Unilingua

[Michael Everson]

At 00:20 -0700 2005-09-16, Tex Texin wrote:

>Why do we want to register things that have no practical use or 
>significance, for which there are almost no documents to give the 
>tag to, and yet make our software tables larger and require more 
>time to explain what it represents than the value of recognizing the 
>code?

Minority languages have practical use and significance, and 639-3 
handles all of them. There aren't many conlangs, and I can't see any 
harm in supporting them with language codes.

>Isn't it ok to have some number of documents for which we say, yes 
>the contents are in a language which isn't covered by tags, so if 
>you want a
>description it needs to be annotated in some other way.

I don't think it's OK, no.

>We should reel the registry back into being something that internet 
>engineers need for practical internet applications and have some 
>form of 80/20 rule related to language categorization.

Internet engineers are having to be dragged kicking and screaming 
into IDN support, now aren't they? Sorry, Tex, but I think it's right 
that resources are made available for support of

>I recognize the needs of
>linguists to distinguish languages with subtle but important differences,
>but I don't see that general software or internet applications should be
>burdened with the overhead.

Real people speak minority languages, and wish them to be supported 
on computers. Go to http://www.google.com/advanced_search? Google 
still doesn't allow searches in Irish, never mind kiSwahili or 
Bambara.

>This has all got to fit in my watch someday. The registry should not 
>be a museum for every possible variant that ever existed or was 
>postulated.

Unilingua isn't a variant, is it?.

>Maybe in addition to 50 documents to register a tag we should 
>require there be 50 engineers that testify they care to recognize 
>the distinction. (kidding, but only slightly...)

This isn't funny.
-- 
Michael Everson * http://www.evertype.com