[Zope3-dev] Re: [Zip]Re: About Zope3 HTTP-input handling

[Heiichiro NAKAMURA]

Hi Tino,


On Wed, 23 Oct 2002 22:30:06 +0200
Tino Wildenhain <tino@wildenhain.de> wrote:

> I actually dont understand the use of more then one locales
> object in the tree. This looks like you want different parts
> of the site having different languages and stuff.

That's right. In today's global economy, even a small-to-mid
company often has deals with several foreign companies.
So, for instance, in Japan, one company often has different
contents for each target region (and it's the recent trend
they want to have contents for other asian countries, Chinese,
Korean, Vietnamine, etc).

Another point is, it is not rare that the content itself is
also significantly "localized", so that almost nothing is shared
(including page template) as the result of effort to satisfy
audience of different cultural preferences.
"Different content for different target audience" makes sense
in today's CRM era ;)

Also, different browser/device often has different set of
requirements (including encoding setttings, formats, etc) and
different behaviours (I can't assume HTTP/1.1 as de facto simply
because many devices don't follow it). Note I don't assume that
IE/NN are the only major browsers since many different Internet-enabled
divices are upcomming.



> My opinion is more like the HTTP language negotiation (rfc2616)
> thinks about it: every variant should transparently be accessible
> at the same URL. The philosophy is: different content: different URL,
> same content (but many representations, say language/encoding variants)
> same URL.

First of all, RFC2616 is a specification, not good as the representation
of the current real situations unfortunately.

In today's environment, it becomes hard to assume the way the users
browse web sites. multiple device (including PDA/mobile-phone), multiple
localtion, multiple situations, etc.


So I made the assumption:
   "Accept-Language" header (and others) doesn't always represent what the
    user really wants to see.

The followings are the situations of it:

Senario 1: A French person visit Japan and go to the "Internet Cafe"
  and trys to see a French web page, but since the PC in the cafe is
  protected (so that he/she can't change the configuration of web browser)
  and can't reach the French page.

Senario 2: A Japanese non-tech person want to see English version of
  a web page since the English version has the newest content (press releas=
e,
  etc).

Senario 3: An Internet-enabled PDA which doesn't implement
  HTTP/1.1 perfectly and no way to change the HTTP Header.

(I can list many more..)

Therefore, the implicit automatic negotiation (including HTTP language
negotiation) shouldn't be the (only) determinant of language and encoding,
although it can be used as one of the suggestion of language to determine
which language should be displayed on the first visit.
It is desirable (actually requirement in most case) to give web-visitors
the explicit choice of language on the home page, and implicit automatic
negotiation is often just a "nice" feature, not a reqiuement.

As far as I know, user's requirement is often fairly complex, and
I belive the platform (such as Zope) shouldn't impose one way for lang-nego=
tiation,
and provide rather flexibility,
and hope Zope3 to be the best platform to build the multilingual web site,
as it should be going to be a good differenciator from other application
servers.


=2E.. are my opinions.



Regards,
---
Heiichiro NAKAMURA <nheiich@quantumfusion.com>