Doug Ewell dewell at
Tue Sep 28 05:10:34 CEST 2004

Georg Schweizer <gschweizer at gmx dot at> wrote:

> The code "gem-CH" should not be used for Swiss variants of
> "gem"-languages other than Alemannic. In this extremely rare case
> the choice of a more specific code is recommended.

I really don't like this part in particular, because the tag "gem-CH" is
already defined by the generative mechanism to mean "Germanic languages
(Other) as used in Switzerland."  Restricting it to one specific
language violates the generative principle, and could possibly cause
valid tagging of some real data to become invalid, which I thought was a
cardinal no-no.

I know the sign-language tags registered in 2001 set a precedent for
this, by taking a generated tag like "sgn-US" (literally "sign languages
as used in the United States") and restricting it to mean only American
Sign Language.  But I was under the impression that this was a rather
special case, owing to the special nature of sign languages.  I would be
surprised if this approach were considered desirable for ordinary
written and spoken languages like Schwyzerdütsch.

Perhaps this particular request could be put on hold until (a) ISO
639/JAC can be consulted and (b) RFC 3066bis is approved (assuming that
is imminent, of course).

-Doug Ewell
 Fullerton, California

More information about the Ietf-languages mailing list