dewell at adelphia.net
Wed May 28 00:20:14 CEST 2003
Mark Davis <mark dot davis at jtcsv dot com> wrote:
> Here is my best guess as to where people stand on the 9 registrations.
> Please speak up if I got you in the wrong place, since I had to guess
> based on email statements; it was sometimes unclear to me what the
> situation was.
> Mark Davis, Peter Constable, John Cowan, Addison Phillips, Keld
> Simonson, Roozbeh Pournader, Ken Whistler, Martin Duerst, John Clews,
> James Seng, Harald Tveit Alvestrand, Francois Yergeau, Doug Ewell
Umm, sort of. As I said before, what I'd really like to see, much more
than the nine registrations, is an extension of RFC 3066 to allow
generative use of script codes, exactly the way it currently allows
generative use of country codes. That way, I could write "sr-Cyrl" or
"sr-Latn" or just plain "sr", depending on how broadly or narrowly I
want to specify things, just as I can write "en-US" or "en-GB" or just
plain "en" today.
Of course, the generative mechanism would also allow me to write
something ridiculous like "sr-Hang". But surprise, I can already write
the equally ridiculous "sr-KP" and that hasn't caused the sky to fall.
Registering the proposed language-script codes, and then extending 3066
to allow such things to be constructed without the need for
registration, might be an odd situation. The registered codes would not
have to be deprecated -- they'd be more legal than ever, in a way -- but
they'd probably have to be mentioned explicitly in the new RFC.
"yi-Latn" is already in this boat.
So I guess I can be considered on the "pro" side, in the sense that I
believe the codes would be useful. But registering a selected set of
nine codes shouldn't be the end of it, and there's a chance that the
"productive script code" model might meet resistance later on from
people who say, "Just register the language-script codes you need, one
by one, the way you did with yi-Latn, sr-Cyrl, az-Arab..."
BTW, regarding the policy question about mixing country and script
subtags, I think it's clear that the precedent set by "de-AT-1901" and
company should be followed, making the order "language-country-script".
Others have already pointed out that you can't always parse these
left-to-right or right-to-left for "importance" anyway.
More information about the Ietf-languages