[Zope] i18n site and search robots

Tino Wildenhain tino@wildenhain.de
Wed, 23 Jul 2003 09:13:04 +0200


Hi,

Dieter Maurer wrote:
> Gilles Lenfant wrote at 2003-7-22 15:50 +0200:
>  > This is not strictly speaking a Zope problem, but certainly lots amond you
>  > faced and fixed this.
>  > I made a i18n site with Localizer that runs fairly good, including its i18n
>  > search engine.
>  > But what about external searche engine robots (google, infossek...)
>  > How to "tell" them that they may browse and index the pages in french,
>  > english, spanish (...), changing their http header "Accept-Language" ?
> 
> Not sure, whether this is the most elegant way, but:
> 
>   You could have "language access folders", e.g. "en", "fr", "de".
> 
>   Requests that go through these folders select the corresponding
>   language. A ("SiteAccess") AccessRule in the folders ensures
>   that "Accept-Language" is correctly set in "REQUEST.environ"
>   and that even "absolute_url" generates the correct language
>   specific URLs.
> 



According to the W3C standard, the server would
1.) issue a vary: Accept-Language  header on each request
2.) if no accept-language header is sent, definition requires to send
     300 "Multiple Choices" as status and provide a list of available
     variations
In the multiple choice answer, the list could consist of the said links
to the language-acess folders Dieter proposed.


This would make a good crawler switch.