.sw.it.aoyama.ac.jp   mailto:duerst@it.aoyama.ac.jp
Received on Sunday, 12 December 2010 09:22:19 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Saturday, 11 June 2011 00:14:10 GMT