Swiss german, spoken

JFC (Jefsey) Morfin jefsey at
Tue Jun 14 15:39:52 CEST 2005

At 05:36 14/06/2005, Harald Tveit Alvestrand wrote:
>--On 13. juni 2005 13:23 -0700 Karen_Broome at wrote:
>>Will a MIME type help me distinguish between Arabic, Devangari, Latin,
>>Traditional Chinese, or Simplified Chinese in the XML format?
>Nope - but if that's the question, you've already decided that you've got 
>written text.....
>>Can it be
>>used with the xml:lang attribute? That's where I need this data.
>I may be at the tail end of a losing battle, but I still haven't bought 
>off on the idea that all information containing an information object 
>needs to be put into the xml:lang attribute. I've been having this 
>discussion since RFC 1766, so it's not a new debate.....

Hi! Harald,
Interesting since this is precisely what langtags seem to intend. It would 
be far more flexible (application wise) to separate concepts and values of 
different kinds. With an xml:script, xml:region, etc. Network wise giving a 
name to a lingustic community and to its common interperson protocol is 
however also of interest, but not necessarily of the same nature (similar 
to a structural TLD).

>>There is
>>often a one-to-many relationship between the spoken language and its
>>written variants and these written variants must be described. Nothing I
>>do is intended for use in e-mail.
>Yep. I think we don't have a disagreement - when I read through the part 
>of the debate that was downloaded when I uploaded my comment, it seems 
>that the consensus was pretty strong that scripts were properties of text, 
>and "written/spoken" was a property of the media.....

Why this dichotomy? Texts are (and are more and more) part of the media. 
There may be the typographers and the publishers, the shooters and the 
distributors, etc. but there is only one public. A language may be 
scripted, written, spoken, talked, what ever you want, it is first used.


More information about the Ietf-languages mailing list