W3C home > Mailing lists > Public > www-international@w3.org > October to December 2004

Re: German as used in Liechtenstein (was: Re: Language Identifier List up for comments)

From: Martin Duerst <duerst@w3.org>
Date: Wed, 22 Dec 2004 18:17:51 +0900
Message-Id: <>
To: "JFC (Jefsey) Morfin" <jefsey@jefsey.com>(by way of Martin Duerst <duerst@w3.org>), www-international@w3.org

Hello Jefsey,

Others have already explained why full and complete exactness as
you are looking for isn't appropriate. I fully agree, and want to
add some more arguments here:

- Languages are know to be very difficult to define and delineate.
- On the Internet, we need something that works and does it's job.
   Asking for too much precision is a recipie for problems; we already
   have a hard time convincing people that they should tag their content
   with language; if one needed a PhD in linguistics to do this correctly
   (rather than just the willingness and time to do a reasonable, common-
   sense job based on some good advice), the percentage of usefully
   tagged content would decrease much further.
- For network technology and data exchange to work, there best be a
   network effect. The network effect is greatest if with low effort
   (read simple tagging), we can achieve a good return (reasonable
   results for searching, text-to-speach, finetuned rendering,...).

Regards,   Martin.

At 07:23 04/12/22, JFC (Jefsey) Morfin wrote:
 >Dear Martin,
 >I am sorry to object. Tagging means to deliver the users (end user, 
developper, application) with an exact information grid. IMHO exact means 
independent from subjective and accidental external information and 
complete. The identification is about "language.country". There are 239 
country codes in ISO-3166 and 6000 or so languages in ISO 639, this means 
that there are 239x6000 tags.
 >Now, you can come with equivalence tables from your own expertise or from 
authoritative concerned bodies. These tables will be subject to changes 
(for example if a law changes on the way to spell some words) and may be 
subject to some additional tag elements such as:
 >- legal
 >- common use
 >- accepted international
 >- academic
 >- historic
 >- etc.
 >These for example could be maintained as a semantic web application using 
the RFC 3066bis taging as it workframe, concerned academic authoritative 
sources providing updated correspondance evaluations and comments. Word 
processors, OS, etc. manufacturers could then indicate if they support it 
or not, etc.
 >So, you may be fully right, but this should be addressed in an 
equivalence table or matrix or web. In a nutshell: we discuss a tag, then 
this tag can be differently used in different frameworks.
 >Or am I wrong?
Received on Wednesday, 22 December 2004 09:31:22 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 21 September 2016 22:37:24 UTC