LANGUAGE TAG REGISTRATION FORM

[Michael Everson]

I am beginning to think that it was a great mistake to encode 
yi-Latn. It appears indeed to have opened the gates to a great 
dumping ground of fixes for underspecifed software. Mark even 
specifically cites the "precedent set by yi-Latn". Thanks. Now the 
Language Tag Review can be *forced* to encode things, right?

For yi-Latn I took Peter Constable's advice and Peter Edberg's 
concept of default script and decided that it might make sense to 
encode some of these different entities. Now we're looking at a whole 
lot of duplicate language codes conflating this stuff.

I think we should go back to first principles and figure out whether 
this RFC is supposed to do this or whether it is supposed to be a 
simple extension to ISO 639.

Mark, Addision, you guys go talk to Peter Constable, Peter Edberg, 
Ken Whistler, Rick McGowan, John Cowan and whomever else you like. 
I'll meet Ken and Rick in Baltimore before the UTC and they can talk 
to me about consensus.

Right now I feel that we have NO consensus, and I feel very uneasy 
about whether these conflations are "language tags" or not. Perhaps I 
am wrong, but something feels wrong about this. I think we need to go 
back to Peter Edberg's contribution and sort out all the bugs before 
we do this.

I suppose this will make me unpopular.
-- 
Michael Everson * * Everson Typography *  * http://www.evertype.com