Request: Language Code "de-DE-1996"

Peter_Constable@sil.org Peter_Constable@sil.org
Tue, 23 Apr 2002 16:26:45 -0500


On 04/23/2002 02:09:26 PM "A. Vine" wrote:

>As far as I know, these are language tags.

In the paper I'm presenting at IUC21, I'm basically suggesting that
"language tags" is too ambiguous and that we need to be clear about what
we're referring to since it's often not an individual language but rather
some derivative notion, usually a writing system or orthography.


>  That they have extra info regarding
>orthography is handy for certain apps.  I can easily see a reason to use
>de-1901, de-DE-1901, de-CH-1901, as distinct tags due to the language
>information (there are differences among the Germans, even at a national
level,
>more so at a regional level, but we haven't defined regions yet).
>
>de-1901 is when you have no better information, or nothing more specific
resides
>in the data.

Presumably an author knows when she/he is creating the data what the
specific orthography is that they're using. The uncertainty can only
existing when one is inspecting exsting, untagged data. Is it really likely
that one will have data from which it can be determined that the spelling
follows 1901 conventions but it can't be determined which particular
country's spelling conventions were used?


> This is a situation which exists.  The alternative is not to tag
>at all, or to tag potentially incorrectly.  Why do that?

I'm certainly not suggesting that we not tag data and especially not that
we tag it incorrectly.


>Also, considering the
>history of this kind of tagging, people are bound to assume that if
there's a
>de-DE-1901, then there's a de-1901, and they'll use it, whether or not
it's
>defined.

By that reasoning, people should be assuming a tag "scouse" from the
existence of "en-scouse", "lojban" from the existence of "art-lojban",
"sgn-nl" from the existence of "sgn-BE-nl", etc. I am quite unconvinced.


>Sooooo, are you thinking the tag is only for orthography?  Or that
orthography
>is the more important information, over and above country-specific?

No. I'm thinking that the tag is used for various different kinds of
things, and therefore when we create the tags we need to be clear about
what *kind* of thing we're intending the tag to denote. I'm also wanting to
avoid ending up with ambiguous tags that refer to non-things that might get
created simply because it seemed like a good thing at the time.




- Peter


---------------------------------------------------------------------------
Peter Constable

Non-Roman Script Initiative, SIL International
7500 W. Camp Wisdom Rd., Dallas, TX 75236, USA
Tel: +1 972 708 7485
E-mail: <peter_constable@sil.org>