RE: Accept-Language and q values

> >[Carrasco]  
> >For most applications one does not probably needs the q and the order
> >is sufficient; indeed, I do not know of any browser that supports this.
> >But then, one cannot express "English" *or* "Spanish" on the same level;
> >i.e., without order. I cannot see when one could need to express
> >"no preference", but one never knows.
> >
> 
> 
> I work with situations everyday where you would want to express 
> "no preference" as you call it.
> 
> That sistuation is public access PCs in public libraries and educational 
> institutions. The municiplaity I work for, although very small, has at
> least
> 63 langauge groups. We attempt to provide multilingual public internet 
> access in our libraries.
> 
> We cannot make assumptions about the language our users are using. 
> Therefore best for us if the langaueg preferences of the browser were set
> to null.
> Getting a web page served up based on the browsers langauge preferences, 
> could mean getthing the wrong page served up.
> 
[Carrasco]  

Yes. Just set the browser so there is no language preference and
(for Apache) set the default language in the server to a dummy
language that just send the list of available languages.

Schematically the situation is a follow:
 
 1) Browser with no language preference

   1.1) Server with default language
      Send the default language.
      It works.

   1.2) Server without default language
     Send the list of available language
     In Apache, use the work around with default to a dummy language

 2) Browser with more than one "equal" language
   It cannot be expressed without the q.
   I do not know of applications with this need.
   I do not know of browsers that allows the setting of q.
   Browsers allow only an order of preference.

   2.1) Requested languages not available and server with default language
     Send the default language.
     It should work.

   2.2) More than one of the requested languages available
     Have a look to RFC 2296
       "HTTP Remote Variant Selection Algorithm -- RVSA/1.0"
       http://gewis.win.tue.nl/~koen/conneg/rfc2296.html 

Regards  
Tomas

Received on Tuesday, 1 February 2000 06:07:23 UTC