Re: International Document Server Support
Tony Sanders <sanders@bsdi.com>
Errors-To: sanders@bsdi.com
Errors-To: sanders@bsdi.com
Message-id: <199312062110.PAA06464@austin.BSDI.COM>
To: Jonathan Abbey <broccol@arlut.utexas.edu>,
Tim Berners-Lee <timbl@nxoc01.cern.ch>
Cc: www-talk@nxoc01.cern.ch
Subject: Re: International Document Server Support
In-Reply-To: Jonathan Abbey's message of Mon, 06 Dec 1993 13:28:10 CST.
Errors-To: sanders@bsdi.com
Reply-To: www-talk@nxoc01.cern.ch
Organization: Berkeley Software Design, Inc.
Date: Mon, 06 Dec 1993 15:10:09 -0600
From: Tony Sanders <sanders@bsdi.com>
> > Well, you wouldn't use "language/english" but pretty close. What you
> > do is pass this information along with the Accept: headers, e.g.:
> > Accept: plain/text,language="ISO3316",charset="..."
>
> So, English is now officially called "ISO3316"? Sort of loses some
> of the flavor.. 8-)
He he! Just to clarify ISO3316 is 'language codes' and ISO639 is 'country
codes'. What you end up with is something like en_US or en_UK. Un*x
based software could use $LANG to choose a default since at least one
vendor is using it for this purpose.
http://info.cern.ch/hypertext/WWW/Protocols/HTTP/HTRQ_Headers.html
still says ISOXXX under Accept-Language. I think that's where it was,
can't reach the server to verify, darn it.
TimBL should make a ruling on this so we can all head in one direction:
Someone else pointed out Accept-Language: but I think this is the wrong
approch. I think this information should be in the Accept: header along
with the other client profile information. I can see the arguments for
special casing headers for some client profile but I would opt to be
consistent. I'll be happy either way, I just want us to make an informed
decision.
--sanders