draft-phillips-langtags-08, process, sp ecifications, "stability", and extensions

[Harald Tveit Alvestrand]

[removing the ietf list, on the theory that all who care have subscribed to 
ietf-languages by now]

--On torsdag, januar 06, 2005 07:42:33 -0800 Peter Constable 
<petercon at microsoft.com> wrote:

> RFC 3066 permits tags that have all kinds of internal structures. That
> is a problem as it will never allow us to derive much useful information
> from a tag with any confidence

actually that was quite deliberate at 1776 time - I wanted to point out 
that the thing we were sending out was a *tag*, not a compendium.
For any given tag, I saw the world divided sharply into two groups:

- Those who knew what the tag meant, and could do something appropriate 
with it. They would have no reason to "derive information from the tag".
- Those who did not know what the tag meant, and would treat it like any 
other unknown tag

The purpose of a public IANA registry was that someone who cared could look 
up a tag in the registry and figure out what it meant, so that it could do 
something appropriate.

The "language-range" construct is a departure from the principle.

3066bis has more-or-less the same principle, but applies it to the 
*components* of a tag rather than to the whole tag.
And you still can't derive script info from the tag "sr".

                    Harald