Re: i18n help: Content negotiation?

["D. Dale Gulledge" <dsplat rochester rr com>]

Jeff Waugh wrote:
> 
> <quote who="D. Dale Gulledge">
> 
> > For anyone looking a large-scale example of use of both on a site that is
> > working well, SourceForge has done it.  I'm sure that at least some of the
> > people on this list are aware of that.  While I'm assuming that they are
> > using Apache, I know they are using content negotiation.
> 
> Yeah, unfortunately, that's a massively dynamic solution, using content
> negotiation as 'hints'. I fear a similar solution is going to be the only
> way.

I wonder if Google relies on something less intensive.

There are a variety of degrees of language selection for a web site:

1) None, you get the language the site is in.  For sites that serve a
particular, limited market, or for which translation is too expensive,
this is a viable option.
2) Manual language selection.  At some point you can pick your language
and the site honors your selection.  Cookies are the obvious mechanism
for this.  It puts more of a burden on the server.
3) Content negotiation.  This requires that the user's browser supports
it, and that the server does.  These days, that isn't a huge burden.  It
may not offer non-default languages to wget and robots.
4) Content negotiation with manual overriding, which is what we are
discussing.  It is obviously possible.  It can create some interesting
situations.  For example, I get different languages from SourceForge
depending on whether I am logged in.  Obviously, it is possible to do
it, and do it pretty well.  But as you have pointed out, there is a
price to be paid by the server.

-- 
D. Dale Gulledge, Sr. Programmer,
dsplat@rochester.rr.com
C, C++, Perl, Unix (AIX, Linux), Oracle, Java,
Internationalization (i18n), Awk.